geekhack

geekhack Community => Other Geeky Stuff => Topic started by: noisyturtle on Sat, 30 November 2019, 17:36:29

Title: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: noisyturtle on Sat, 30 November 2019, 17:36:29
I'm at the point where my current hardware is starting to not be enough for modern titles and applications. I can still run them, but it's a bit of a slog.
Current rig: IBM 2500k (oc'd)/ GTX 970 (oc'd)/ 16 GB ram

Now I COULD spend about $600 and get a GTX 1070ti and a IBM 3770k, which is just about the limit of what my mobo can handle and basically double my performance (no modern gpu, or anything above 3k IBM series. It can run a 1080ti but the price difference between the 1070 and 1080 is ridiculous,) which would put it about in the middle of modern machines. Maybe get another 2-3 years out of it.

Or spend 2.5x that amount and just build a new PC that would inevitably last slightly longer.

How do you decide when the time is right? Tech keeps improving and getting cheaper at a really fast rate these days, perhaps waiting 2-3 more years is better?
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: absyrd on Sat, 30 November 2019, 17:52:18
I decided when I was trying to game again plus go back to school. I had a 2500k and 960. I sold the gpu and gave my dad the 2500k.

Ran through a few intel and amd setups, and now I'm working with a 3600x and rx 570 (soon to be 5700 xt). I'm having fun starting over. It has been worth it.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: tp4tissue on Sat, 30 November 2019, 17:57:54
Threadripper is 1 wurd..   

If you don't have threadripper, then you need it,  that's how the decision is made.

(https://i.imgur.com/fupGm31.gif)

GPU, in the mean time, 5700xt is best buy,  1660ti is the best buy from Nvidia.

not huge on GPU upgrad, because Tp4 mainly watches movies.


Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: noisyturtle on Sat, 30 November 2019, 18:17:39
Threadripper is 1 wurd..   

If you don't have threadripper, then you need it,  that's how the decision is made.

Show Image
(https://i.imgur.com/fupGm31.gif)


GPU, in the mean time, 5700xt is best buy,  1660ti is the best buy from Nvidia.

not huge on GPU upgrad, because Tp4 mainly watches movies.




believe it or not I've actually never in my life owned an AMD system of any sort, and always feel weird when I use them.

well that's not technically totally true, I do own AMD stock which has been performing very well ;)
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Sniping on Sat, 30 November 2019, 18:26:02
I've had good luck buying and selling used. I refresh my computer every 3 years or so by looking for a new rig and selling off the old one to recoup some costs. It's a bit of a game and takes some hunting but I think it's pretty fun and you can score on a nice rig. I usually search for a part that's a hard requirement for the system, so usually a GPU or ITX and see if the rest of the build fits my needs, and of course for the right price. PC hardware for sure gets outdated long before it wears down, so buying used is a pretty safe bet in my eyes. Maybe I'll get burned one day, but so far so good for me. My current computer is sweet and I bought it for basically the price of the GPU new.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: fohat.digs on Sat, 30 November 2019, 19:45:41
Since I am not a gamer my requirements are not as stringent, but I do audio and video editing so I need decent power.

I bought an excellent huge tower case at least 15 years ago and have been upgrading incrementally ever since. As the father in a family of 4, the process was originally that each time I upgraded the hand-me-downs cascaded downwards and the oldest pieces dropped off.

When my son got to be about 12-13 and started gaming in earnest, his system moved to the top of the food chain and I was first in line for the obsolescent pieces.

That scenario has worked quite well. Motherboard upgrades happen about every 2-3 years, video cards about the same, although staggered in time, hard drives last maybe a little longer, and fans and power supplies kind of give up at random.

This scheme has allowed me to keep multiple computers up and running without ever spending more than a couple of hundred dollars at any one time, but requires that you have a daisy chain wherein there is a hierarchy of user needs. I do like the idea of buying a really good case and power supply and working from there.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: noisyturtle on Sat, 30 November 2019, 19:50:41
Since I am not a gamer my requirements are not as stringent, but I do audio and video editing so I need decent power.

I bought an excellent huge tower case at least 15 years ago and have been upgrading incrementally ever since. As the father in a family of 4, the process was originally that each time I upgraded the hand-me-downs cascaded downwards and the oldest pieces dropped off.

When my son got to be about 12-13 and started gaming in earnest, his system moved to the top of the food chain and I was first in line for the obsolescent pieces.

That scenario has worked quite well. Motherboard upgrades happen about every 2-3 years, video cards about the same, although staggered in time, hard drives last maybe a little longer, and fans and power supplies kind of give up at random.

This scheme has allowed me to keep multiple computers up and running without ever spending more than a couple of hundred dollars at any one time, but requires that you have a daisy chain wherein there is a hierarchy of user needs. I do like the idea of buying a really good case and power supply and working from there.


And it gives your son an excuse to get upgrades, what a racket! :P
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Leslieann on Sun, 01 December 2019, 01:05:01
I'm at the point where my current hardware is starting to not be enough for modern titles and applications. I can still run them, but it's a bit of a slog.
~trim
How do you decide when the time is right? Tech keeps improving and getting cheaper at a really fast rate these days, perhaps waiting 2-3 more years is better?

Get the 1070ti and 3770k but I will add that if you have a spare slot, get an NVME to PCIE adapter (about $10 on Ebay) and an NVME ssd ($40) for your OS.

This will bring you up pretty close to a modern system, allowing you to save for a better/newer mobo/ram/cpu which the ssd and 1070 will plug into later without waiting to get the benefits.


How do I decide?
My last 2 upgrades happened because I had people asking for a nice, used gaming system, so I sold them my mobo/ram/cpu. I've almost never done a completely new system, it's always incremental.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Larken on Sun, 01 December 2019, 03:59:49
Incremental upgrades until it stops being a good proposition in terms of value, then I do a full refresh, reusing certain parts if they're still good. I may buy some extra components to put together a spare system with the older parts for family or as a work system, that's about it.

Imo, the jump from 2500K to 3770K is not worthwhile unless you can get the chip really cheap, since it is already several generations old, and not a huge jump from your 2500K.   

Graphics card wise, usually it's a case of 'you gotta do what you gotta do'. The 970 is definitely still a decent card, but if you aren't getting the performance you desire in games, the 1070Ti is a decent upgrade.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: tp4tissue on Sun, 01 December 2019, 07:22:23
Unless the 1070ti / 1080 is really really cheap,  you wouldn't buy them over the 1660ti,

5700xt is top price/ perf at the moment.

2500k on z68 doesn't support nvme natively, so you'd have to run a modded bios, they're usually available for the popular motherboards.

Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Sintpinty on Sun, 01 December 2019, 13:57:56
I decide when i want to. It's sad knowing someday due to the effects my GPU might be unable to run the Egg Hunt, which i must dedicate 40+ hours and then defeat a boss to get e g g
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Leslieann on Sun, 01 December 2019, 14:18:31
And it gives your son an excuse to get upgrades, what a racket! :P
It's a never ending race.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Leslieann on Sun, 01 December 2019, 14:30:56
Some people brought up some good points (3770k price matters a lot) and TP has some points but is out to spend all your money but then I re-read Ops first post.
Op doesn't game much, so I'm going to revise my advice.

Find a new or used RX 570 or 580, you can get them for $100 or less, this will easily handle your gaming needs.
Then dump the rest into a cpu upgrade.

I'm assuming you have probably $300-400 or so to spend considering what you were looking at.
If you can find one, look into buying someone else's old system, especially between now and Feb since many will be upgrading between now and then. You might even get one with a GPU. If you want new though, you can just about get a $100 gpu and still have enough for a Ryzen 5 (2nd gen), 16gb ram and a mobo if you are careful. I'd try and get a single stick of 16gb. leaving you room for a second stick (4 sticks are more problematic) and make sure the ram is 3000 speed or better. If you live near a Microcenter, check them now and watch later as they have good cpu prices and offer $40 off as well if you buy a matching mobo. Amazon and Newegg can't touch them on this, but you have to do it in store. If you are willing to pass on the GPU you can get something even better.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: tp4tissue on Sun, 01 December 2019, 15:27:36
IDk bout 580s,  they're not easy to buy, only a few models are safe buys after mining crazy.

Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Leslieann on Sun, 01 December 2019, 15:40:45
Dk bout 580s,  they're not easy to buy, only a few models are safe buys after mining crazy.
On a 1080 where you are spending $400 or something, sure, but when you can get a 580 for $80-100, meh.

Replacement fans are cheap and easy to get and when was the last time you saw a gpu die from capacitor failure? If you get a year out of it you got more than your moneys worth.
If you are still concerned get a 4gb card, those are less likely to have been used for mining.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: tp4tissue on Sun, 01 December 2019, 15:59:37
I dn' think NT would get a 580, because 970 is almost the same as 580.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: noisyturtle on Sun, 01 December 2019, 16:39:31
Decided just to just do the jump from 970 to 1070, that will basically double my in-game performance. CPU and other stuff I'd maybe get another 5-10% out of desktop performance, but it's not really worth the price. But a used 1070ti is like $300, that will give plenty of time to save for a new computer next year.

I do like the idea of years from now making a "best parts from 2016" build though. I used to have a "best parts from '95" comp that was really fun to time machine in.

Suppose the time to upgrade comes down to, Can you afford to upgrade? Does your current setup meet minimum requirements for modern titles? What is the most impactful part you could upgrade, and how long would that carry you?
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: fohat.digs on Sun, 01 December 2019, 16:59:05
Since I am lucky to have a MicroCenter in my city, I buy motherboards and CPUs together and get a good bundle price. That was usually my most expensive increment because I sometimes did it to accommodate a Windows leap, but for now I haven't upgraded my mobo/CPU since moving to 10. When that time comes, I suppose that I will have to buy Windows again (groan).

But for a non-gamer, my quad-core processor and 8 GB of RAM seems like it should carry me for another couple of years if I needed it to. And, depending on my work situation, I would relish the opportunity to ditch MicroSoft and go Linux FOSS to eke another couple of years out of a weak system.




 
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: _rubik on Sun, 01 December 2019, 18:07:37
I'm in the same boat at you OP (maybe not 2500k old, though). Running an i5-4670k and GTX 970 that hasn't been upgraded in ~5 years. It can handle things for the most part, but struggles a bit on newer titles. I've always upgraded my mobo and cpu in tandem, but I'm so far behind the current chipsets that I may have to do a full overhaul in the near/distant future.

For now,  I think I'm just going to pick up a 2070 and call it a day. Most of what I do on that machine anyway is gaming and a wee bit of rendering, so nothing too CPU intensive.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: tp4tissue on Sun, 01 December 2019, 18:24:08

For now,  I think I'm just going to pick up a 2070 and call it a day. Most of what I do on that machine anyway is gaming and a wee bit of rendering, so nothing too CPU intensive.

u need 2070 SUPER, if you gonna get 2070,



Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: tp4tissue on Sun, 01 December 2019, 18:27:39
Decided just to just do the jump from 970 to 1070, that will basically double my in-game performance. CPU and other stuff I'd maybe get another 5-10% out of desktop performance, but it's not really worth the price. But a used 1070ti is like $300, that will give plenty of time to save for a new computer next year.

I do like the idea of years from now making a "best parts from 2016" build though. I used to have a "best parts from '95" comp that was really fun to time machine in.

Suppose the time to upgrade comes down to, Can you afford to upgrade? Does your current setup meet minimum requirements for modern titles? What is the most impactful part you could upgrade, and how long would that carry you?


Get 2060 Super if ur going green..

Can't mention enuff, 5700xt = best buy for this bracket.

The only down side is , some driver issues at the moment. u know black screen crashes, that sort of thing.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: tp4tissue on Sun, 01 December 2019, 19:13:00
Hrrrrm.. some chatter about Nvidia-Killer ... hrrrrmm.. but rumor says end of 2020.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: noisyturtle on Sun, 01 December 2019, 19:55:06
the 5700 xt is nice, but it's also $200 more than a used 1070 ti and too new to find used
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: tp4tissue on Sun, 01 December 2019, 20:52:11
the 5700 xt is nice, but it's also $200 more than a used 1070 ti and too new to find used

Not stopping u dawg.. U good with 1070ti..


Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Leslieann on Sun, 01 December 2019, 21:57:56
Nvidia is better on Windows
AMD is better on Linux.

They work fine for the most part though.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: absyrd on Mon, 02 December 2019, 04:49:51
the 5700 xt is nice, but it's also $200 more than a used 1070 ti and too new to find used

1070 ti will rip. Go for it.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: cicada on Mon, 02 December 2019, 04:57:23
The generule of thumb is to get the Sapphire version if you go AMD, I don't think it really matters with Nvidia cards. AMD is crazy good this gen though.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: cicada on Mon, 02 December 2019, 05:01:48
Dk bout 580s,  they're not easy to buy, only a few models are safe buys after mining crazy.
On a 1080 where you are spending $400 or something, sure, but when you can get a 580 for $80-100, meh.

Replacement fans are cheap and easy to get and when was the last time you saw a gpu die from capacitor failure? If you get a year out of it you got more than your moneys worth.
If you are still concerned get a 4gb card, those are less likely to have been used for mining.

I recently got a 580 for my other rig and performance is stellar... maxing most games at 1080p60fps for $80 is almost too good to be true.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Sintpinty on Mon, 02 December 2019, 07:14:52
Since I am lucky to have a MicroCenter in my city, I buy motherboards and CPUs together and get a good bundle price. That was usually my most expensive increment because I sometimes did it to accommodate a Windows leap, but for now I haven't upgraded my mobo/CPU since moving to 10. When that time comes, I suppose that I will have to buy Windows again (groan).

But for a non-gamer, my quad-core processor and 8 GB of RAM seems like it should carry me for another couple of years if I needed it to. And, depending on my work situation, I would relish the opportunity to ditch MicroSoft and go Linux FOSS to eke another couple of years out of a weak system.

A Microcentre? the only thing that i have is Canada Computers which does have components but with expensive prices.

I hate the CAD. USD better
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Sintpinty on Mon, 02 December 2019, 07:15:23
I'm in the same boat at you OP (maybe not 2500k old, though). Running an i5-4670k and GTX 970 that hasn't been upgraded in ~5 years. It can handle things for the most part, but struggles a bit on newer titles. I've always upgraded my mobo and cpu in tandem, but I'm so far behind the current chipsets that I may have to do a full overhaul in the near/distant future.

For now,  I think I'm just going to pick up a 2070 and call it a day. Most of what I do on that machine anyway is gaming and a wee bit of rendering, so nothing too CPU intensive.


Upgrade the CPU! That old I5 will bottleneck it
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Sintpinty on Mon, 02 December 2019, 07:16:31
Hrrrrm.. some chatter about Nvidia-Killer ... hrrrrmm.. but rumor says end of 2020.

Nvidia is dead?

My GTX 1060 might need an upgrade. I am considering giving my laptop to my sister and saving up the money for over 1000 dollars, since i have 700 from paychecks
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Leslieann on Mon, 02 December 2019, 18:38:22
I'm in the same boat at you OP (maybe not 2500k old, though). Running an i5-4670k and GTX 970 that hasn't been upgraded in ~5 years. It can handle things for the most part, but struggles a bit on newer titles. I've always upgraded my mobo and cpu in tandem, but I'm so far behind the current chipsets that I may have to do a full overhaul in the near/distant future.

For now,  I think I'm just going to pick up a 2070 and call it a day. Most of what I do on that machine anyway is gaming and a wee bit of rendering, so nothing too CPU intensive.

Upgrade the CPU! That old I5 will bottleneck it
Bottleneck smottleneck.
If you are hitting enough fps on an old cpu on the games you play it doesn't matter if the cpu is bottlenecking until you can upgrade it. Most people are perfectly happy so long as the system averages around 60fps without dipping below 30 on the games they play. In most cases a 4th gen I5 will do that on most games, I would imagine so would a 3rd or even 2nd gen I7 would as well.
You don't need an e-sports competition grade system to have fun.

That said, beware buying anything too high end too far in advance, not just because of bottlenecks but because you end up over paying and getting less. It's often better to buy two mid level cards spread across time than a single high end. By that I mean if you paid $1000 for a 1080 2 years ago and it was heavily bottlenecked you could have instead bought a 1060 for $200 at the time and gotten similar performance and then later bought a 2060 Super for $400 and gotten better performance then the 1080. Not only was the price spread out over 2 years instead of a single purchase, the 2060 is more modern and has features and optimizations the 1080 doesn't. You would have saved $400 without any performance loss and in fact gained performance later. This doesn't even account for the fact that you would be able to use a lower watt power supply and paid less on your electrical bill saving even more money.

I'm not saying the 2070 is too far, it's not badly priced, just that it will be wasted if you wait 2-3 years to upgrade the rest of the system, in which case a 1660 now/3060 later would be a smarter purchase.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: tp4tissue on Mon, 02 December 2019, 23:36:32
I have a 5ghz 2500k myself, Even @ that clock, it DOES bottleneck esports titles like CSGO/ Dota/ League of Legends, even Doom (2016).

It's not a problem for games like tombraider.

Starcraft 1 runs perfect,  Starcraft 2 not nearly as gud' as newer gen.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: noisyturtle on Mon, 02 December 2019, 23:53:47
I have a 5ghz 2500k

 :eek: that's some nice overclocking
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: absyrd on Tue, 03 December 2019, 04:45:10
Amazon had mid-range x570 mobo, 3700x, and 500gb evo plus nvme for $500 for a few hours yesterday. Damn good deal. Wish I had held out for the holidays.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: PadawanGeek on Thu, 05 December 2019, 18:41:56
My upgrade path is pretty weird, I tend to make a grand gesture by upgrading to the best I can afford, then making GPU upgrdaes along the way to keep the system relevant for gaming. Back in late 2011/early 2012, I built a pretty solid rig:
i7 3960X @4.2Ghz
Asus Rampage IV Extreme
4x 4GB Corsair Dom Plat 2133Mhz
2x HD7970 (followed byt 2x R9 290X, followed by 2x RX VEGA64....no reduced to 1x RX VEGA64 due to lack of CF support in games)
Seasonic X-1250

This rig was my main gaming rig, but is now my 2nd rig (having 6C/12T really helped with its longevity), it's hooked up to my Samsung 49" Freesync 2 monitor and it's doing a damn nice job handling games even now.

Now, this year has been a Ryzen year, and I couldn't help but feel it was time for a new build, again, I went with what was available back in June, so I jumped right in. I was in Toronto at the time, so I had to buy parts piecemeal, but finished my build when I got back home in August, this is my main rig now:

AMD Ryzen 9 3900X
Gigabyte X570 Aorus Xtreme (AGESA 1.0.0.4)
4x 8GB DDR4 XPG D60G 3200CL16 (have a set of 4x 8GB Patriot DDR4 3733CL17 RAM incoming, my niece is gonna bring it to me when she visits me here)
Samsung 4TB 860 EVO + 2TB 860 QVO + 2x 1TB 850 EVO + WD 6TB Black HDD
Corsair HX1000 Plat
PowerColor RX VEGA64 Red Devil (Collected this RMA'ed card while I was in Toronto, still a useful card for 3440x1440 gaming, awaiting NAVI23 or Ampere)

As you can see, I'm more or less in line with the belief that getting the most cores available at the time and upgrading along the way helps ensure the longevity and usefulness of a DIY system. I'd expect my R9 3900X to last at least 3 years or more. Gonna skip AMD/Intel for the next few years while I try to eke out as much performance outta my main rig...
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: romevi on Thu, 05 December 2019, 18:44:40
I think once you have enough money where it's no object, you should start buying new builds every year. Ensure you are flaunting your annual lunch-money purchases by posting pictures on the internet.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Leslieann on Thu, 05 December 2019, 20:03:24
My upgrade path is pretty weird, I tend to make a grand gesture by upgrading to the best I can afford,
That's much more common than you think especially now that CPU gains have slowed considerably.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: suicidal_orange on Fri, 06 December 2019, 02:13:44
My upgrade path is pretty weird, I tend to make a grand gesture by upgrading to the best I can afford,
That's much more common than you think especially now that CPU gains have slowed considerably.
It's what I've always done, though seems Padawan spent more while I bought cheaper low-high end (Pentium D 805 @3.9ghz from 2.66 stock when the best available was 4.0 and heat limited not much higher, Core 2 quad/Xeon ES which was similarly very slow and OC'd, i7 2600k...) then I had money so bought (and killed somehow) an i7 5775c followed by the only time I've put a second CPU in a mobo - my current i7 4790HQ.

GPU wise I similarly bought fun stuff, Radeon 9500 unlocked to 9700, geforce 6800 unlocked to nearly ultra then converted to Quadro.  Finished uni and pretty much stopped gaming so picked up a geforce 6600GT which was fanless and had the GPU on top of the card.  That didn't work following a watercooling leak... next fanless 7300GT because it was cheap, 5775c iGPU and now running a fanless 1050ti bought to diagnose the 5775c.

When is it time to 'upgrade'?  These days when something breaks I complain lots and replace it, thankfully it's usually been so long the only thing worth keeping is the case and PSU.  I hope to skip DDR4 completely.  *Did I just ask for system failure on Christmas Eve?!*
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: yui on Fri, 06 December 2019, 02:43:49
to answer the original question with my point of view, i do incremental upgrading as long as i do not have to completely strip the PC down to upgrade the part (like if i change motherboard or case), and i usually lag behind a few gens and pick old server stuff (12 cores for 10 euros and 64GB of ram for 80euros would could say no to that?) and upgrade usually happen when the new cheap server stuff really outpace what i have (really looking forward to cheap epyc in a few years time if it is like opterons we will have 128 cores for the price or 2 to 3 beers) or when i am bored because the pc runs too good and start building something stupid and outdated.

and sorry tp4 but epyc > threadripper (it is in the name how can you beat something so epic)
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: tp4tissue on Fri, 06 December 2019, 06:57:16
and sorry tp4 but epyc > threadripper (it is in the name how can you beat something so epic)

It's difficult to appropriate epyc for home use. I can't think of any home scenarios where an epyc would be more cost effective vs running on the cloud.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: 4sStylZ on Fri, 06 December 2019, 08:09:20
I did full refresh and I try to keep it work for many years. Usually 5 years or more.
The only thing that I keep is my good high tower (Silent base 800 Be Quiet) because I don’t think that case can be outdated so fast.
The only thing that I buy is disk space sometimes.

I have to admit that I have changed my graphical card because of my new screen but that’s an exception.

I prefer to have a good config for a couple of years or more and deal with all the problem at the same times. When I was incrementally upgrading I was hesitating about what to buy, what to upgrade and sometimes I wasn’t correcting my issues because of the hesitation of investing too much cash on a PC that I will probably leave soon.

When you got many issue for months, you can make a coherent build at 1 time and I find that more easy.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Sintpinty on Fri, 06 December 2019, 08:15:22
Threadripper is 1 wurd..   

If you don't have threadripper, then you need it,  that's how the decision is made.

Show Image
(https://i.imgur.com/fupGm31.gif)


GPU, in the mean time, 5700xt is best buy,  1660ti is the best buy from Nvidia.

not huge on GPU upgrad, because Tp4 mainly watches movies.


Why do you have a soccer GIF
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Sintpinty on Fri, 06 December 2019, 08:16:01
I think once you have enough money where it's no object, you should start buying new builds every year. Ensure you are flaunting your annual lunch-money purchases by posting pictures on the internet.

What's having money?
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: suicidal_orange on Fri, 06 December 2019, 09:44:04
I think once you have enough money where it's no object, you should start buying new builds every year. Ensure you are flaunting your annual lunch-money purchases by posting pictures on the internet.

What's having money?

It's what happens between getting a full time job and getting a mortgage, then you have many years without until you've paid off the house then you can buy a sports car.  Not sure what the female equivalent last step is :))
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Sintpinty on Fri, 06 December 2019, 11:12:54
I think once you have enough money where it's no object, you should start buying new builds every year. Ensure you are flaunting your annual lunch-money purchases by posting pictures on the internet.

What's having money?

It's what happens between getting a full time job and getting a mortgage, then you have many years without until you've paid off the house then you can buy a sports car.  Not sure what the female equivalent last step is :))

D:
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: romevi on Fri, 06 December 2019, 14:13:08
I think once you have enough money where it's no object, you should start buying new builds every year. Ensure you are flaunting your annual lunch-money purchases by posting pictures on the internet.

What's having money?

It's what happens between getting a full time job and getting a mortgage, then you have many years without until you've paid off the house then you can buy a sports car.  Not sure what the female equivalent last step is :))

D:

:D
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Leslieann on Fri, 06 December 2019, 17:41:04
It's what I've always done, though seems Padawan spent more while I bought cheaper low-high end (Pentium D 805 @3.9ghz from 2.66 stock when the best available was 4.0 and heat limited not much higher,
I know that chip all too well.
I broke 4ghz on mine but I was still a bit shy of the records, destroying a few parts in the process.

I.M.O. this was peak overclock, we had the cpu overhead and cooling to take advantage of it. While we have even better cooling today nothing has that much overclocking headroom.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: suicidal_orange on Fri, 06 December 2019, 18:11:51
It's what I've always done, though seems Padawan spent more while I bought cheaper low-high end (Pentium D 805 @3.9ghz from 2.66 stock when the best available was 4.0 and heat limited not much higher,
I know that chip all too well.
I broke 4ghz on mine but I was still a bit shy of the records, destroying a few parts in the process.

I.M.O. this was peak overclock, we had the cpu overhead and cooling to take advantage of it. While we have even better cooling today nothing has that much overclocking headroom.

Nice!  I was but a skint student so couldn't risk going further for fear of killing it and having to starve.  Or was it fear of keeping my housemates awake with the vantek tornado fans trying to keep it cool (92x38mm and far from slow - one took a chunk out my finger and kept on spinning :eek:)
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: NewbieOneKenobi on Wed, 11 December 2019, 04:28:56
Well, you want to upgrade the platform when:

a/ It's cheap to do so.
b/ You have to.
c/ You have spare cash.

You want to max out on the existing platform if:

a/ The top of what it can handle is attractively priced. This mostly means maxing out on the CPU for old mobos, especially if you have good cooling for overclocking.
b/ The bang for buck in upgrading is clearly better than in replacing the platform. OR marginally inferior while allowing you to save a lot of cash by not replacing something you don't need to replace.
c/ It's just one or two components, so you aren't buying a lot of old tech.

In your situation I would not want to buy a 3770K due to:

- the disadvantageous price-to-performance ratio of the top i7's of old generations
- too little performance upgrade from 2500K

I would probably want to replace the GPU right away given a good bargain — not above 1070ti performance level (not necessarily nVidia) unless very attractively priced, but not necessarily below it, either, to save yourself the hassle of small increments, however theoretically perfect on paper.

A 1070ti-level GPU should be decent enough to last you a bit on your new platform when you get around to getting it.





Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Prelim on Wed, 11 December 2019, 09:47:55
same boat here, running a 2500K@4.6Ghz + 16Gb 1866Mhz + GTX 970 (it does the job since I only play at 1080p)

Thinking about upgrading just the GPU for a GTX 1660 this Xmas, and then upgrading the barebone (mobo+cpu+ram) to a 9600K or even i5-10500K next year!
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: absyrd on Wed, 11 December 2019, 15:51:06
1660 ti is a beast, especially considering its power draw.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Leslieann on Wed, 11 December 2019, 16:16:46
Inventory of 1660 cards of all variants will be limited until after New Years.
This was why there was no good GPU sales, they didn't need them, any expected GPU sales were cancelled before we even reached Black Friday and most inventory was gone before then.

It's not just us, world inventory of 1660 and 2060 is quite low (we wiped out Canada), with mostly single fan cards remaining. If you find one for a good price don't wait or you will probably miss it. Word is that Newegg or EVGA is expecting a shipment around the 15th or 16th but it won't last (keep checking both). Also beware inflated prices which may not come back down for a month or more for this reason, I was seeing insane prices for what was still available and even those sold. It's not as bad as the mining craze but it's still an issue as people were really expecting deals and everyone missed the boat for how much demand there would be.

That said, if you can find a good deal on a used 1070, preferably never used for mining, grab it, but don't pay more than $200, even for a prime example. I recommend the EVGA XC series which has better cooling than most. Basically you will get the same performance as a 1660 TI for about 2/3rds the price. This is what I ended up doing after missing out on a good 1660 Super. Mine is super low hours (it sat on a shelf most of it's life) and the added VRAM is better considering I have dual 1440p screens. Don't be afraid to beat up on 1070 owners by comparing price/performance to the 1660, if they want to sell they are going to have to come to grips with reality.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: noisyturtle on Wed, 11 December 2019, 16:25:39
The 1070 ti still blows the 1660 ti out of the water in terms of performance. At minimum +20-5% across the board in stress and rendering tests.

The price point is very attractive though. What a weird stop-gap card the 1660 series is  :confused:
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Leslieann on Wed, 11 December 2019, 18:25:49
The 1070 ti still blows the 1660 ti out of the water in terms of performance. At minimum +20-5% across the board in stress and rendering tests.

The price point is very attractive though. What a weird stop-gap card the 1660 series is  :confused:
TI for TI, yes but not if you do 1070 standard vs 1660 TI which is what most will compare due to similar price.
1660 isn't a stopgap, the only 10 series possibly still made is the 1060 and those under it. 1070 and above have long been discontinued and most are out of the supply chain as well.


Best bang for the buck:
$100 or less look for a used 580
Yes, these were often mining cards, but they are extremely cheap.
$130 new 570
$160 new 580
$200 look for used 1070
Mining cards can be had for less, haggle, use the 1660 as an excuse to lowball. These were used for mining but not the extent faster cards were as they didn't have as a good of a rate of return. Most probably lost money mining on these due to market changes. If no warranty for this much scares you when spending this much consider a 1660 Super, it's only a teeny bit slower.  I bought mine locally and saw and heard it run before I bought it for $200, had it been mined or a lower class model I would have paid less but I think I did okay.
$240 buy a 1660 Super or used 1070 TI
If you need low wattage get a 1660 Super, skip the standard 1660 ($10 difference!). The 1070 Ti is a better choice (more than 10% faster) but only if you can find a good one and that could be difficult, these were heavily used for mining when 1080's became scarce and there's no warranty. As always it comes down to your comfort buying used for this price, personally, if there's no heat issues I'd look for the perfect 1070 TI but continue saving for the 2060 (8% faster) and buy whichever happened first.
$300 buy a 2060
Unless you need low wattage for a small from factor, in which case get a 1660 TI, there's only about $10 difference in price.  You can find used GTX 1080s for this price range and they perform similar, however most are former mining cards and at this price, without a warranty, just say no. Gamers with good, low use cards are asking around $400. There isn't even an argument to be made for the extra ram for high res/refresh as it doesn't help here.
$400 buy a 2060 Super
Gamers are offering good used 1080's at this rate but the Super smokes the 1080 and comes with a warranty. Once again even the larger memory can't save it at this point, the Super is just that good.

1080's are simply a stupid investment at the moment as you can buy something off the shelf with a warranty that performs as good or better for the same amount. The same applies to new and used 1060's which are in a horrible predicament.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: absyrd on Thu, 12 December 2019, 04:16:20
I'm still scared of used GPUs even when people claim "light use". I just always assume they were mining.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Leslieann on Thu, 12 December 2019, 07:00:49
I'm still scared of used GPUs even when people claim "light use". I just always assume they were mining.
So assume it was mined and price accordingly, if it wasn't you got an even better deal.

Modern GPUs rarely fail and should last years even under hard load, it's almost all solid state electronics and packed with thermal and voltage protection, if it worked yesterday it's probably going to work tomorrow, the one exception really is fans and those are cheap. There is one one trick I do for GPUs and that is to look at a sellers other listings and feedback, if they are clearing out lots of high end cards they were mining, if it's just the one, odds are they were just gaming with it and if they have a wide spread they're a recycler and it's difficult to say but they are the least likely to hassle you over a problem. It's not foolproof, but it does give you an edge. As for problems Ebay and Paypal almost always side with a buyer, so much so that bigger sellers don't even argue most of the time that there is a problem because they know they will lose.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Sintpinty on Thu, 12 December 2019, 09:38:58
I'm still scared of used GPUs even when people claim "light use". I just always assume they were mining.
So assume it was mined and price accordingly, if it wasn't you got an even better deal.

Modern GPUs rarely fail and should last years even under hard load, it's almost all solid state electronics and packed with thermal and voltage protection, if it worked yesterday it's probably going to work tomorrow, the one exception really is fans and those are cheap. There is one one trick I do for GPUs and that is to look at a sellers other listings and feedback, if they are clearing out lots of high end cards they were mining, if it's just the one, odds are they were just gaming with it and if they have a wide spread they're a recycler and it's difficult to say but they are the least likely to hassle you over a problem. It's not foolproof, but it does give you an edge. As for problems Ebay and Paypal almost always side with a buyer, so much so that bigger sellers don't even argue most of the time that there is a problem because they know they will lose.

You should see space invaders
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: NewbieOneKenobi on Thu, 12 December 2019, 11:01:29
Well, I think I may be joining the upgraders' ranks this Christmas. A z370 mobo + i5-9600K can be had for less than a 7700K (the equivalent of USD 300). Make it i5-9400 + z730 mobo, and it's going to cost the same as a 7600K (the equivalent of USD 200). So it makes little sense for me to upgrade within my z710 mobo's range, while I'm afraid the good prices on z730 + gen 9 could not last forever.

However, anything I'm going to play will probably do fine in 1440p with the 6600 and 1070ti.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Leslieann on Thu, 12 December 2019, 14:44:20
Well, I think I may be joining the upgraders' ranks this Christmas. A z370 mobo + i5-9600K can be had for less than a 7700K (the equivalent of USD 300). Make it i5-9400 + z730 mobo, and it's going to cost the same as a 7600K (the equivalent of USD 200). So it makes little sense for me to upgrade within my z710 mobo's range, while I'm afraid the good prices on z730 + gen 9 could not last forever.

However, anything I'm going to play will probably do fine in 1440p with the 6600 and 1070ti.
If you can, wait until after 4th quarter profit results are posted unless you see a really good sale price.

Based on the numbers I've seen thrown around almost no one is building low-mid tier Intel and odds are they'll be doing price rollbacks after quarterly results come out. Only trouble is I'm not sure if they will do it for this quarter or next (they may wait to see a pattern, though it's pretty clear). Enthusiasts have all but shunned Intel mid and low end. Microcenter employees have said they estimated AMD selling 9 to 1 with only the Intel 9900k and up selling and Gamers Nexus reported the same for their Amazon Affiliate account but this time with verifiable numbers.

It's a good choice if the price is right.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Larken on Thu, 12 December 2019, 20:53:57
There is one one trick I do for GPUs and that is to look at a sellers other listings and feedback, if they are clearing out lots of high end cards they were mining, if it's just the one, odds are they were just gaming with it and if they have a wide spread they're a recycler and it's difficult to say but they are the least likely to hassle you over a problem. It's not foolproof, but it does give you an edge.

Bought an rx 580 earlier this year, did the same thing. Seller was selling the one rx 580, still had warranty on it. Went down to collect it, found out on the spot that:

1. It was a former mining card.
2. Receipt for warranty was a photocopied one for 22 rx580s and the seller overstated the warranty period.
3. Turns out the seller bought the rx 580 from a miner previously, so it was a third hand card.

In the end I still got it (between the sunken travel time, the price was still borderline acceptable considering everything. But it definitely made it less of a great deal). I definitely wasn't happy about the lack of transparency on the seller's part. If it wasn't a Sapphire Nitro+ 8GB model, I would've passed on it.

Was not foolproof indeed.

Got home, stress tested the card to make sure it works, took it apart to clean out the dust+grime and applied new thermal paste. Slight corrosion on the heat pipes, which couldn't be helped.

On the bright side, the card has worked great over the last year. Got me back into gaming after a long hiatus due to obsolete hardware.

I'm looking to upgrade to a Ryzen 3600 or 3700x soon-ish. Pretty excited, coming from an FX-8320 that's served me well over the last decade.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: noisyturtle on Thu, 12 December 2019, 21:03:10

AMD selling 9 to 1 with only the Intel 9900k and up selling and Gamers Nexus reported the same for their Amazon Affiliate account but this time with verifiable numbers.


that doesn't sound right, one out of every 10 processors sold are IBM?
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: NewbieOneKenobi on Fri, 13 December 2019, 00:25:08
Well, I think I may be joining the upgraders' ranks this Christmas. A z370 mobo + i5-9600K can be had for less than a 7700K (the equivalent of USD 300). Make it i5-9400 + z730 mobo, and it's going to cost the same as a 7600K (the equivalent of USD 200). So it makes little sense for me to upgrade within my z710 mobo's range, while I'm afraid the good prices on z730 + gen 9 could not last forever.

However, anything I'm going to play will probably do fine in 1440p with the 6600 and 1070ti.
If you can, wait until after 4th quarter profit results are posted unless you see a really good sale price.

Based on the numbers I've seen thrown around almost no one is building low-mid tier Intel and odds are they'll be doing price rollbacks after quarterly results come out. Only trouble is I'm not sure if they will do it for this quarter or next (they may wait to see a pattern, though it's pretty clear). Enthusiasts have all but shunned Intel mid and low end. Microcenter employees have said they estimated AMD selling 9 to 1 with only the Intel 9900k and up selling and Gamers Nexus reported the same for their Amazon Affiliate account but this time with verifiable numbers.

It's a good choice if the price is right.

Thank you. That also gets me some time to look at used 6–7th gens (for a quick BuyNow to later put my locked 6600 through a long bidding auction to recover most of the spend). No time to play games or do any overclocking before Christmas and probably all the way to New Year and slightly beyond.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Leslieann on Fri, 13 December 2019, 01:23:30
Was not foolproof indeed.
Got home, stress tested the card to make sure it works, took it apart to clean out the dust+grime and applied new thermal paste. Slight corrosion on the heat pipes, which couldn't be helped.
On the bright side, the card has worked great over the last year. Got me back into gaming after a long hiatus due to obsolete hardware.
I'm looking to upgrade to a Ryzen 3600 or 3700x soon-ish. Pretty excited, coming from an FX-8320 that's served me well over the last decade.
That's a bummer, on the other hand, like you said, it still functioned.

The FX line wasn't as bad as people made them out to be but that's going to be a heck of an upgrade, be sure to get an NVME drive at the same time.

Thank you. That also gets me some time to look at used 6–7th gens (for a quick BuyNow to later put my locked 6600 through a long bidding auction to recover most of the spend). No time to play games or do any overclocking before Christmas and probably all the way to New Year and slightly beyond.
You're welcome, it's a bit of a gamble, but worst case I'd call it a draw waiting, best case, you get a nice price drop.

Keep your eye on Ryzen 2000 deals as well (Microcenter had that and mobo for well under $200) and if you are willing to consider used, there will be a LOT of used Intel and AMD stuff going up after Christmas with everyone getting new stuff. Used cpu and ram is usually pretty safe to buy, it's motherboards that tend to die first, however that's more of an age issue. So long as it's new-ish you should be fine, especially from an individual and not a recycler. Recyclers are actually the worst for boards as they rarely test or include I/O plates and who knows what it was used for, whereas an indivdual who built their own I5 or I7 probably spent the money on a decent PSU to protect it.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: NewbieOneKenobi on Sat, 14 December 2019, 09:43:14
You're welcome, it's a bit of a gamble, but worst case I'd call it a draw waiting, best case, you get a nice price drop.

Yeah. And I really need to work till Christmas Eve and probably beyond, so I'll wait for end-of-year invoices to come in, do the balance sheet, see how much I have to play with. Meanwhile I guess generation 10 could pop up?

Quote
Keep your eye on Ryzen 2000 deals as well (Microcenter had that and mobo for well under $200)

Nice Z370 boards can be had for under $100 where I live right now. No bells and whistles but solid gaming brands with at least one M.2 slot, SLI capacity and everything else I need (though I'd rather have two M.2 slots).

Quote
and if you are willing to consider used, there will be a LOT of used Intel and AMD stuff going up after Christmas with everyone getting new stuff.

Hopefully, yes. Just these last days since our last post or two 7700Ks have decreased by ~ one sixth. Hence maxing out on my CPU slot would currently be the best bang for the buck in terms of immediate power gain. If I see them drop further, I'll jump at it.

Quote
Used cpu and ram is usually pretty safe to buy, it's motherboards that tend to die first, however that's more of an age issue. So long as it's new-ish you should be fine, especially from an individual and not a recycler. Recyclers are actually the worst for boards as they rarely test or include I/O plates and who knows what it was used for, whereas an indivdual who built their own I5 or I7 probably spent the money on a decent PSU to protect it.

I have enough RAM, I think, but the locked i5-6600 is a bit too mid-range right now.

I'll see what they all have in store after Christmas, but I'll probably go with a z370 mobo and either i5-9400KF on a very good deal (I've seen some of those — generally costing the same with a mobo that a comparably powerful older i5/i7 costs without) or a 9600K on a good deal. Or Ryzen, but I'd probably look at series 3 only.  I'm in the good position that I only need to replace the mobo + CPU. The RAM is already 3000-ish DDR4, and the SSD is an NVM Samsung PCIE4 drive I got off some laptop seller whose client wanted a capacity upgrade back when those things were still a novelty.

I'm not really going to pay extra just to avoid a mobo swap and be stuck with an older chipset etc. Plus, I guess I'll recover more money if I also sell the mobo. Unlikely go for less than $50, and the CPU probably for $100-ish, so that's a 50% extra, and about the difference in price between 7700K and 9700K.

Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Larken on Sat, 14 December 2019, 11:25:56
That's a bummer, on the other hand, like you said, it still functioned.

The FX line wasn't as bad as people made them out to be but that's going to be a heck of an upgrade, be sure to get an NVME drive at the same time.


Was an early adopter of the FX line. If I was in the US, I'd probably be eligible for compensation in the class action suit against AMD. Given my mileage on it, I shouldn't complain, but objectively it wasn't great even on release and never lived up to what AMD promised.

I would've gone Sandy Bridge (which was king back then) but I bought into the whole AM3+ thing (had a 1055t thuban prior to the FX upgrade). Between lacklustre gains in new chips that followed Sandy Bridge and how processor speeds wasn't much of a factor for most games until recently, I kept pushing off my overhaul. 3rd gen Ryzen is indeed going to be a heck of an upgrade.

Based on what I read, I'm not sold on the NVME drive. Seems like the consensus is that it only really benefits use cases where extremely large files are common (i.e. 4K video editing). Doesn't seem worthwhile to give up 2 sata ports for it (seems to be the case when using a NVME drive for the MSI Tomahawk Max); I'm one of those desktop dinosaurs who still has 6 hard drives in my case and am looking to add 4 more via a LSI 9211 (still trying to figure out if the cooling will work out though)
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Leslieann on Sat, 14 December 2019, 17:06:32
I'll see what they all have in store after Christmas
I always tell people buy the best they can when they're ready, prices are always in flux, especially this time of year and if you keep waiting for the next/better thing you will be in perpetual wait mode because something else is just around the corner.


I'm not sold on the NVME drive. Seems like the consensus is that it only really benefits use cases where extremely large files are common (i.e. 4K video editing). Doesn't seem worthwhile to give up 2 sata ports for it (seems to be the case when using a NVME drive for the MSI Tomahawk Max); I'm one of those desktop dinosaurs who still has 6 hard drives in my case and am looking to add 4 more via a LSI 9211 (still trying to figure out if the cooling will work out though)
You're right about the drive being best for large stuff, but large stuff is where things are heading.
The biggest benefit is room, which is why almost everything is switching to it. If you thought going from 3.5in to ssd saved space in a case, it's even better with NVME as you don't even need the drive cage at that point. Not that matters to you at the moment.

That said, with that many drives, I would give SERIOUS consideration to building yourself a home file server.
This frees up your desktop to be whatever you want (small box), offers a bit of extra security since it's not used for browsing and it can be a different OS entirely, and you can offload quite a lot onto it freeing your desktop. Not only can you offload work, but you can shut down or put your desktop to sleep (which is a power hog) while still accessing files from any other devices (great if you experiment with different OS). It can even save money on your power bill, even if you have to buy some parts to build it and it can be built in anything without regard for noise and looks since you can now move it to another spot or room entirely. This doesn't even cover all the benefits, I would be hard pressed to give mine up.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Larken on Sat, 14 December 2019, 20:56:50
That said, with that many drives, I would give SERIOUS consideration to building yourself a home file server.
This frees up your desktop to be whatever you want (small box), offers a bit of extra security since it's not used for browsing and it can be a different OS entirely, and you can offload quite a lot onto it freeing your desktop. Not only can you offload work, but you can shut down or put your desktop to sleep (which is a power hog) while still accessing files from any other devices (great if you experiment with different OS). It can even save money on your power bill, even if you have to buy some parts to build it and it can be built in anything without regard for noise and looks since you can now move it to another spot or room entirely. This doesn't even cover all the benefits, I would be hard pressed to give mine up.

Yea, I've been thinking about building my own home file server eventually (just so I don't have to spend as much time doing cable management for my desktop each time I tinker with my desktop). Would you have any advice on how to go about it, OS/software wise? I know nothing about it yet, but FreeNAS/UnRaid seems to be the most popular, and I do have enough old parts to throw one together for pretty much nothing.

NVME is an interesting thought, but yea, it's not something I'd benefit from at the moment. Though with prices of NVME drives and 2.5" ssds being what they are right now, going NVME would be the better option if I was buying new and decide to shift the rest of my hard-drives off to a separate server. I'll probably wait a bit to see how the b550 chipsets turns out with regards to PCI-E 4.0 support for NVME drives to decide.

Thanks for the advice! Really enjoyed the discussion.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Leslieann on Sun, 15 December 2019, 02:03:12
Thanks for the advice! Really enjoyed the discussion.
You're welcome.
A home file server can be as simple or crazy as you want, hardware and software-wise.

Ignore the hype.
Unraid has one massive issue, at least in my eyes... Well, two actually. It's not free and it only boots from a USB stick, who thought it was smart to boot a raid off crappy flash drive is beyond me. Also Freenas runs on BSD and it lacks a LOT of drivers. I've been planning to look into Openmediavault, it's one of the founders of Freenas who got annoyed with the problems of BSD and built a new version based on Debian. There is plugins for virtual machines, media center (to run a tv) as well as something similar to unraid.

If you have the parts, stop procrastinating and just start toying with it.
Start simple, Windows, there's far less security and permissions issues and most people understand it more. This does leave you a bit more open to infection and such but a good place to start and understand all you can do with it. Find out what features you need and want and then go from there.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: NewbieOneKenobi on Sun, 15 December 2019, 04:07:01
I'll see what they all have in store after Christmas
I always tell people buy the best they can when they're ready, prices are always in flux, especially this time of year and if you keep waiting for the next/better thing you will be in perpetual wait mode because something else is just around the corner.

Wonderful advice in all areas of life. ;)

Thankfully, here in this country we have a nation-wide auction platform that is very reputable and very advanced — full-blown buyer protection plus a lot of cool functions such as the ability to track searches and get daily notifications when something new pops up in the category. So it doesn't cost me that much time or effort to keep tabs on a bunch of 7–9th-gen CPUs, but yes, I've experienced it very empirically that bargain-hunting is not worth it.

See, I'm a freelancer, like probably a large minority of other posters here. Jobs, or at least offers, inquiries, RFPs and all that jazz come like busses — sometimes there are none for an hour, sometimes six at once. When I get in PC upgrading mode, it interferes with my productivity and prevents me from being able to finish previous projects early and taking on new ones early, thereby making more money in the relevant month — usually obviously by more than the difference that exists between the prices of an i5 and an i7 within the same range or e.g. RTX 2060 vs RTX 2070 or whatever. Hence if it's time for an upgrade, it doesn't really pay to get submerged in the whole seeking-comparing-haggling mindset, it's better to just focus on making the money first, then spending it. I felt this when buying my last computer a couple of years ago, and to some extent also now, that is last week — wasted a bunch of hours scouting keyboards instead of focusing on my work and then buying the best.

It really is the most productive, gainful strategy to focus on the best, provided it doesn't cross the point of diminished returns by a very bad margin.

People tend to focus on not overspending, but that's more psychological distractionst than rationality. People also want the best bang for the buck, bu they fail to realize chasing it is not efficient — it's not the best bang for the buck when it comes to investing their time. For example when your lost wages due to the time invested in saving money via bargains are worth more than the savings (my own case when buying the last computer).

The difficult challenge here is to actually learn, then make the somewhat bold decisive step of nipping the whole analysis in the bud to lock on the top of the line as a target, the task being to get the money for it. Producing more money is often faster, easier, more efficient than making decisions about how to spend the limited resources one already has. This is probably abundance vs scarcity, which would explain why most people don't and won't get it — same reason why most of us aren't millionaires or even successful career people.

From a practical standpoint, overspending on a powerful CPU or GPU can truly be somewhat of a waste, but as long as one isn't lapsing into an episode of legal insanity, one should be fine. It's probably easier to live with 'yeah, well, I overspent a hundred bucks, let me just put some more ultra settings on for the sake of it' (or enjoy some other non-necessary perks) than 'crap, I skimped on a hundred bucks while already spending half a thousand, and now the whole experience is marred by the corners I cut, with my performance tangibly affected by the negligible savings'. It's easier for me, at least, to let go of the wasted bucks than the missing fps.

Plus, if you overspend a little up front you can avoid intermediate upgrades.

However, of course, overspending cuts into your budget for your next rig, so it's still smarter to not buy things that won't be needed, e.g. gaming features if never gaming, or server/workstation/rendering power while mostly just gaming and doing office work.

So I'm trying to apply this to my situation and right now consciously making the decision to just focus on my work and only keep tabs on the prices, stop overthinking. But easier said than done, of course.






Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: noisyturtle on Sun, 15 December 2019, 16:03:19
Well I just got a refurbished 3770k off eBay for $100 ($90 with the 10% off coupon)

The new cpu along with a 1070ti should keep my rig running current applications for at least another year. The gpu and cpu together already nearly double my performance so that's fine until I can save a few grand for a completely new rig. I COULD upgrade to 32gb ram for another lil 2-4% increase, but it's hardly worth dropping another $200 on :))

This has turned out to be a pretty helpful thread, love this community  :thumb:
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Larken on Sun, 15 December 2019, 16:51:12
Ignore the hype.
Unraid has one massive issue.....


Thanks, that was educational. I'll have to do more reading on OMV.

If you have the parts, stop procrastinating and just start toying with it.
Start simple, Windows, there's far less security and permissions issues and most people understand it more.


While I have parts for a separate system lying around, I don't have spare copies of Windows lying around, though I could probably just grab one off ebay. Inertia is present especially since the separate file server is a nice-to-have for me, and not a need-to-have right now. Definitely will take what you said into consideration.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: TacticalCoder on Sun, 15 December 2019, 17:27:28
Based on what I read, I'm not sold on the NVME drive. Seems like the consensus is that it only really benefits use cases where extremely large files are common (i.e. 4K video editing).
The day I read about NVMe M.2 drive is when I knew it was time to upgrade. So four years ago I bought a mobo with an integrated M.2 NVMe slot and a Samsung 950 Pro PCIe 3.0 x4 lanes SSD.

Single biggest boost I ever experienced (I'm a software dev and I don't game at all).

A few months ago, after nearly four years of daily using that first NVMe drive, I upgraded to another NVMe drive: a M.2 PCIe 3.0 x4 lanes 500 GB Samsung 970 Evo Plus, for my main/OS drive. Sweet begeezus that thing is fast.

And these are cheap consumer drives (I think I paid my 970 Evo Plus 80 EUR VAT excluded), not crazy expensive Optane SSDs or anything like that.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Leslieann on Sun, 15 December 2019, 18:35:41
Thanks, that was educational. I'll have to do more reading on OMV.

While I have parts for a separate system lying around, I don't have spare copies of Windows lying around, though I could probably just grab one off ebay. Inertia is present especially since the separate file server is a nice-to-have for me, and not a need-to-have right now. Definitely will take what you said into consideration.
You're welcome.

It didn't start out as a necessity for me either but it quickly became apparent I should have done it sooner, I think you will find the same thing. It's not good having one system doing everything because the system you use is always under constant threat and load. Not the place you want all your data stored.

As for Windows, it is perfectly legal to download and use Win10 isos from Microsoft and use them for 30 days, it's even okay to use them beyond that. It won't offer full customizing but will still work. If you want to feel better, grab a key from Ebay for a few bucks.  Ms will even give you a legit license if you upgrade to it from a real or fake win7 or 8 license.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Leslieann on Sun, 15 December 2019, 19:40:51
The day I read about NVMe M.2 drive is when I knew it was time to upgrade. So four years ago I bought a mobo with an integrated M.2 NVMe slot and a Samsung 950 Pro PCIe 3.0 x4 lanes SSD.
Single biggest boost I ever experienced (I'm a software dev and I don't game at all).
A few months ago, after nearly four years of daily using that first NVMe drive, I upgraded to another NVMe drive: a M.2 PCIe 3.0 x4 lanes 500 GB Samsung 970 Evo Plus, for my main/OS drive. Sweet begeezus that thing is fast.
And these are cheap consumer drives (I think I paid my 970 Evo Plus 80 EUR VAT excluded), not crazy expensive Optane SSDs or anything like that.
NVME is simply a protocol, more precisely the removal of one. Sata requires translating the chipset to sata then from sata to memory controller to memory chip, then everything goes back in reverse. NVME removes the sata layer and sata speed limits but the memory controller and memory chips are often the limiting factor. 90% of NVME drives use the same memory controllers and memory chips as the sata drives do, they get around this limitation by stacking controllers and memory chips, this is why the faster ones cost more, they're basically using multiple ssds in raid.

All this is to say NVME is only faster if built to be faster, most are not. Most are simply the same controller and chips slapped onto a new pcb minus the sata controller, yes, they are actually charging more for less. The R&D was almost nothing, the PCB cost was almost nothing, they simply cost more because they knew people would pay it.


At any rate, most systems aren't capable of really even pushing a 950 much less a 970 except specific load cases, your use is one of them. I also suspect you have a system capable of it as well considering you stay up on top of the fast stuff.

As for Optane, it's benefits are oddly specific and limited, and for the cost, just buy a fast ssd.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: NewbieOneKenobi on Mon, 16 December 2019, 11:47:37
Well I just got a refurbished 3770k off eBay for $100 ($90 with the 10% off coupon)

The new cpu along with a 1070ti should keep my rig running current applications for at least another year. The gpu and cpu together already nearly double my performance so that's fine until I can save a few grand for a completely new rig. I COULD upgrade to 32gb ram for another lil 2-4% increase, but it's hardly worth dropping another $200 on :))

This has turned out to be a pretty helpful thread, love this community  :thumb:

Congrats!

Myself I've decided to stop looking and focus on working. My clients are flooding me this week, so I'm working 16-hour days and contemplating energy drinks + sleepless nights to cope with the load, just not gonna start too early to avoid detrimental effects kicking in too early. I'm probably going to be forced to start declining any new projects till Christmas. Hence I've decided I'm done scouting and researching, just gonna keep pinging 7700K's, 8600–87000Ks and 9600–9700K's for bargains. Basically any hour or two lost scouting/researching would be costing me the same as a one-step CPU increment, so I might as well focus on work and get myself the better CPU right away, or just focus on work, cash in on the invoices and see then if I want to economize to save some money or change my guilty pleasure of choice and buy something else instead.

It's really a tough challenge not looking at reviews, benchmarks, etc., combatting procrastination. One thing to find in benchmarks is the varying behaviour of older i7's with more cores vs newer i5's with better single-core performance. Solution: no 7600K compromise, only 7700K if slotting out on my current mobo, and will keep checking 8700K/9700K prices too or just grab a 9600K, since it has six cores anyway, so can't lag behind an old 4-core 8-thread i7 much.

I had half a mind to getting the 7700K when I signed on three new urgent projects today (sleepless nights coming), but the good deals are gone in the meantime, so whatever — I need to force myself to stop thinking about this to avoid sinking into a procrastination black hole and taking too much time off work. Besides, at 37 it's high time I started to learn earning before spending (sigh… bachelor's financial management…).

Edit: I've started looking more seriously at Ryzen 3600X, but then I've just realized I already have an aftermarket cooler for socket 1151 (adapted from LGA775, har har), and 3600x goes about head to head in both performance and price with 9600K, also same number of cores, so I guess I'm gonna stick with Intel.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: absyrd on Mon, 16 December 2019, 15:54:20
Damnit. 5500 reviews underwhelming.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: tp4tissue on Mon, 16 December 2019, 20:28:21
Damnit. 5500 reviews underwhelming.

U wan dat, 5700xt
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Leslieann on Mon, 16 December 2019, 21:29:30
Damnit. 5500 reviews underwhelming.
Were you expecting a 2080 killer?
It slots precisely where it should against Nvidia based on price.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: NewbieOneKenobi on Tue, 17 December 2019, 00:43:34
Don't know what the prices are where you are, but in my neck of the woods I sometimes see great prices on 1080 and especially 1070ti, so unless you need RTX that can be the way to go, perhaps with some OC, especially for the 1070ti. In some cases blower versions are cheaper by such a large margin that you can buy a great aftermarket cooler within the difference.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: absyrd on Tue, 17 December 2019, 00:59:29
Damnit. 5500 reviews underwhelming.

U wan dat, 5700xt

Yes.

Damnit. 5500 reviews underwhelming.
Were you expecting a 2080 killer?
It slots precisely where it should against Nvidia based on price.

I was expecting a lower price. I was expecting a bit of a jump in performance. I guess the memory is expensive?

Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Larken on Tue, 17 December 2019, 06:00:57
Edit: I've started looking more seriously at Ryzen 3600X, but then I've just realized I already have an aftermarket cooler for socket 1151 (adapted from LGA775, har har), and 3600x goes about head to head in both performance and price with 9600K, also same number of cores, so I guess I'm gonna stick with Intel.

Having been considering the Ryzen for a while now, I'll save you some time and potential grief. If you go for Ryzen, skip the 3600x and go for either 3600 or 3700x. 3600x is a really poor buy, as is the 3800x (if you have workloads that require more cores than what the 3700x can provide, just go for a 3900x or a 3950x)

Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: tp4tissue on Tue, 17 December 2019, 07:48:34
Edit: I've started looking more seriously at Ryzen 3600X, but then I've just realized I already have an aftermarket cooler for socket 1151 (adapted from LGA775, har har), and 3600x goes about head to head in both performance and price with 9600K, also same number of cores, so I guess I'm gonna stick with Intel.

Having been considering the Ryzen for a while now, I'll save you some time and potential grief. If you go for Ryzen, skip the 3600x and go for either 3600 or 3700x. 3600x is a really poor buy, as is the 3800x (if you have workloads that require more cores than what the 3700x can provide, just go for a 3900x or a 3950x)



If just playing games, 3600x is fine.

It's also fine for blender, cads , because it's cheaper and faster to buy cloud time.

You have to have something very specific and local to use more than 8 cores. Like a minecraft server.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: NewbieOneKenobi on Tue, 17 December 2019, 09:35:25
Thanks. Well, what I use is MS Office with sometimes humongous and badly optimized files, Firefox, which is capable of eating through any amount of recourses, with anywhere from two to three gazillion tabs, and above all Trados — a badly optimized CAT suite that's based on XML and involves the real-time use of humongous terminology databases (a.k.a. Multiterm, e.g. the entire IATE database) and translation memories. Those files are sometimes huge, and when you convert like 10 pages' worth of memoQ (another translation suite) into a macro-heavy .docx format, it's enough to make my i5-6600 cry (along with its 16 GB RAM). When you stay within the native software, it's much better (these things somehow still all worked 15 years ago, people still did 500-page files back on Core 2 Duo and before), but it can get laggy, especially with those large terminology databases that get pinged anew every new sentence for all expressions contained in that sentence. This makes gamings rigss cry in my and other translators' experience. But I don't really know if more cores would help or just better single-core performance. I suspect the latter. Otherwise it's just games for me.

The irony of my software is that all of the things I do can still be done on a 2005 rig, it's not like you won't be able to do them (except for the macro-heavy Office files viewed in Office — that's gonna clog any Whateverisbestnowlake(, but they're still going to be a pain almost no matter what sort of ridiculous power you have in your rig.

Games-wise I thought just getting 9600K was the best solution… until I saw benchmarks from games actually optimized to actually use more cores/i7 functions, which sometimes put those 7700s and 8700s very firmly in the lead. Allegedly 3600x is supposed to be almost on par with 8700K in those games.

However, I think I'll spare myself the headache and just stick with Intel, which I'm more familiar with. Generation 10 is expected to come with a new socket, so I hope gen 11 will still it in the same socket, os might as well skip 1151 and wait for that, getting 10600K or 10700K the moment it comes out.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Leslieann on Tue, 17 December 2019, 16:55:49
"Intel is better for games"

This is something Youtubers and others REALLY, REALLY need to stop saying or at least explain it better.
Intel may be faster, but both are so ridiculously far and above what you actually need for a good experience and future proofing that it's completely irrelevant. They only say Intel is better for games because it helps differentiate the two, in the real world the only way to even spot the difference is with a benchmark.


As for a 9600 vs 8700k, in most instances even your current cpu really isn't that much slower, I had a 6700k before my 8700k, there was a speed jump, but most of that was going from a 500MB transfer rate sata ssd to a 3500MB nvme drive. The 6700 was plenty fast, but the best way to explain it is this... While you can get most of the same performance in games as I do, given equal ssd, ram and gpu (we're ignoring the nvme drive for the moment), I get that performance gain while also streaming video and running a virtual machine in the background. The ssd added the speed, the extra cores added the ability to do more at the same time. I'd advocate upgrading, but mostly for the NVME drive and less for the core count.

By the way, Firefox gets laggy at 9gb of ram on windows, doesn't matter how much more ram or cpu you have this will happen, by 15gigs it's pretty much unusable. While I haven't tested on Windows Chrome is probably similar as it gets laggy about the same time as Firefox on Linux. Browsers and how the OS handles them are the problem and no amount of hardware will fix that. I documented some of my browser/memory experimentation here (https://geekhack.org/index.php?topic=103373.0).
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: tp4tissue on Tue, 17 December 2019, 18:56:49
I still think , for a JUST GAME pc, 9900KS is a decent choice,  even though it's a dead socket.

Esports titles respond very well to Intel, and since these are the games most people actually play, intel's still leading.

The main advantage being , there's no guessing game for overclocking, it's all very well worked out, along with compatible OC ram.


For Ryzen, there's some grinding for ram compatibility, and while OC isn't a problem, there's more tweaking involved for very little gain.

For Upgrades, Ryzen is the only way to go. 5ghz is on the horizon, but maybe 2 more iterations away.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Leslieann on Wed, 18 December 2019, 02:31:25
I still think , for a JUST GAME pc, 9900KS is a decent choice,  even though it's a dead socket.
AM4 is also a dead socket.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: NewbieOneKenobi on Wed, 18 December 2019, 07:02:28
"Intel is better for games"

This is something Youtubers and others REALLY, REALLY need to stop saying or at least explain it better.
Intel may be faster, but both are so ridiculously far and above what you actually need for a good experience and future proofing that it's completely irrelevant. They only say Intel is better for games because it helps differentiate the two, in the real world the only way to even spot the difference is with a benchmark.


As for a 9600 vs 8700k, in most instances even your current cpu really isn't that much slower, I had a 6700k before my 8700k, there was a speed jump, but most of that was going from a 500MB transfer rate sata ssd to a 3500MB nvme drive. The 6700 was plenty fast, but the best way to explain it is this... While you can get most of the same performance in games as I do, given equal ssd, ram and gpu (we're ignoring the nvme drive for the moment), I get that performance gain while also streaming video and running a virtual machine in the background. The ssd added the speed, the extra cores added the ability to do more at the same time. I'd advocate upgrading, but mostly for the NVME drive and less for the core count.

By the way, Firefox gets laggy at 9gb of ram on windows, doesn't matter how much more ram or cpu you have this will happen, by 15gigs it's pretty much unusable. While I haven't tested on Windows Chrome is probably similar as it gets laggy about the same time as Firefox on Linux. Browsers and how the OS handles them are the problem and no amount of hardware will fix that. I documented some of my browser/memory experimentation here (https://geekhack.org/index.php?topic=103373.0).

Yeah, I get good performance in games, though not exactly all ultra @ 1440p @ 60 fps. I have a 32'' 1440p monitor, so sometimes I fake 4K by scaling, and things do look vastly better, though fluidity is lacking. In some cases even high fps feels a bit choppy, like the performance, which is technically there, is putting a big strain on the system and making it chafe under the load. Plus, the usual gripe of having a good mobo but a non-K processor (stupid choice on my part some years ago). I question the need for a $300–400 platform upgrade, while upgrading within s1151 feels like paying half that for only a small gain.

I guess I can always finish whatever games I can already top out, first, and only then go for a platform upgrade. Or just swap the CPU if there's a bargain in the meantime.

One thing I've noticed about used 6–8000 CPUs is they tend to cost the same with the mobo as without, so that's also something to consider.

But meanwhile I'm drowned in work anyway. So perhaps I'll just have to rein the upgradeitis in. ;)
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: tp4tissue on Wed, 18 December 2019, 08:13:08
1440p is still the go to game resolution, but then we're stuck with a non-4K monitor for Movies/Shows.

1440p on a 4K monitor looks really awful, because the scaling algorithm blemishes/ tones down the highlight/ texture detail of the games.

This was also a problem with 1080p on 4K, but it's alot better now with Integer scaling available on newer cards.

AMD is next lvl enabling integer all the way down to 7xxx series.

200+ fps is a sweet spot for gaming fluidity. nothing can really hold that on latest titles @ 4K
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: NewbieOneKenobi on Wed, 18 December 2019, 10:13:18
I have a 1440p monitor that can do 75 fps, but my 1070ti on this i5-6600 won't do that unless perhaps in older titles, with lowered settings, and so on. Will see when I get around to finishing some of the older games I have.

But lately I've been playing The Witcher 3, and using virtual 4K resolution significantly improves the experience, except the framerate goes down to like 30 on a mix of high and ultra settings. I suppose I could work out a satisfactory compromise if I had more knowledge about how the various settings work, basically what can be turned off for a noticeable performance gain without a noticeable detriment to image quality.

I suppose Kingdom Come is going to be more demanding.

I have a ton of different things I should or at least more usefully could spend money on, but I guess I've been caught by ugradeitis.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: tp4tissue on Wed, 18 December 2019, 10:54:46
I have a 1440p monitor that can do 75 fps, but my 1070ti on this i5-6600 won't do that unless perhaps in older titles, with lowered settings, and so on. Will see when I get around to finishing some of the older games I have.

But lately I've been playing The Witcher 3, and using virtual 4K resolution significantly improves the experience, except the framerate goes down to like 30 on a mix of high and ultra settings. I suppose I could work out a satisfactory compromise if I had more knowledge about how the various settings work, basically what can be turned off for a noticeable performance gain without a noticeable detriment to image quality.

I suppose Kingdom Come is going to be more demanding.

I have a ton of different things I should or at least more usefully could spend money on, but I guess I've been caught by ugradeitis.


Witcher is not an fps. so, fast doesn't make a huge difference.

But 240hz VAs are out there now.   Just waiting on one with ULMB.

I'd honestly prefer a thicker monitor w/ higher contrast ratio, but they don't seem to be making'um.

Current samsung curved VA contrast goes upwards of 3800:1,  TVs are pushing 6000-7000 now.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: NewbieOneKenobi on Wed, 18 December 2019, 16:31:57
Witcher is not an fps. so, fast doesn't make a huge difference.

No fps for me, though some of the RPGs I play have relatively fast action that's sometimes cast in the first-person perspective. Plus rallies and racers.

Quote
But 240hz VAs are out there now.   Just waiting on one with ULMB.

First time I hear, but sounds like my dream.

Quote
I'd honestly prefer a thicker monitor w/ higher contrast ratio, but they don't seem to be making'um.

Somebody still makes CRT monitors in India, from what I've heard. ;) But those are for industrial applications.

Quote
Current samsung curved VA contrast goes upwards of 3800:1,  TVs are pushing 6000-7000 now.

Delicious. My VA from several years ago tops out around 1000, which is nominally worse than my IPS from last year. I'm not sure I've even looked at anything above 1000 or 1200 yet.

A while ago I bought a 32'' IPS from AOC because I wanted a 1440p monitor and my old Dell 30'' suffered terminal damage. It's crisp enough so I can use 100% scaling in Windows and the 1000-ish contrast does show, though in a different way than the nominally slightly inferior value does on my VA. I'd probably ideally prefer a VA with extended colour range and HDR and yes, as little blur as possible. Next time maybe. And I'll probably go down to 27''. Possibly curved but not sure.

But for something like 200-ish refresh rates I'd need to clone my GPU. Possibly twice.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: yui on Thu, 19 December 2019, 03:35:35
basically what can be turned off for a noticeable performance gain without a noticeable detriment to image quality.

if you use 4k you should be able to disable or strongly reduce anti-aliasing, this thing hogs performances (might be my old hardware and not be as true of newer cards) and depending on if you like them or not chromatic aberrations and motion blur (and other "cinematic" enhancements) could be disabled and depending on their implementation increase performances by a fair margin. the only way to really know for sure is to experiment because different graphics cards, games and driver revisions can have different results when it comes to some effects (like PhysX on Amd cards running on the CPU if at all)
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: NewbieOneKenobi on Thu, 19 December 2019, 04:20:27
basically what can be turned off for a noticeable performance gain without a noticeable detriment to image quality.

if you use 4k you should be able to disable or strongly reduce anti-aliasing, this thing hogs performances (might be my old hardware and not be as true of newer cards) and depending on if you like them or not chromatic aberrations and motion blur (and other "cinematic" enhancements) could be disabled and depending on their implementation increase performances by a fair margin. the only way to really know for sure is to experiment because different graphics cards, games and driver revisions can have different results when it comes to some effects (like PhysX on Amd cards running on the CPU if at all)

Thank you. I guess I should also get around to OC'ing the card already. It's a 1070ti from Colorful that has a beefed-up power section to 8+8 as opposed to the normal 8+6.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: yui on Thu, 19 December 2019, 07:01:21
If i remember correctly those cards can overclock enough to surpass the stock 1080 so yeah might be a good idea if you have sufficient cooling in your case, and are not slowed down by an other part of your build
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: NewbieOneKenobi on Thu, 19 December 2019, 09:40:56
If i remember correctly those cards can overclock enough to surpass the stock 1080 so yeah might be a good idea if you have sufficient cooling in your case, and are not slowed down by an other part of your build

If i remember correctly those cards can overclock enough to surpass the stock 1080 so yeah might be a good idea if you have sufficient cooling in your case, and are not slowed down by an other part of your build

The i5-6600 CPU is a bit of a bottleneck for 1070ti from what I've seen in some online tool… It really sucks I bought a non-K CPU with a strong OC/gaming mobo plus already had a strong CPU cooler. Can't really OC much even through BCLK. Anyway, I should have enough airflow in the case and room for more fans to get a serious OC going — just never get around to it.

At this point I could get a K CPU for this mobo, but given the prices of used 6700/7600/7700K, it makes more sense to swap the mobo as well.

I think I could buy a 6600K for $125 and sell my 6600 for $100-something, for a net $25 expenditure, which from my current standpoint would be very reasonable, but I'd rather go more steps up. The problem? 7600K is like $75 more expensive than 6600K and is actually like $20 more expensive than 8600K. So basically you're supposed to pay a heavy premium for not having to buy a new mobo. The price of an old 6700K is about the same as of a new 9600K.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: tp4tissue on Thu, 19 December 2019, 09:55:50
You need OC for Esports titles which run at 200 to 500fps.  Those are the only applications that could be said to take advantage of higher frequency.

For example Doom 2016, it's locked to 200fps, and a 1070ti will dip below 150fps quite frequently. There's really no reason to pay extra for an OC cpu for a GPU bound game.

Whereas, CSGO, vanilla will get you 200-400fps , overclocked will get you 300-600fps,  That's a huge difference worth paying for.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: noisyturtle on Fri, 20 December 2019, 00:35:14
I just spent the last... you know what, I'm embarrassed to even say how long, but I spent that obscene amount of time attempting to install this new cpu. Got it in there once, only to realize 'You dummy, you forgot to update the bios!'. So I take it back out, clean it off, reinstall my old 2500k, then proceed to tear my skin off trying to flash the new bios version. Now I'd never had to do this on my current rig, so little did I know a "REALLY GREAT FEATURE" called Fast Boot had been automatically enabled. Now I dunno if you're familiar with Fast Boot, but it's this wonderful thing ASUS mobo's have that disables all your usb devices until AFTER Windows has launched. "Wait a sec" you say wisely spying the glaring flaw "then how are you supposed to access the bios when all the usb devices have been disabled?" And by gum, you'd be absolutely right! You freaking can't! And because it's a bios setting you cannot have the sliver of human decency and convenience to be able to change that while in Windows. Oh what fun, oh what joy I had today trying to figure that complete bull**** out! A waste of an entire day, and what's more I wound up being too frustrated and tired after finally getting my bios updated to even attempt reinstalling the 3770k.

I already can predict this will end tomorrow with the DRAM led still lighting up with no post when using the new cpu, and me defenestration my computer right out the third story of the building.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: yui on Fri, 20 December 2019, 02:54:14
I guess that if you unplug your boot drive you should be able to access bios, maybe, or keep an old ps2 keyboard handy.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Leslieann on Fri, 20 December 2019, 03:37:27
I just spent the last... you know what, I'm embarrassed to even say how long, but I spent that obscene amount of time attempting to install this new cpu. Got it in there once, only to realize 'You dummy, you forgot to update the bios!'. So I take it back out, clean it off, reinstall my old 2500k, then proceed to tear my skin off trying to flash the new bios version. Now I'd never had to do this on my current rig, so little did I know a "REALLY GREAT FEATURE" called Fast Boot had been automatically enabled. Now I dunno if you're familiar with Fast Boot, but it's this wonderful thing ASUS mobo's have that disables all your usb devices until AFTER Windows has launched. "Wait a sec" you say wisely spying the glaring flaw "then how are you supposed to access the bios when all the usb devices have been disabled?" And by gum, you'd be absolutely right! You freaking can't! And because it's a bios setting you cannot have the sliver of human decency and convenience to be able to change that while in Windows. Oh what fun, oh what joy I had today trying to figure that complete bull**** out! A waste of an entire day, and what's more I wound up being too frustrated and tired after finally getting my bios updated to even attempt reinstalling the 3770k.

I already can predict this will end tomorrow with the DRAM led still lighting up with no post when using the new cpu, and me defenestration my computer right out the third story of the building.
Been there done that, I have answers.
Hold shift when hitting restart it will give you an option to enter EFI. As mentioned you can also use a PS2 keyboard, pull the drive or my favorite, clear the cmos/reset bios. Newer boards have the ability to fast boot while leaving the keyboard/mouse enabled. Yeah, that was a dumb way to do it and they should have seen it coming.

Good luck.

One thing to beware of on this era of boards where we switched over to EFI is that some GPUs are not fully compatible.
I replaced a very good psu, almost the GPU before reverting to a different mobo I had which worked before finding out it was a gpu/efi/bios compatibility issue (I had updated the biso since the last time I used that board/gpu combo). So if you can't get a gpu signal after, don't panic. It may not be easily fixed to where everything works, but don't freak out thinking something went bad.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: NewbieOneKenobi on Fri, 20 December 2019, 06:42:43
I just spent the last... you know what, I'm embarrassed to even say how long, but I spent that obscene amount of time attempting to install this new cpu. Got it in there once, only to realize 'You dummy, you forgot to update the bios!'. So I take it back out, clean it off, reinstall my old 2500k, then proceed to tear my skin off trying to flash the new bios version. Now I'd never had to do this on my current rig, so little did I know a "REALLY GREAT FEATURE" called Fast Boot had been automatically enabled. Now I dunno if you're familiar with Fast Boot, but it's this wonderful thing ASUS mobo's have that disables all your usb devices until AFTER Windows has launched. "Wait a sec" you say wisely spying the glaring flaw "then how are you supposed to access the bios when all the usb devices have been disabled?" And by gum, you'd be absolutely right! You freaking can't! And because it's a bios setting you cannot have the sliver of human decency and convenience to be able to change that while in Windows. Oh what fun, oh what joy I had today trying to figure that complete bull**** out! A waste of an entire day, and what's more I wound up being too frustrated and tired after finally getting my bios updated to even attempt reinstalling the 3770k.

I already can predict this will end tomorrow with the DRAM led still lighting up with no post when using the new cpu, and me defenestration my computer right out the third story of the building.

Argh. Sorry to hear that, especially just before Christmas. Soldier on, mate. Wish you luck.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Steezus on Fri, 20 December 2019, 07:25:27
I just spent the last... you know what, I'm embarrassed to even say how long, but I spent that obscene amount of time attempting to install this new cpu. Got it in there once, only to realize 'You dummy, you forgot to update the bios!'. So I take it back out, clean it off, reinstall my old 2500k, then proceed to tear my skin off trying to flash the new bios version. Now I'd never had to do this on my current rig, so little did I know a "REALLY GREAT FEATURE" called Fast Boot had been automatically enabled. Now I dunno if you're familiar with Fast Boot, but it's this wonderful thing ASUS mobo's have that disables all your usb devices until AFTER Windows has launched. "Wait a sec" you say wisely spying the glaring flaw "then how are you supposed to access the bios when all the usb devices have been disabled?" And by gum, you'd be absolutely right! You freaking can't! And because it's a bios setting you cannot have the sliver of human decency and convenience to be able to change that while in Windows. Oh what fun, oh what joy I had today trying to figure that complete bull**** out! A waste of an entire day, and what's more I wound up being too frustrated and tired after finally getting my bios updated to even attempt reinstalling the 3770k.

I already can predict this will end tomorrow with the DRAM led still lighting up with no post when using the new cpu, and me defenestration my computer right out the third story of the building.

While you're loaded into Windows you can have the computer reboot straight into BIOs. I keep fast boot on so anytime I need to access the BIOs I just do it through Windows.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: fohat.digs on Fri, 20 December 2019, 07:52:09
This seems to be a good place to pose this inquiry:

Sometime in the coming weeks or months I plan to buy a new motherboard/CPU/RAM (which will also necessitate buying Windows, which I have avoided so far via the free upgrade route to 10 from 7/8) for the first time in about 5 years.

I want something with reasonable power but I am not a gamer, so my real criteria are somewhat future-proofing and dependability, although I have no objection to buying "last year's model" to economize.

There is a MicroCenter in town, so I am very fortunate in that regard with all the specials and package deals that they usually offer. Without getting into specific models or numbers, considering a budget in the range of $200-250 for motherboard and processor + $100-150 for RAM, what architecture should I look at?

On Intel vs AMD I am agnostic and have had good experiences with AMD in the past.

Thanks!

 
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: tp4tissue on Fri, 20 December 2019, 08:37:09
This seems to be a good place to pose this inquiry:

Sometime in the coming weeks or months I plan to buy a new motherboard/CPU/RAM (which will also necessitate buying Windows, which I have avoided so far via the free upgrade route to 10 from 7/8) for the first time in about 5 years.

I want something with reasonable power but I am not a gamer, so my real criteria are somewhat future-proofing and dependability, although I have no objection to buying "last year's model" to economize.

There is a MicroCenter in town, so I am very fortunate in that regard with all the specials and package deals that they usually offer. Without getting into specific models or numbers, considering a budget in the range of $200-250 for motherboard and processor + $100-150 for RAM, what architecture should I look at?

On Intel vs AMD I am agnostic and have had good experiences with AMD in the past.

Thanks!

 


You prolly need Threadrippa to stay relevant , because it's really easy for software developers to (unnecessarily) Use a crap ton of CPU power due to lazy development.

For example, Battle.net downloads,  some asshol accountant on their end figured they could save on bandwidth bill by high-compression.

This high compression leads to 100% CPU utility while downloading their games on the consumer end.  So now the download speed is dependent on how fast the CPU is on the buyer end, using a crap-ton of Electricity, JUST SO Blizzard can work out a slightly improved balance sheet.  That compression they're using is insanely power intensive.

It really should be illegal for them to do this.

But this is the kind of scenario we're going to run into in the future as AMD opens the door for more cores.

This power-creep will happen on everything, not just games.

I recommend at least a 3900x if cost is a concern.

GPU,  Right now, Turing 20xx series on Nvidia has alot of driver issues. Without competition for the top bracket from AMD, Nvidia is still sandbagging on features and performance.  I'd wait this out until AMD brings parity.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: tp4tissue on Fri, 20 December 2019, 08:38:01
I already can predict this will end tomorrow with the DRAM led still lighting up with no post when using the new cpu, and me defenestration my computer right out the third story of the building.

Yea, but this is ONLY THE BEGINNING,

Overclocking is another 10-20 hours of FUNNNNNN....
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: NewbieOneKenobi on Fri, 20 December 2019, 08:56:42
This seems to be a good place to pose this inquiry:

Sometime in the coming weeks or months I plan to buy a new motherboard/CPU/RAM (which will also necessitate buying Windows, which I have avoided so far via the free upgrade route to 10 from 7/8) for the first time in about 5 years.

I want something with reasonable power but I am not a gamer, so my real criteria are somewhat future-proofing and dependability, although I have no objection to buying "last year's model" to economize.

There is a MicroCenter in town, so I am very fortunate in that regard with all the specials and package deals that they usually offer. Without getting into specific models or numbers, considering a budget in the range of $200-250 for motherboard and processor + $100-150 for RAM, what architecture should I look at?

On Intel vs AMD I am agnostic and have had good experiences with AMD in the past.

Thanks!

 

If you aren't a gamer, then AMD seems to be the better choice these days — if you need workstation power. I'm not really familiar with the 1st & 2nd gen Ryzens, but 3500-3900(x) ones seem to pack more workstation performance than Intel in the same price bracket, plus cheaper mobos, in the upper middle end and perhaps lower upper end. When you take it even further, Intel does seem to be cheaper for the same performance with the 9700K, except for extreme workstation use.

If you choose intel, it probably makes the most sense to buy 9th gen and skip older generations.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Steezus on Fri, 20 December 2019, 10:13:47
This seems to be a good place to pose this inquiry:

Sometime in the coming weeks or months I plan to buy a new motherboard/CPU/RAM (which will also necessitate buying Windows, which I have avoided so far via the free upgrade route to 10 from 7/8) for the first time in about 5 years.

I want something with reasonable power but I am not a gamer, so my real criteria are somewhat future-proofing and dependability, although I have no objection to buying "last year's model" to economize.

There is a MicroCenter in town, so I am very fortunate in that regard with all the specials and package deals that they usually offer. Without getting into specific models or numbers, considering a budget in the range of $200-250 for motherboard and processor + $100-150 for RAM, what architecture should I look at?

On Intel vs AMD I am agnostic and have had good experiences with AMD in the past.

Thanks!

 


You prolly need Threadrippa to stay relevant , because it's really easy for software developers to (unnecessarily) Use a crap ton of CPU power due to lazy development.

For example, Battle.net downloads,  some asshol accountant on their end figured they could save on bandwidth bill by high-compression.

This high compression leads to 100% CPU utility while downloading their games on the consumer end.  So now the download speed is dependent on how fast the CPU is on the buyer end, using a crap-ton of Electricity, JUST SO Blizzard can work out a slightly improved balance sheet.  That compression they're using is insanely power intensive.

It really should be illegal for them to do this.

But this is the kind of scenario we're going to run into in the future as AMD opens the door for more cores.

This power-creep will happen on everything, not just games.

I recommend at least a 3900x if cost is a concern.

GPU,  Right now, Turing 20xx series on Nvidia has alot of driver issues. Without competition for the top bracket from AMD, Nvidia is still sandbagging on features and performance.  I'd wait this out until AMD brings parity.


You recommended him a CPU that costs twice as much that he budgeted for both the mobo and CPU. He also said that he does not play any video games so I'm not sure why you're recommending top-tier parts. I just uninstalled Hearthstone and reinstalled it on Battle.net to test your claims and my 3900x did not use more than 10% of the CPU while downloading. Rather than recommending the best parts to him, why not extract more information so you can recommend parts that are tailored to fit his needs.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: tp4tissue on Fri, 20 December 2019, 11:44:07

You recommended him a CPU that costs twice as much that he budgeted for both the mobo and CPU. He also said that he does not play any video games so I'm not sure why you're recommending top-tier parts. I just uninstalled Hearthstone and reinstalled it on Battle.net to test your claims and my 3900x did not use more than 10% of the CPU while downloading. Rather than recommending the best parts to him, why not extract more information so you can recommend parts that are tailored to fit his needs.

Bnet download cpu use depends on the connection speed.

I have 1 gigabit connection.  it downloads at ~80MB/s and that chews in 100% 6 core 12 thread cpu.

If your connection speed is only 100-200 Mbit, whicn is only 25MB/s  that obviously won't take much CPU, because you're not decompressing at a high rate.

My point is simply, it's criminal for them to use compression that way, because it greatly magnifies the total power use from delivery to no great purpose except Superficial savings on bandwidth..

TERRIBLE for the environment.

I am not sure on hearth stone, or if they use variable compression method per game/ per region, but try a larger game, like overwatch.


Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Sniping on Fri, 20 December 2019, 12:24:49
I think the most important factor to me is just having fast storage, if I had that kind of budget I'd only leave $100 for ram and maybe $100-$150 for a CPU and just spend the rest on fast storage. if we're talking development here you'll for sure want fast storage for quickly opening up workspaces and IDE's. I think $100 affords around 16gb ram which is a pleasant amount. 
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: NewbieOneKenobi on Fri, 20 December 2019, 14:11:11
Sigh. I've just found out that one year ago and even half a year ago used 7700Ks sold for 1/3 less than they do now, at least where I live.

For ease of reference, once again, my current rig is i5-6600 (non-K), Asrock Fatal1ty K6 z170 (https://www.asrock.com/MB/Intel/Fatal1ty%20Z170%20Gaming%20K6/), 2x8GB G.Skill Ripjaws 3000/15 (DDR4), (Colorful) GTX 1070ti (8+8).

Options include:

7700K, used: ~$310, perhaps 260 if I keep waiting and am lucky ($50 less than a 9700K, WTF!)
6700K, used: ~$210 and going up
7600K, used, delidded + liquid metal applied, vendor claims does 5 GHz < 70C: ~$210
7600K, used, with MSI Tomahawk z270: ~$230 (this has 2 M.2 slots, which I find useful due to already owning a 256GB NVM drive (and only 1 slot on my current mobo; with the Tomahawk I could add a new SDD instead of swapping)
6600K, used: ~$130

9600K, used: ~$200 cheapest, which is great, but the mobos start at at least $130–140.

Edit: i3-8350K, used: ~$120, this probably undermines the position of 7600K on the list

I won't bore you with the details, but AMD is not an option.

My own mobo goes for around $100-125 used (not sure who in his right mind would pay the price, though), the non-K 6600 often sells better than the K (people like low TDPs?). So by not bundling them, and by putting them through slow bidding auctions without a BuyNow option, I could probably easily recover the cost of e.g. buying a 7600K with a z270 mobo. If not, the loss would be minimal.

And of course there's the option to try to hack my bios to allow me to OC the heck out of the CPU I already have, though I'd need to buy two 12cm Noctuas for my CPU heatsink. Not sure how far non-K OC goes.

What would you guys suggest?

Due to December being a good month financially, I'm not as cash-limited as I usually am, but I'd normally find it hard to justify spending $500 on just the CPU and mobo. I have other expenses, too, so to avoid crossing the point of highly diminished returns would be ideal — basically the best bang for the buck would be ideal. The 7700K, while topping out on my existing slot, is not ideal, as it's just a single-digit percentage improvement in performance over 6700K or 7600K in many applications while costing 50% extra.

I've just checked, and for work I need as much single-core performance as possible, because SDL Trados coders apparently can't learn to use multi-core in 2019. I'll probably OC the CPU for work because of this, just to avoid slowdowns and lag when translating huge files (20K words plus heavy macros, lots of formatting, etc.) segment by segment with humongous XML-based memories and IATE-sized multiterm dictionaries attached (all of which gets parsed every time you confirm a sentence, which means up to a dozen times a minute).

For games, well, it seems i7's are finally gaining a large lead over i5's, obviously probably due to more cores. But at least games use multiple cores these days and any i5 has 4 or 6 of them, even mine.

There's also, I guess, the option to wait for 10000 CPUs to come out. But let's say I were buying right now, from the list above. What would you suggest?

At this point I've half a mind to grab the cheapest 6600K possible and then put my 6600 on a long auction just for the meantime. That's a $80 difference to a 7600K/6700K! And I'd probably come close to a net 0 loss.

What would you do?

Edit: I've just realized that i3-9100 (4-core) is about on par with 6700K/7600K before you start OC-ing. Those little things are ridiculously cheap at $90-ish. That and the cost of the mobo would be about the same range as 7600K with no mobo. The i3-8350K is about $25 more expensive, and it's basically identical to the i5-7600K. For me, it's the kind of OC'ing CPU I wouldn't really lose my sleep over bricking if something went wrong. Afterwards, I could keep an eye out for Coffee Lake i7s.

However, I'm quite excited about the delid 7600K, which I guess I could push harder. But then, I wouldn't have a new mobo (and the ability to sell my current one). Still, the difference in performance between 7600K and 6600K — a single-digit percentage — is not large enough to justify the difference in price as per the above lists, and I'm not sure the good delid on the 7600K makes up for it versus just grabbing a 6600K as cheap as possible.

Heck, I've just realized my i5-6600 could probably sell for more than I'd have to pay for the 8–9th generation i3's.

One seller is selling the 8350K and MSI Z370 Gaming mobo for $205 total. Do I take them?

The reasoning: I pay $205 for these things now. I slow-sell the Asrock Fatal1ty and the locked 6600 for as much as I can, through a long bidding auction, meaning certainty that I will at least sell, perhaps unprofitably, but probably not far below $180 in total. Let's double the difference, so let's say I'm paying $50 net plus some time.

However, I might as well splurge out already and set my eyes on the 9600K. What say ye?

EDIT: Just found another 8350K for $90 and got that MSI Z370 Gaming Plus mobo from the first guy for $80.

If this works, I'm gonna have a mobo that has my exact RAM on its QVL and room for the whole range of Coffee Lake up to i9, and a processor that clocks at 4 Ghz and still has 4 cores, great for a lot of games I play, plus the work. I'm going to send it for professional delidding, buy two 12 cm fans (or actually 14 + converters / 14 on 12 mounts…) for the Thermalright Ultra and fry.

Oh, and sell the Asrock and the i5-6600. Who knows if not for about the same money I've just paid. Probably not much less.

Just hope nothing's broken.

Will keep you guys posted. Thanks for putting up with me so far.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Leslieann on Fri, 20 December 2019, 17:47:38
You prolly need Threadrippa to stay relevant

Drugs are bad, m'kay.


If you aren't a gamer, then AMD seems to be the better choice these days
Only if you are talking very top end, like 2080 GPUs.
Average people are buying AMD, not Intel. Intel can't hardly give away I3, I5 with I7 beings sold to people who already have an I3 or I5 and want an upgrade.

Basically the only Intel selling well to gamers is the 9900k and KS.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Leslieann on Fri, 20 December 2019, 18:01:01
I think the most important factor to me is just having fast storage, if I had that kind of budget I'd only leave $100 for ram and maybe $100-$150 for a CPU and just spend the rest on fast storage. if we're talking development here you'll for sure want fast storage for quickly opening up workspaces and IDE's. I think $100 affords around 16gb ram which is a pleasant amount.
Ryzen 3600X with either current or last gen B series motherboard and 32gigs ram. This puts you about on par with an Intel 8700K, and it will be a beast.
Or, go 16gigs ram and and get a 3700 or 3700K putting you between a 9th gen I7 and I9 then later add another 16gigs.

Just make sure the ram is certified to work (motherboard manufacturers have lists), while not as necessary as it was on earlier Ryzen, it makes for a painless install.

As for Windows,
Is yours oem or retail?  If it's a retail copy you can download tools to extract the activation codes and move it to a new system. I'd pull the code, upgrade to Win10 then pull the new code just to be hassle free. Worst case, buy a code off Ebay for $2-10.


I think the most important factor to me is just having fast storage, if I had that kind of budget I'd only leave $100 for ram and maybe $100-$150 for a CPU and just spend the rest on fast storage. if we're talking development here you'll for sure want fast storage for quickly opening up workspaces and IDE's. I think $100 affords around 16gb ram which is a pleasant amount.
You can add an NVME slot to any board with a free pcie slot for less than $20. It's not directly bootable, but that is easy to fix if you have a second drive, or just use it purely as a data drive. Beware, anything older than 3rd gen will probably be severely bottlenecked by the chipset, still better than a sata drive, but it won't be close to nvme data speeds probably. Make sure it's not a normal ssd on a card, check dtaa rates because even if it's bottlenecked you want higher speed for later upgrades.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Leslieann on Fri, 20 December 2019, 18:07:17
Asrock Fatal1ty K6 z170[/url], 2x8GB G.Skill Ripjaws 3000/15 (DDR4), (Colorful) GTX 1070ti (8+8).

7700K, used: ~$310, perhaps 260 if I keep waiting and am lucky ($50 less than a 9700K, WTF!)
Some 170 boards support some 7th gen chips but generally they only support 6th gen. Check first.


What would you guys suggest?
Either sell the board/chip combo and buy newer or simply invest in 32gigs ram and/or faster ssd which can be moved to a newer setup later.

The price to performance you are looking at by simply switching to a K series isn't worth what you are going to spend.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: fohat.digs on Fri, 20 December 2019, 20:18:57

As for Windows,
Is yours oem or retail?  If it's a retail copy you can download tools to extract the activation codes and move it to a new system. I'd pull the code, upgrade to Win10 then pull the new code just to be hassle free. Worst case, buy a code off Ebay for $2-10.


Was this to me? I have a full retail 8.1 that has been used on at least 2 different motherboards in the past.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Larken on Fri, 20 December 2019, 21:00:21
This seems to be a good place to pose this inquiry:

Sometime in the coming weeks or months I plan to buy a new motherboard/CPU/RAM (which will also necessitate buying Windows, which I have avoided so far via the free upgrade route to 10 from 7/8) for the first time in about 5 years.

I want something with reasonable power but I am not a gamer, so my real criteria are somewhat future-proofing and dependability, although I have no objection to buying "last year's model" to economize.

There is a MicroCenter in town, so I am very fortunate in that regard with all the specials and package deals that they usually offer. Without getting into specific models or numbers, considering a budget in the range of $200-250 for motherboard and processor + $100-150 for RAM, what architecture should I look at?

On Intel vs AMD I am agnostic and have had good experiences with AMD in the past.

Thanks!

 

Ryzen 3600 + a b450 mobo is pretty much the go-to recommendation right now for gamers, but imo it's a pretty good value for non-gamers as well. You could go for the X version, but general consensus is that it is poor value; the only differences appears to be a better stock cooler (Wraith Prism for the X) and a slight bump in stock clock speeds (which is pointless imo because the auto-OC algorithm (PBO2) these CPUs have makes them run similarly). You'd want rams that's at least 3200mhz for Ryzen 3000 series chips; these chips benefit significantly from fast memory. Personally, I'd get the non-X 3600 and use the price difference to get an aftermarket cooler that performs better than the Prism (which to be fair, is a decent stock cooler, just not amazing).

Future proofing wise in terms of upgrade paths, you're not going to get much. AMD only promised to support AM4 compatibility up to 2020; so you're going to get 1 more generation of chips to choose from in the future should you want a drop in chip upgrade before AMD releases a new socket (still better than Intel in this aspect).

Hardware wise, there are some who would say going 8 cores (i.e. 3700x) would be more future proof, but YMMV. The performance increase is nice, but performance/$ is generally on a diminishing return curve past the 3600 on Ryzen 3rd gen, unless you do have a use for the extra cores (for some applications, the extra performance/$ is can be close to linear depending on your choice of brands). I think where you land in terms of chip choice really depends on your budget and whether you need the power.

Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: noisyturtle on Fri, 20 December 2019, 21:48:29
Weirdest thing with this new cpu I've never seen. I've got 4x4gb ram and it won't boot with the full 16 in there. I tried it with just one to make sure it wasn't the cpu and it still didn't boot (red blinking DRAM led on the mobo) I then tried each of the 4 sticks alone in each slot and still got the red blinking DRAM led. Just for fun I try using only 2 sticks and it worked! Why and how? And they only work in two specific slots the furthest from the cpu, not any other way.

It's so weird, but for now I'm gonna have to get by with 8gb ram until I can either figure this out or replace them with 2 8's instead.

If anyone knows why this would happen please enlighten me. I tried forcing specific wattage for ram in bios which did nothing, it just won't boot with all 4 sticks.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Larken on Fri, 20 December 2019, 23:01:51
Try lowering the speed or looser timings. Some times that solves the issue. Had to run my ram slower than it was rated for when using 4 sticks instead of 2 in my current system or it wouldn't post as well.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: tp4tissue on Fri, 20 December 2019, 23:09:09
Try lowering the speed or looser timings. Some times that solves the issue. Had to run my ram slower than it was rated for when using 4 sticks instead of 2 in my current system or it wouldn't post as well.

Also, try overclocking the chipset / memory controller, or giving it more voltage.

might also try to disable auto-training, because that could auto set/ detect mem-timings which would not work.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: noisyturtle on Sat, 21 December 2019, 01:05:03
Try lowering the speed or looser timings. Some times that solves the issue. Had to run my ram slower than it was rated for when using 4 sticks instead of 2 in my current system or it wouldn't post as well.

Also, try overclocking the chipset / memory controller, or giving it more voltage.

might also try to disable auto-training, because that could auto set/ detect mem-timings which would not work.


eh, just attempted overvolting the ram and lowering their frequency to 1333 respectively. Neither attempt worked, but I appreciate the suggestions since I'd not tried those yet.

It's frustrating knowing all 4 sticks work but refuse to work together even though they have been doing so for about 7 years. I'll probably just wind up ordering a 2x8 set, this is more annoyance than it's worth by this point. I thought it must be the bios update, but they were all working fine when I had the 2500k in there. It really makes no sense to me.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: romevi on Sat, 21 December 2019, 02:07:34
Currently using CRT as my second monitor. Maybe it's time to backgrade.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: NewbieOneKenobi on Sat, 21 December 2019, 02:49:41
Currently using CRT as my second monitor. Maybe it's time to backgrade.

My hero!
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Leslieann on Sat, 21 December 2019, 04:14:11
Was this to me? I have a full retail 8.1 that has been used on at least 2 different motherboards in the past.
I thought it was, but I was in a hurry.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: tp4tissue on Sat, 21 December 2019, 08:39:23
Currently using CRT as my second monitor. Maybe it's time to backgrade.

I've avoided wasting my CRT Gun-Life on anything.  Only turns it on for Retro stuff like 4:3 anime, typing of the dead,  SEGA Dreamcast, etc.

This technology will never come back, what we see now is pretty much it.

The entire back end of the industry is already gone.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: NewbieOneKenobi on Sat, 21 December 2019, 13:47:40
Either sell the board/chip combo and buy newer or simply invest in 32gigs ram and/or faster ssd which can be moved to a newer setup later.
The price to performance you are looking at by simply switching to a K series isn't worth what you are going to spend.

Bought the i3-8350K for the $90 (private chap who's already sent it) and spent another $95 an an MSI MPG Z390 Gaming Plus with two M.2 slots (outlet from a huge retail chain) — yay!

Hence my next purchases are going to be a bunch of fans and a larger NVM SSD.

My options are:

A-Data SX8200pro 1 TB for $160 / 0.5TB for $80 (already faster than my current drive (https://ssd.userbenchmark.com/Compare/Adata-XPG-SX8200-Pro-NVMe-PCIe-M2-1TB-vs-Samsung-SM951-NVMe-PCIe-M2-256GB/m638791vsm30950))

vs

Samsung SM970pro 512GB for $176, allegedly 80% faster than my existing drive?

This is the SXC vs SM970pro comparison (https://ssd.userbenchmark.com/Compare/Adata-XPG-SX8200-Pro-NVMe-PCIe-M2-1TB-vs-Samsung-970-Pro-NVMe-PCIe-M2-512GB/m638791vsm498971l) according to User Benchmark. It shows 27% extra sustained write and 70% extra mixed I/O. My work involves working with translation memories and terminological databased that be gigabytes-large and are parsed hard every time you confirm a sentence (or more rarely paragraph) as translated and move on to the next one. The biggest limiting factor is the CPU (the software is 1-core, 32-bit), but I guess a faster SSD would help. Just not sure if the difference is worth it.

And since you mentioned RAM — some $75 for another 2x8GB of what I already have (or sell mine for 60 and get something better) — would you say it's worth it or should I put the money elsewhere? (The i3-8350K is supposed to be a fun project for some tinkering and last me until there's a sweet deal on a 9700K or higher, unless I decide I don't need it.)
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Leslieann on Sat, 21 December 2019, 15:46:59
A-Data SX8200pro 1 TB for $160 / 0.5TB for $80 (already faster than my current drive (https://ssd.userbenchmark.com/Compare/Adata-XPG-SX8200-Pro-NVMe-PCIe-M2-1TB-vs-Samsung-SM951-NVMe-PCIe-M2-256GB/m638791vsm30950))

vs

Samsung SM970pro 512GB for $176, allegedly 80% faster than my existing drive?


And since you mentioned RAM — some $75 for another 2x8GB of what I already have (or sell mine for 60 and get something better) — would you say it's worth it or should I put the money elsewhere? (The i3-8350K is supposed to be a fun project for some tinkering and last me until there's a sweet deal on a 9700K or higher, unless I decide I don't need it.)
I have the Adata (512Gb), works fine and from what I've seen the differences are minimal at best. I had the same choice, and picked the Adata for this very reason, the price/performance made zero sense to get the 970. Moreover, I HIGHLY doubt that cpu will be able to come close to saturating either in almost any workload other than a straight file copy.

Get the ram, I always say more ram is better than less ram* or even faster ram (unless you're Tp who seems to think everyone is a AAA pro gamer).


*Just don't go over 32gigs unless you have a specific use case as very few programs are written to take advantage of that much ram. Frankly most can't even use much more than 16, especially on Windows. Linux and Mac can use it better, but due to how those handle memory, they also have less need for it.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: tp4tissue on Sat, 21 December 2019, 19:16:44
The Adata is a good Mid-tier nvme drive.

The sx8200 pro/ s11pro / hp ex950 / sabrent rocket,  are all essentially the same drive with slightly modified firmware tweaks. the usually holiday sale price is $100 for 1TB

I'd still recommend the 1tb WD Sn750, which has much higher sustained write performance end to end,  whereas because of the way SLC caching works on the adata, it has severely limited performance when the drive is near full, whereas the sn750 holds  onto near peak performance.

Amazon / newegg has the sn750 512gb for ~$70, i believe WD's site is still selling for $130 for the sn750 1tb. 
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Leslieann on Sun, 22 December 2019, 04:16:27
The WD had not even had reviews when I got the Adata but I also subscribe to the idea of never coming close to filling any SSD.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: tp4tissue on Sun, 22 December 2019, 08:17:23
The WD had not even had reviews when I got the Adata but I also subscribe to the idea of never coming close to filling any SSD.


Tp4 only babies CRTs because they're irreplaceable.  SSDs, well they're kinda like air filters, eventually, you just gotta swap um' out, so , crank it.


(https://i.imgur.com/9cLx1zH.gif)
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: NewbieOneKenobi on Sun, 22 December 2019, 10:30:26
I've also just realized that: (1) Samsung 970 Evo is practically just as good as 970 pro while almost twice less expensive, and (2) still has AES encryption, while (3) 1TB SSDs from Samsung perform much faster than 500GB ones from the same line. Obviously there's also the problem of performance limits when the drive is near full. So another option for me is the 'buy and forget' 1TB 970 EVO. I can always sell my existing drive afterwards to help bridge the budget. Probably makes more sense to swap in the 256GB for a 1TB than buy a 500GB to run with the 256GB.

However, I'm wondering if I shouldn't instead choose a 4th-gen Corsair (whatever be that particular mate's precise product name, yarr!) coming with a heatsink already attached. Still, heatsinks can be bought separately if needed.

Sabrent has no luck making it to Poland, though I could get it from German Amazon with free shipment. The WD Sn750 is available, however.

The 4th gen nominally has much better speeds but somehow this doesn't seem to show in tests, where 4th gen drives still place behind Samsung's 970s.

I can get a 970 evo plus 1 TB for $175–190, which I'm inclined to do. Should I? (This is probably comparable to like $120 in America due to price inflation resulting from the EU's VAT, customs, and so on. The usual regular price is much higher with the large retail chains but small online sellers sometimes take it down a quarter or third.

Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: tp4tissue on Sun, 22 December 2019, 10:52:50
I wouldn't pay the premium for 970 evo plus,  Even though I have 2x 1TB of um in my other pcs

I got Sn750 Heatsinked version 1TB for $100 on BF..   That's a way better deal.

Evo,  evo plus, sn750,   They're all pretty much the same in terms of consumer loads, their main advantage is sustained performance even if the drive is full, allowing you to actually use the whole capacity without detriment, stuttering.

It just comes down to price per TB,  I don't think the extra performance in the 970 evo /ev+ is worth the premium, because it's hard to generate sustained random i/o
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: NewbieOneKenobi on Sun, 22 December 2019, 12:05:20
Ouch. Yeah, well, $100 for 1TB is lovely, though it's unlikely to happen in Europe. Importing from a non-VAT country generally means the VAT will be applied on the border along with customs, basically meaning 30% extra, if not more. Hence it's actually better to buy from commercial importers than to import privately from abroad. Sigh.

Re: writes vs reads, sustained vs not, vs mixed, is something I can't yet wrap my mind around, but my applications include:

- keeping large, huge or outright humongous files open and parsing/pinging/searching them all the time for small chunks of data
- updating those files with small chunks of data several times a minute
- occasionally importing input files or exporting output files that are sometimes huge
- whatever's done in games (RPG, RTS, strategies, car racers)

Durability would be good to have. Not in the sense of longevity but rather pure reliability to avoid data loss from some sort of failure (temperature, power, etc.), so that the drive doesn't die on me in the middle of an ASAP/round-the-clock project where being set back by an hour or at least a couple of hours can potentially matter a great deal, which is a high-severity low-probability risk.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: tp4tissue on Sun, 22 December 2019, 12:10:47
Are you sure those applications do operations to disk and not in ram ?
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Leslieann on Sun, 22 December 2019, 16:07:06
I've also just realized that: (1) Samsung 970 Evo is practically just as good as 970 pro while almost twice less expensive, and (2) still has AES encryption, while (3) 1TB SSDs from Samsung perform much faster than 500GB ones from the same line.
The SSD is broken into multiple memory chips (on larger ones), you can assign a controller to each chip, each one adding it's own bandwidth in sort of a mini raid. So if a 500 has 2 controllers, a 1tb has 4, doubling the available bandwidth.


Durability would be good to have. Not in the sense of longevity but rather pure reliability to avoid data loss from some sort of failure (temperature, power, etc.), so that the drive doesn't die on me in the middle of an ASAP/round-the-clock project where being set back by an hour or at least a couple of hours can potentially matter a great deal, which is a high-severity low-probability risk.
If you're doing a truckload of caching and need the writes, look into a small Optane drive, at least some of them are built specifically for it. Though if it was me, I'd either use a normal drive for caching or get more ram.



Tp4 only babies CRTs because they're irreplaceable.  SSDs, well they're kinda like air filters, eventually, you just gotta swap um' out, so , crank it. [/size][/color][/font]
Most of the time the only effort I put into this is buying large enough to where it doesn't matter.

I've filled them on accident a few times over the years (early adopter), but most of the time it was due to something running away/going nuts or me just not emptying trash. The only one I've had fail was due to Microsoft's decisions regarding Win10, on the bright side, it pushed me to Linux.

Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: HoffmanMyster on Mon, 23 December 2019, 17:32:16
Ooh, this thread is relevant to my interests... I'm rocking my 2012 build still, with no updates performed so far (aside from adding an SSD about a month after building it, along with a buttload more storage space).  Debating planning for some incremental upgrades (GPU/CPU) vs full CPU/GPU/RAM/Mobo. 

Current Specs
CPU: i5-3570K
GPU: Radeon HD 7850
RAM: 16 GB (4x 4 GB) DDR3-1600

I think I could definitely benefit from a new GPU (I do game a bit, but not cutting-edge and I don't mind reduced graphics - however, I do enjoy nice graphics if it's not crazy expensive).  Not sure if it's worth doing anything more than that though?  It's a lot harder for me to justify CPU/RAM upgrades without a Mobo upgrade as well because I'm marching further down an obsolete path there... 

And no, tp, I don't want to hear the good word of "Threadrippa" today.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Sintpinty on Mon, 23 December 2019, 18:09:30
I wish i would have spare cash ):
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: tp4tissue on Mon, 23 December 2019, 19:15:19

Current Specs
CPU: i5-3570K
GPU: Radeon HD 7850
RAM: 16 GB (4x 4 GB) DDR3-1600



Depending on the game you want to play,  a 1070ti or 1660ti like noisyturtle just got is more/less a good price/ perf lvl.

Upgrading the CPU is not cost effective. faster 2133mhz / 2400mhz ram will help with Min-FPS if that's important, but prolly too expensive at this point.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Sintpinty on Mon, 23 December 2019, 21:43:07
Ooh, this thread is relevant to my interests... I'm rocking my 2012 build still, with no updates performed so far (aside from adding an SSD about a month after building it, along with a buttload more storage space).  Debating planning for some incremental upgrades (GPU/CPU) vs full CPU/GPU/RAM/Mobo. 

Current Specs
CPU: i5-3570K
GPU: Radeon HD 7850
RAM: 16 GB (4x 4 GB) DDR3-1600

I think I could definitely benefit from a new GPU (I do game a bit, but not cutting-edge and I don't mind reduced graphics - however, I do enjoy nice graphics if it's not crazy expensive).  Not sure if it's worth doing anything more than that though?  It's a lot harder for me to justify CPU/RAM upgrades without a Mobo upgrade as well because I'm marching further down an obsolete path there... 

And no, tp, I don't want to hear the good word of "Threadrippa" today.

Clearly that system does need an upgrade. You've gone that long?
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Larken on Mon, 23 December 2019, 22:52:58
Ooh, this thread is relevant to my interests... I'm rocking my 2012 build still, with no updates performed so far (aside from adding an SSD about a month after building it, along with a buttload more storage space).  Debating planning for some incremental upgrades (GPU/CPU) vs full CPU/GPU/RAM/Mobo. 

Current Specs
CPU: i5-3570K
GPU: Radeon HD 7850
RAM: 16 GB (4x 4 GB) DDR3-1600

I think I could definitely benefit from a new GPU (I do game a bit, but not cutting-edge and I don't mind reduced graphics - however, I do enjoy nice graphics if it's not crazy expensive).  Not sure if it's worth doing anything more than that though?  It's a lot harder for me to justify CPU/RAM upgrades without a Mobo upgrade as well because I'm marching further down an obsolete path there... 

And no, tp, I don't want to hear the good word of "Threadrippa" today.

Depends on how much money you want to throw at it,  but a rx580 would be a nice upgrade for the money. My system was built somewhere around the same year, with a much crappier cpu (fx 8320) with a 5870 gpu. Swapped to an 580 earlier this year and I could run most aaa games on decent settings at 1080p. Ac Odyssey was probably the worst since it is poorly optimized to begin with, but my cpu actually held it back even more (I was getting lower fps compared to online benchmarks for the 580 with a better cpu).
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: noisyturtle on Mon, 23 December 2019, 22:58:24
If you can get a new 1660 ti for the same price as a used 1070 ti, why would anyone ever get the 1660?
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Leslieann on Tue, 24 December 2019, 02:24:52
Current Specs
Depending on the game you want to play,  a 1070ti or 1660ti like noisyturtle just got is more/less a good price/ perf lvl.
Upgrading the CPU is not cost effective.
This...
Your setup is aged, but it's not down for the count in my opinion, especially if it still works for you, so I'd start with the gpu. Being on DDR3 your upgrade path is a lot more limited than if you were already on DDR4 or even an older 2nd gen, which brings up another issue.

If you do wait on the cpu you need to keep in mind DDR5, which will be going mainstream in about 2 years (it will be here sooner, just high priced, not great performance). If you can wait, you will be able to pick up used DDR4 systems for cheap, otherwise you will be buying DDR4 just before we start the switchover or over spending on DD5.  If you can't buy a new mobo/cpu/ram in the next 8-10 months or so I'd probably consider holding off and seeing how things develop over the next few months.

In other words, do it now, or consider waiting even longer.
If it was me, I'd be looking at new and used 2k series Ryzen.


If you can get a new 1660 ti for the same price as a used 1070 ti, why would anyone ever get the 1660?
Warranty, availability, fear of getting a nearly dead mining card.

Plus, a LOOOOT of 1070 and up owners have not gotten the clue that their prices are way out in left field. I saw someone trying to sell a 1080 for $750 and another selling a 1080 TI for $1200 last night (go see what that gets you on Newegg if you don't know). I blame Nvidia and to a lesser extent AMD for this, it kind of left then high and dry for a market, but at the same time, this is what you get when you buy bleeding edge stuff, especially when the market was completely out of whack when you bought it. I'm sorry you spent $1500-2k on a 1080, but that's not my problem.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: NewbieOneKenobi on Tue, 24 December 2019, 07:25:02
The SSD is broken into multiple memory chips (on larger ones), you can assign a controller to each chip, each one adding it's own bandwidth in sort of a mini raid. So if a 500 has 2 controllers, a 1tb has 4, doubling the available bandwidth.

If you're doing a truckload of caching and need the writes, look into a small Optane drive, at least some of them are built specifically for it. Though if it was me, I'd either use a normal drive for caching or get more ram.

Thanks. Well, I think I'll grab a 1TB drive just for that gain, having seen some of the benchmarks. No overkill like 2TB, but 1TB yes.

I'm also almost regretting the purchase of the i3-8350K, but it was so cheap I'm almost certain to make no loss on reselling it, so no big deal really. And more comfort waiting for a good deal on a 9600K or 9700K.

Now getting a bunch of Noctua fans for the CPU and the case.

Any opinion on that, incidentally? I'm definitely going to have one fan blowing into the heatsink, and make it a high-static-pressure fan. Not sure about putting either a high-pressure or high-flow fan on the other side — good suction vs good dissipation. The fan on the end side of the CPU cooler will obviously be just a couple of inches away from the case exhaust fan (14cm), and I could get a high-flow one there. Probably high-flow ones for roof exhaust. Just kind of having last-minute doubts about high-flow fans because of the potential performance loss due to exhausting through mesh, where HP fans could prove better in the end/net result.

With the fans I'm not going to be saving money. I'm not really itching to buy like 6 fans for $50 each, but hypersensitivity to sound + working on this thing with low ambient and needing the ability to concentrate puts things in perspective.

A year or so ago I bought two 14 cm 'Silent' Wings 3 for cause intake… ugh. Can't stand them at anywhere above 500 rpm, and they have a hard time avoiding touching each other in the front of my case, after I used them to replace a broken 20cm Phanteks fan, the one I got with my Enthoo case. It's probably going to be the same with the roof — also 2x14 or 1x20.

One more less than spectacularly intelligent thing I did last year, I bought a PSU that stays totally passive unless it gets a really heavy load. It does so like 7–8'' away from the GPU. Sigh.

Are you sure those applications do operations to disk and not in ram ?

Yeah, the work software does.

I'm probably going to focus some attention on the RAM too in the coming weeks. More, faster, or both.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Leslieann on Tue, 24 December 2019, 18:18:47
I'm also almost regretting the purchase of the i3-8350K, but it was so cheap I'm almost certain to make no loss on reselling it, so no big deal really. And more comfort waiting for a good deal on a 9600K or 9700K.


Now getting a bunch of Noctua fans for the CPU and the case.

Any opinion on that, incidentally? I'm definitely going to have one fan blowing into the heatsink, and make it a high-static-pressure fan. Not sure about putting either a high-pressure or high-flow fan on the other side — good suction vs good dissipation. The fan on the end side of the CPU cooler will obviously be just a couple of inches away from the case exhaust fan (14cm), and I could get a high-flow one there. Probably high-flow ones for roof exhaust. Just kind of having last-minute doubts about high-flow fans because of the potential performance loss due to exhausting through mesh, where HP fans could prove better in the end/net result.

With the fans I'm not going to be saving money. I'm not really itching to buy like 6 fans for $50 each, but hypersensitivity to sound + working on this thing with low ambient and needing the ability to concentrate puts things in perspective.
I don't think it matter too much what fan you use on the back of the heatsink, it's more about assisting the HSP fan on front. I've never gone looking for data, but I don't think it matters too much, couple degrees at most from what I've seen.

My whole take on case fans changed once I went to mini ITX and put it on my desk
On large cases you just need flow. Don't worry too much about how or even so much about how much, just get some in and some out. That may seem over simplified, but really the difference is minimal at best despite people putting so much thought into it. This is where I think people make a mistake, the heat sinks need lots of air flow (Noctuas are good here!) but the case just needs flow in general. Good, larger slow spinning fans are already quiet and do the job fine, except in those critical spots. So long as you are getting flow with no hot spots you will get diminishing returns from then on. On a bigger box, occupy all the slots with something, tune your fan curves and it will probably be fine. Blah Blah Blah 3 degrees... does it really matter? If 1 or two degrees separate you from stable and not, you're not stable.

On mini ITX cases this does NOT apply. Don't try and fudge your way through even if your plan is to stuff a bunch of Noctuas, start with what others have already done. Ignore this at your own peril. Seriously, some of these boxes are critical in how air flows. Installing a fan a certain way can actually raise temps dramatically, even dangerously.


Speaking of quiet...
When you get quiet/slow enough, the bearing type matters more than the fan itself (you want fluid dynamic bearings).
(From an old post)
Quote
At 700 rpm I could hear the fan ball bearings and air flow over the fan blades
At 600 rpm I could mostly only hear the air flow over the fan blades
At 500 rpm I could hear sleeve bearing noise
At 450 rpm I could hear the air flow turbulence over the case edges/fan grill and only just hear the sleeve bearings if you listened.
In all cases I could hear if there was an imbalanced blade and some fans simply wouldn't spin down below 500 or even 600rpm. At this point some would start pulsing the motor which has a really annoying drum/drone effect. If they ran at all.

And this goes with what I said above about large cases,
Stop being super concerned with temps, you don't need to stay under 60C, these parts are designed to last years at max temp. Do you honestly care if your CPU lasts 12 or 15 years as opposed to 20? This was difficult for me to adjust to after years of spending tons to stay under 55 or 60c while silent and when all temp monitors act like 60 is the devil, but really there's no sense in that. These days my fans never go more than 30% or 40% unless the system exceeds 60C. Yes, the system breaks 70c on occasion and sometimes gets close to 80c... But it's well within limits, it's silent and will more than outlast how long I or even the next owner expect to keep it and by then it will be so far outdated it won't matter.

Learn to trade some longevity for silence.
It's hard to look up and see 72c on the CPU or GPU and not get concerned, but that's really a walk in the park for them and nothing to be concerned with.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: absyrd on Wed, 25 December 2019, 04:52:23
Agreed about silence > a few C max temps benching.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: tp4tissue on Wed, 25 December 2019, 12:02:40
Agreed about silence > a few C max temps benching.

2080 Ti , Max overclock, Max Fanspeed.

Put in adjoining room, snake industrial Displayport cable through wall.

Quiet + FTW Graphix.


(https://i.imgur.com/UzALKpm.gif)

That or Bolt mod CPU heatsink to GPU, easily dissipate 200Watt+ @ minimal noise vs GPU standard sinks.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Sintpinty on Wed, 25 December 2019, 12:14:36
Agreed about silence > a few C max temps benching.

2080 Ti , Max overclock, Max Fanspeed.

Put in adjoining room, snake industrial Displayport cable through wall.

Quiet + FTW Graphix.


Show Image
(https://i.imgur.com/UzALKpm.gif)


That or Bolt mod CPU heatsink to GPU, easily dissipate 200Watt+ @ minimal noise vs GPU standard sinks.


bro you actually can afford that?
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: NewbieOneKenobi on Wed, 25 December 2019, 14:56:41
On large cases you just need flow. Don't worry too much about how or even so much about how much, just get some in and some out.

Large case (between mid and big tower) but poor placement I can't really help — the PC is under my desk in the corner of the room. It's not a real tragedy, but I've never had spectacular flows.

Quote
Speaking of quiet...
When you get quiet/slow enough, the bearing type matters more than the fan itself (you want fluid dynamic bearings).
(From an old post)
At 700 rpm I could hear the fan ball bearings and air flow over the fan blades
At 600 rpm I could mostly only hear the air flow over the fan blades
At 500 rpm I could hear sleeve bearing noise
At 450 rpm I could hear the air flow turbulence over the case edges/fan grill and only just hear the sleeve bearings if you listened.
In all cases I could hear if there was an imbalanced blade and some fans simply wouldn't spin down below 500 or even 600rpm. At this point some would start pulsing the motor which has a really annoying drum/drone effect. If they ran at all.

Yeah. During the day I don't really mind the flow sound as much as the bearing sound, and different bearings sound different. When there are other environmental sounds and some general ambient, discernible but civilized sounds coming from underneath the desk don't really matter so much. At night, though, or during very quiet hours I prefer to keep the PC dead silent so I can focus better. This is not normally difficult to achieve as long as the GPU is capable of passive mode and the bearing on the CPU fan isn't broken. What's a bit more challenging is playing certain games I prefer to play without sound. And, of course, I don't want the the fans to outscream my games under heavy loads. I would normally be somewhat reluctant to spend too much money on getting my ideal gaming conditions, but work is a different matter. Water would be overkill, but good air is not something I'm going to skimp much on.

Quote
Stop being super concerned with temps, you don't need to stay under 60C, these parts are designed to last years at max temp.

Yeah, I just don't want to get throttled under big loads or have to interrupt my long gaming sessions. Due to how freelancing works, I sometimes end up playing for like 12–16 hours straight in between jobs — which is when I switch to working for like 12–16 hours stright. ;)

Quote
Do you honestly care if your CPU lasts 12 or 15 years as opposed to 20?

I need 5 years max. I don't care about longevity. I also don't really mind the CPU dying on me after the 2nd or 3rd year unless replacements are expensive.

Quote
This was difficult for me to adjust to after years of spending tons to stay under 55 or 60c while silent and when all temp monitors act like 60 is the devil, but really there's no sense in that. These days my fans never go more than 30% or 40% unless the system exceeds 60C.

Right now: CPU 32C, mobo 36C, CPU fan 460 rpm (up to 600 is inaudible anyway, possibly 800, and 1000 will still be drowned out by any other fan that can be heard), front intakes are 580 and 610 (both same model but need different settings), and that's probably it. This is always sufficient for work, and for games it depends. Latest AAA titles could probably make this box loud.

Quote
Yes, the system breaks 70c on occasion and sometimes gets close to 80c... But it's well within limits, it's silent and will more than outlast how long I or even the next owner expect to keep it and by then it will be so far outdated it won't matter.

At this point in my life, obviously I want the power from the hardware, but the no. 1 consideration with temps is the ability to maintain absolute silence for work and as much silence as possible in gaming.

Quote
Learn to trade some longevity for silence.

Yup. Both longevity and even speed, though I'd like to keep speed compromises to a minimum.

I'll probably end up swapping the the 8350K for a 9600K some time soon, if the overclock ends up causing too high temps in the case, resulting in GPU fans (3x80/90mm) spinning too fast too early. Normally, though, I'd just keep my eye out for good 8700k/9700k deals.

Agreed about silence > a few C max temps benching.

Benching is just benching. ;) Bench, post, brag, get the aaahs, then bump it a couple notches down and enjoy the silence gain.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Leslieann on Wed, 25 December 2019, 16:29:52
I would normally be somewhat reluctant to spend too much money on getting my ideal gaming conditions, but work is a different matter. Water would be overkill, but good air is not something I'm going to skimp much on.
One thing edited out while doing that was the fans I use, I meant to tell you, you don't need Noctuas, especially if it stays under your desk.

Arctic(?) makes some nearly as quiet for half the price with similar flow, but so long as you don't need max flow, don't be scared to look at others. Noise, bearing and rpm are related, almost any slow spinning fan with the right bearing will be darn quiet no matter who makes it. If you can punch out your case for bigger fans, do that first as bigger flows a lot more for less rpm.

On my case pretty much everyone uses Noctuas and while I use them on the cpu and rear fan, I actually have a pair of really cheap Rosewill 140mm Fluid Dynamic Bearing fans in bottom, they cost 1/4 what Noctuas do. So long as I keep them turned down they are just as quiet, which means a loss of flow so I upsized to 140mm to compensate (they barely fit, had to shave 1/2mm on two tabs).  Basically same air and noise as Noctua 120's for 1/4 the price. The rear fan on mine is a Noctua, mostly because the Zalman 92mm I had wouldn't spin down enough even with slow speed adapters. I needed some flow here to keep the GPU heat from saturating the mobo and psu, but but anything more than just movement was creating a lot of noise due to obstructions.


Contrary to popular belief water isn't any more quiet than air.
You have the added pump noise and heat and most of the time people add a second or even third fan. Not only are these fans not deep inside where the sound is muffled, they are high pressure fans beating air against a lot of corners. Most AIO owners would have been better off just buying a high end air cooler, but most see a high end air cooler and think for just a little more they could get a water cooler.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: absyrd on Wed, 25 December 2019, 17:42:06
Pump plus resistance from rad is SO much louder than a quiet air setup, IME.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: fohat.digs on Wed, 25 December 2019, 17:55:13

Being on DDR3 your upgrade path is a lot more limited than if you were already on DDR4 or even an older 2nd gen, which brings up another issue.

If you do wait on the cpu you need to keep in mind DDR5, which will be going mainstream in about 2 years


Is there a particular build/socket/chip set that might put me at a lower but viable sweet spot - as somebody whose primary "demanding" tasks are decent quality sub-professional-level audio and video editing?

   
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: tp4tissue on Wed, 25 December 2019, 20:47:49
Most AIO owners would have been better off just buying a high end air cooler, but most see a high end air cooler and think for just a little more they could get a water cooler.

Air coolers don't have pump whine, so for quietness at the low end it's definitely more quiet.

But AIO still gets better peak cooling.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Leslieann on Thu, 26 December 2019, 02:34:44
Most AIO owners would have been better off just buying a high end air cooler, but most see a high end air cooler and think for just a little more they could get a water cooler.

Air coolers don't have pump whine, so for quietness at the low end it's definitely more quiet.

But AIO still gets better peak cooling.

Not as often as you think
The top end Noctuas can pretty much match or beat any AIO in noise and cooling.

Problem is size and fitment, some of the better air coolers have limited case compatibility while most any modern case will support a dual 120 rad and most people prefer the looks and ease of maintenance an AIO provides (I have to remove the heatsink to disconnect or remove almost any connectors, ram or ssd). Water and aluminum is not as efficient as heat pipes and copper at removing heat, guess what AIO's have, and worse, newer AIOs have moved to thinner radiators.

Here's an interesting video on it.
Here's a good video int.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Larken on Thu, 26 December 2019, 05:29:47
Contrary to popular belief water isn't any more quiet than air.
You have the added pump noise and heat and most of the time people add a second or even third fan. Not only are these fans not deep inside where the sound is muffled, they are high pressure fans beating air against a lot of corners. Most AIO owners would have been better off just buying a high end air cooler, but most see a high end air cooler and think for just a little more they could get a water cooler.

I see this quite often, even when I'm talking with buddies who are hardware enthusiasts. For some reason, they always think that AIOs are quieter than air coolers (to be fair, there are a lot of noisy air coolers out there). Personally, I prefer air coolers for the reasons mentioned above. Actually just ordered a Noctua DH-15 chromax for my upcoming cpu upgrade.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: tp4tissue on Thu, 26 December 2019, 08:23:07
Not as often as you think
The top end Noctuas can pretty much match or beat any AIO in noise and cooling.

Problem is size and fitment, some of the better air coolers have limited case compatibility while most any modern case will support a dual 120 rad and most people prefer the looks and ease of maintenance an AIO provides (I have to remove the heatsink to disconnect or remove almost any connectors, ram or ssd). Water and aluminum is not as efficient as heat pipes and copper at removing heat, guess what AIO's have, and worse, newer AIOs have moved to thinner radiators.

Here's an interesting video on it.
Here's a good video int.

Their system is not scientific, but the range is probably too thin for their measurement setup.

Different blocks are also suited for different die sizes. It's also possible that the AIO could cool an even larger total tdp while the noctua can not, this wasn't tested
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: absyrd on Thu, 26 December 2019, 09:47:49
Contrary to popular belief water isn't any more quiet than air.
You have the added pump noise and heat and most of the time people add a second or even third fan. Not only are these fans not deep inside where the sound is muffled, they are high pressure fans beating air against a lot of corners. Most AIO owners would have been better off just buying a high end air cooler, but most see a high end air cooler and think for just a little more they could get a water cooler.

I see this quite often, even when I'm talking with buddies who are hardware enthusiasts. For some reason, they always think that AIOs are quieter than air coolers (to be fair, there are a lot of noisy air coolers out there). Personally, I prefer air coolers for the reasons mentioned above. Actually just ordered a Noctua DH-15 chromax for my upcoming cpu upgrade.

NH-D15 is dope. Good choice.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Larken on Thu, 26 December 2019, 18:57:23


NH-D15 is dope. Good choice.

Woops. Messed up on the name. But yea, I've been wanting one of them noctuas since the d14, but could never justify it as I already had an aftermarket cooler that performed really well. Needed a new one since I'm moving to am4. Cyrorig r1 and dark rock pro 4 were in contention (on top of being much easier to purchase where I'm at) but I figured I'd probably regret not going for the noctua afterwards. They really do make the best fans.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Leslieann on Thu, 26 December 2019, 19:29:23
Their system is not scientific, but the range is probably too thin for their measurement setup.

Different blocks are also suited for different die sizes. It's also possible that the AIO could cool an even larger total tdp while the noctua can not, this wasn't tested
None of them are scientific, no reviewers have access to that sort of equipment, most are just testing in a room or open air, this is one of the better ones though.


"It's also possible that the AIO could cool an even larger total tdp"
Well yeah, because you can have an infinite sized AIO but that's not realistic now is it?

Stop looking at temps as the end-all-be-all, it's wrong. It doesn't tell you how much the cooler can handle, it's telling you how it performs based on the fans/fan curve it was provided with. Given equal air flow and surface area an air cooler will win pretty much every single time due to efficiency (SCIENCE!). The trouble is fitting it. People need to stop comparing something like a Hyper 212 to a dual 120 AIO, of course the AIO wins. Despite inefficiencies, the AIO has twice the surface area and twice the air flow.


As for block sizes, Asetek, who makes all of the AIOs (they hold the patent and refuse to license it), uses pretty much only one size block. Go look at all the AIOs on Newegg, they all look pretty much the same for this reason. The one real exception (other than the pump in rad AIOs) is the Threadripper purpose built blocks which use the same pump and I'd almost bet a similar internal layout as the rest, which explains why it doesn't really cool much better than a non TR purpose built block. I suspect they may have just made it a larger external surface (acting as a heat spreader) as it means less design and tooling changes.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: 1391401 on Tue, 31 December 2019, 02:51:57
When I was younger I would replace my computer motherboard, CPU, GPU, and often memory every year to two years due to my budget forcing me to buy lower end hardware, the pace of hardware evolution, and the increasing demands of the games I was playing.

In 2009 I built a machine with an i7, 12 GB memory, 320 MB/s SSD and since then have only replaced the GPU since it failed and the CPU since I upgraded to the biggest Xeon that would fit the motherboard (6 core 12 thread).

I also finally upgraded my case from 2001 with a Fractal Design silent case and an AIO cooler.  I finally have a silent PC (my first PC had an Alpha PAL8045 with an 8000 RPM Delta fan on it for those in the know).

I think my demands have slowed down and the pace of hardware evolution has slowed down so the need to upgrade as often just isn't there.  The PC I have gobbles up productivity workloads I throw at it and can still play older games just fine.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: fohat.digs on Tue, 31 December 2019, 08:23:23

In 2009 I built a machine with an i7, 12 GB memory, 320 MB/s SSD

I think my demands have slowed down and the pace of hardware evolution has slowed down so the need to upgrade as often just isn't there. 


In 2009 that was a monstrous beast of a machine, wasn't it?

Not being a gamer, I could now probably go at 6-8 years without feeling the need to modernize.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: jacethesaltsculptor on Tue, 31 December 2019, 09:22:11
I've noticed in my job that other than the great migration to SSD's, the past decade has been a very stagnant one for everything save maybe graphics cards, and SSD sizes. (Monitors too have gotten cheaper and much bigger)

I built my current rig I'm typing to you on around 2012 and it's still going strong, There's few games I can't play at 4K, and I've only upgraded the graphics card and SSD's in the meantime.

I got:

Azza 4000 case
AMD 9590 8-Core at 4.70, (I used to overclock it but didn't really notice a difference. :( It still runs very hot. )
16GB DDR3 @1333
2 TB of storage, and 1TB combined of SSD space. (I bought drives piecemeal.)
Vega 64 Graphics Card
1600 Watt Power supply

The power supply is a hercules sold on Newegg, it was an over-reaction to a crash I was getting with my graphics cards, and I wouldn't do now that I'm older and know better.

I had bought two 6990's and Crossfired them so I technically had 4 GPU's, and it was powerful but not to an extreme extent. Because they were old bitcoin miners they burnt out fast, I owned them for probably a year before they gave up the ghost.

Honestly if you have an i-something, and it's not literally the first gen. (or if it's the i7 first gen) you have an SSD, and a decent graphics card, you'll be hard pressed to max them out in an everyday workload.

Mobile CPU's in laptops will be a different case, but that's also beyond the scope of this post.

I think the Software lastly has a lot to say here.

We're much more rarely getting Triple A games that REALLY push a computer, most of the things that push it now are more hardware related: (E.G. I want a 144hz monitor and need more frames! I want 4k! or I want them both!) Games are being built more and more so anyone can play them on nearly any hardware and the company making the game can make more money. You limit your market when you have to make people go out and buy more hardware to play it. Though I loved doing that back in the day, it was fun for me, it doesn't persuade everyone to do it.

Also OS's have been more about running on Less. Most machines running windows 7 will run 10 fine. I couldn't say the same for the Vista era, a windows 2000 machine would be very unlikely to run Vista well or at all in some cases.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Leslieann on Tue, 31 December 2019, 18:39:43
Hardware has exceeded the software is why.
Vista was peak software and while Win10 is closing in on it, hardware has taken a major jump since Vista (vista was Core 2 and mostly dual core). You could make Vista run well on that era hardware it just took VERY high end parts which most didn't have, coming from old XP machines MS and manufacturers told them it would run on. It didn't and they knew it, but HP and Dell had thousands of low end ($300-$400) systems ready to sell and people would have balked had they been told they actually needed to spend $1500+.

As for stagnation... Yes and no
Monitors are cheaper, unless you want new-ish features then you pay through the nose. Monitors have always been that way, it's a slow industry.
CPUs and GPUS... electrical trace size determine speed and it's a race to zero, getting more and more difficult to extract more speed. Core count was a way to offset it but it's not really a replacement, we need an architecture shift for a major leap at this point. This was why SSDs were so important, it was a major architecture shift.

Anything NVME will probably last you a long time as there's nothing major coming down the pipe.
Yes, more cores, but trust me, once you get 6-8 cores and at least as many threads, more matters less and less. Do I drool over 16core/32 thread, absolutely, will it put my 8700k to shame, only in very limited workloads that few ever really see. Cores won't fix the issue, it's an architecture/software problem that's not easy to solve.


The next big shift is actually going to be on motherboards and connectors.
MATX is dead (some companies are dropping it already), SLI is dead, ATX isn't really needed anymore. Expect ITX to become the dominant form factor soon. All you had to do was pay attention to holiday inventory to see where the market is going, you couldn't find a sale on SFX power supplies and many are still sold out. Seems everyone bought AMD Ryzen, ITX boards and 1660/2060 GPUs. In fact almost the entire North American inventory of full size 1660/2060 cards was depleted in early November. Everyone is shrinking the box, case makers are lagging but they are coming. BTW, EVGA is still trying to recover from the November wipe-out. That's why sales on those GPUs parts were limited on Black Friday/Cyber Monday, they were out of inventory before we even got there.

Longer term is connectors.
I expect a single interface to appear for internal and external connectors, do we really need an internal USB 3.1 usb connector when they could just embed a Type C connector on the board? Same for Sata and RGB (which needs a standardized connector), switch them to Type C. All those connectors are slower than USB 3.2 so we no longer need them. Make a small bank of Type C and call it a day. This is slowly starting to happen with a few companies including an internal usb connector. I don't know how soon this will happen, but the sooner the better. Personally, ditch all the sata slots and give me an extra NVME slot. We also may see a switch to so-dimm (laptop) memory as it makes it easier to fit everything on the smaller boards. This alone would give room for another nvme slot.

While I don't see it happening, I'd also like to see a different cable system, ATX12v is aging and showing it's limitations. We need a better system than that stupid massive 24 pin, 8 pin, and now becoming more common a second 8 pin. It's time to update it. And can we FINALLY get a common pinout/connector for the front panel. Why are we still having to look at manuals to figure out the pinout and connect each manually. This is ridiculous, surely they can standardize this.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: noisyturtle on Wed, 05 February 2020, 21:02:08
After some long term testing and review I am legitimately impressed with the noticeable performance difference going from a i5-2500k to a i7-3770k. Both were overclocked to 4.2 and 4.5 GHz respectively, the i7 runs a bit hotter, topping out at around 160f under max pressure stress testing. Thankfully I don't generally come close to that, and it hovers around 100-120f most of the time in it's OC state with nothing more than a mid-level Scythe CPU cooler.
Win and in-Win apps are running easily double the speed (tested before in Win 7 environment, and again in Win 10,) and current gaming titles like Red Dead Redemption II run noticeably smoother under generally higher settings.

Upgrades all told:
i5 2500k OC 4.2 -> i7 3770k OC 4.6 GHz
16gb -> 32gb RAM
Win 7 -> Win 10 ):
Monitor 60 Hz -> 244 Hz

All told, not including monitor: $185
Staving off having to buy a new computer for another year or so: PRICELESS

I could get even more mileage out of this build if there were a 1070 TI available anywhere on Earth for a reasonable price. If anything they have been going up in price lately?
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Leslieann on Thu, 06 February 2020, 03:27:41
Be sure to price shop any used cards, prices have dropped considerably in the last month.

In about one month 1080 prices dropped almost 20%, I've seen a few EVGA models struggle to reach $270 on auction. This has compressed the 1070, 1070 ti and 1080's into a small space with barely $100 separating them all. Why get a 1070 ti (which for some reason are scarce) when $40 more will get you a 1080 which are all over the place.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Sintpinty on Thu, 06 February 2020, 05:18:16
Be sure to price shop any used cards, prices have dropped considerably in the last month.

In about one month 1080 prices dropped almost 20%, I've seen a few EVGA models struggle to reach $270 on auction. This has compressed the 1070, 1070 ti and 1080's into a small space with barely $100 separating them all. Why get a 1070 ti (which for some reason are scarce) when $40 more will get you a 1080 which are all over the place.

Broke me is hyped
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: tp4tissue on Thu, 06 February 2020, 08:58:26
Be sure to price shop any used cards, prices have dropped considerably in the last month.

In about one month 1080 prices dropped almost 20%, I've seen a few EVGA models struggle to reach $270 on auction. This has compressed the 1070, 1070 ti and 1080's into a small space with barely $100 separating them all. Why get a 1070 ti (which for some reason are scarce) when $40 more will get you a 1080 which are all over the place.

Not sure 1080 is the way to go though.

Recent events, from the 2D side, it seems that Madvr will require a 1080Ti for HDR Tonemapping video playback.

From the 3D side, there's rumor that Nvidia will double Ray-Trace performance in the next gen card, making Ray Trace an actual playable feature.

Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Leslieann on Thu, 06 February 2020, 20:22:15
Not sure 1080 is the way to go though.

Recent events, from the 2D side, it seems that Madvr will require a 1080Ti for HDR Tonemapping video playback.

From the 3D side, there's rumor that Nvidia will double Ray-Trace performance in the next gen card, making Ray Trace an actual playable feature.
Double RTX... *yawn*
Wake me when it's worth actually having before I spend money to double it.

I'm sure at some point it will be a normal thing like shadows were once they became common, but dismissing a heavily discounted last gen card over a feature that is years away from even being a common thing much less a necessity (by which time you could upgrade again for cheap) is just silly.


As for MADVR,
You're talking about specialized software doing specialized encoding and playback while encoding, not sure what that has to do with anything mentioned here. The number of people who need that feature are few and far between and would be far better served on a forum at least dedicated to the software.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: absyrd on Sat, 08 February 2020, 02:40:42
Ray-tracing at a decent price-point seems far out.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: tp4tissue on Sat, 08 February 2020, 07:29:10
Ray-tracing at a decent price-point seems far out.

They can crank it out tomorrow if not for GREEEEED.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: el_murdoque on Sat, 08 February 2020, 08:00:02
How do you decide when the time is right? Tech keeps improving and getting cheaper at a really fast rate these days, perhaps waiting 2-3 more years is better?

First of all, I need to hit that point. My office machine is a potato, but it runs Ubuntu from an SSD and I don't spend time waiting for applications to open or close or do their stuff. It all happens instantly, so I'm happy and in no mood to upgrade.
On my home machine, I do occasionally play games, even though my busy schedule allows less and less of that.
I also use more demanding software, so that machine has a bigger pot to cook.
So when I hit that point where the game I installed does not run smoothly on settings between high and medium -
or applications take a while to launch or I get angry with the lack of power in any other way, it's time to do something.
My personal happiness with the performance of my system is the only benchmark that counts. I've had people commenting on my machines being potatoes before, but that never sparked anything.
So when I want more juice, I decide whats best - new parts or new system.  When I go new system, I usually buy something used off ebay, rip both the new and my old box apart and frankenstein a system out of the best parts. With any luck, there will be enough parts left to build another working system that will go either to a good case or on ebay.

I upgrade parts when either my GPU or CPU are old and the mainboard is able to take a serious upgrade. Right now, I have something like the second best possibility there is in my CPU slot. The best, while being better, is not that much of an upgrade. So as soon as I run into CPU-being-slow-issues,
I'll need a new system. I won't get a new Mainboard and migrate everything else. The work involved usually is the same as building a new system.

Tech keeps improving and getting cheaper at a really fast rate these days, perhaps waiting 2-3 more years is better?
I've built my first machine in 1994, as an upgrade to an outdated 80286 that came pre-build.
I spent about three grand on a 80mHz CPU with a staggering 8 megs of RAM. To round it off, a 400 meg harddrive (so big, you'll never have to worry about space). Tech was improving fast and getting cheaper at a really fast rate back then and it has never stopped until now. It will, with a very high probability go on doing just that.
So basically, it is always the right or wrong moment to buy. "The new <chipset/gpu/cpu/whatever> will be released in three weeks" is a statement that is true at any given time.
 





Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: absyrd on Sat, 08 February 2020, 11:14:09
Just grabbed a 1660 super instead of a 5600 xt. I have no time for driver headaches right now.

Also... someone talk me into a non-cheapo wireless mouse. Which one? Budget?
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: tp4tissue on Sat, 08 February 2020, 13:53:09
Just grabbed a 1660 super instead of a 5600 xt. I have no time for driver headaches right now.

Also... someone talk me into a non-cheapo wireless mouse. Which one? Budget?

There's only Logitech, the only reliable wireless platform.

G305 is Tp4 recommends.  The other ones with rechargeable require disassembly to swap battery.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: absyrd on Sat, 08 February 2020, 14:21:34
Just grabbed a 1660 super instead of a 5600 xt. I have no time for driver headaches right now.

Also... someone talk me into a non-cheapo wireless mouse. Which one? Budget?

There's only Logitech, the only reliable wireless platform.

G305 is Tp4 recommends.  The other ones with rechargeable require disassembly to swap battery.


Perfect. I don't need 40 ****ing buttons on my mouse. And $50 seems reasonable.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Sintpinty on Sun, 09 February 2020, 10:00:08
Just grabbed a 1660 super instead of a 5600 xt. I have no time for driver headaches right now.

Also... someone talk me into a non-cheapo wireless mouse. Which one? Budget?

G305 is the best option. Only has two side buttons and can be modded easily
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Maledicted on Tue, 11 February 2020, 12:09:37
From the 3D side, there's rumor that Nvidia will double Ray-Trace performance in the next gen card, making Ray Trace an actual playable feature.
Double RTX... *yawn*
Wake me when it's worth actually having before I spend money to double it.

I'm sure at some point it will be a normal thing like shadows were once they became common, but dismissing a heavily discounted last gen card over a feature that is years away from even being a common thing much less a necessity (by which time you could upgrade again for cheap) is just silly.

Agreed. RTX is a joke.


There's only Logitech, the only reliable wireless platform.

G305 is Tp4 recommends.  The other ones with rechargeable require disassembly to swap battery.


I've always loved how all of their wireless devices last months and months on disposable batteries, and those unifying receivers are wonderful.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: noisyturtle on Tue, 11 February 2020, 15:46:45
I am intrigued by the passively cooled 1650. It would be neat to do a dead-air build with no fans in the case at all.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: bliss on Tue, 11 February 2020, 16:32:29
It would be neat to do a dead-air build with no fans in the case at all.
Been there, done that - it's worth it!

However, the passive 1650 at 75W is clearly meant for active case cooling, not unlike a server card.

My build uses a 25W Xeon, passively cooled PSU and a GPU where fans stop at idle. With this you can play games for hours no problem (GPU fans spin), while the PC is dead quiet (no fans) when the GPU is not stressed. Tried a XFX R7 250 Passive, got too hot in that PC! I have got a case with top vent to facilitate the stack-effect.

Rather than a passive 1650, I would try a Ryzen 5 3400G with built-in graphics, possibly at TDP-down, with a really large (passive) cooler. Could do fanless  gaming or whatever :thumb:
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: tp4tissue on Tue, 11 February 2020, 17:09:01
It's easier to snake a cable through another room ,  vs a no fan build.

Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: fanpeople on Tue, 11 February 2020, 17:21:32
It's easier to snake a cable through another room ,  vs a no fan build.



I will snake my cable into your room.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: absyrd on Tue, 11 February 2020, 19:49:29
1660 super was a pita. Got 5700 xt. This thing rips so far.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: fohat.digs on Tue, 11 February 2020, 20:12:12
I think that fanpeople may need help if he has sunk to hitting on TP4
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: tp4tissue on Tue, 11 February 2020, 20:31:40
I think that fanpeople may need help if he has sunk to hitting on TP4


Fohat yet again underestimating Tp4's immense PG appropriate charisma
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: fanpeople on Tue, 11 February 2020, 20:32:48
I think that fanpeople may need help if he has sunk to hitting on TP4

It is more of a doggy dominance thing. Nothing sexual, pure domination.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: tp4tissue on Tue, 11 February 2020, 20:44:58
I think that fanpeople may need help if he has sunk to hitting on TP4

It is more of a doggy dominance thing. Nothing sexual, pure domination.

Help.. I need an Adult..
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: fanpeople on Tue, 11 February 2020, 21:41:44
I think that fanpeople may need help if he has sunk to hitting on TP4

It is more of a doggy dominance thing. Nothing sexual, pure domination.

Help.. I need an Adult..

...... tonight.... you!
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Leslieann on Tue, 11 February 2020, 22:07:19
I am intrigued by the passively cooled 1650. It would be neat to do a dead-air build with no fans in the case at all.
Most of these expect some airflow in the box, especially smaller boxes.
That doesn't mean you can't use LARGE, SLOW fans, and if the right ones (Noctua and a few others) you won't hear them, but you generally need something.

I too have thought of doing this, came really close on one of my servers, only fan was a super slow 140mm, but for a gaming/high performance system it means too many compromises.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: tp4tissue on Wed, 12 February 2020, 07:09:21
...... tonight.... you!

You are trying to buy fried fish,  Tp4 only sells newspapers.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Maledicted on Wed, 12 February 2020, 07:35:45
I think that fanpeople may need help if he has sunk to hitting on TP4

It is more of a doggy dominance thing. Nothing sexual, pure domination.

Help.. I need an Adult..

...... tonight.... you!

Is that a hand banana reference? Good old Aqua Teen Hunger Force.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: fanpeople on Wed, 12 February 2020, 19:58:07
I think that fanpeople may need help if he has sunk to hitting on TP4

It is more of a doggy dominance thing. Nothing sexual, pure domination.

Help.. I need an Adult..

...... tonight.... you!

Is that a hand banana reference? Good old Aqua Team Hunger Force.

 :thumb:
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: absyrd on Sat, 15 February 2020, 14:23:28
Just grabbed a 1660 super instead of a 5600 xt. I have no time for driver headaches right now.

Also... someone talk me into a non-cheapo wireless mouse. Which one? Budget?

There's only Logitech, the only reliable wireless platform.

G305 is Tp4 recommends.  The other ones with rechargeable require disassembly to swap battery.


Got the g305. It is friggin light as hell. I'm used to heavy ****. This is gonna take a while to get used to...

Nice mouse for the price, though. Thanks.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: tp4tissue on Sat, 15 February 2020, 20:34:06
Got the g305. It is friggin light as hell. I'm used to heavy ****. This is gonna take a while to get used to...

Nice mouse for the price, though. Thanks.

Do dat Flick shot.  Quake 3 Arena.

Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Leslieann on Sun, 16 February 2020, 02:44:48
For anyone looking to upgrade their SSD and Ram, DO IT NOW!!!!

Manufacturers over supplied the market and began cutting back production before Thanksgiving, supplies are now down to where they want so prices are going to go up. Worse, Sony is sourcing parts for the Playstation 5 right now (and complaining about prices), so once the current supply stateside dries up prices are going to shoot up fast. You would think these companies could plan better but they never do.

Don't be surprised if ram prices jump 30-50% by April and to a lesser degree SSDs. Don't expect prices to get back to where we are until we get towards the end of summer, possibly early Fall.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: suicidal_orange on Sun, 16 February 2020, 05:08:51
For anyone looking to upgrade their SSD and Ram, DO IT NOW!!!!
...
Don't be surprised if ram prices jump 30-50% by April and to a lesser degree SSDs. Don't expect prices to get back to where we are until we get towards the end of summer, possibly early Fall.

Sounds like I'm due a CPU/RAM failure right around May, got a spare mobo so that will survive for a change :rolleyes:
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: tp4tissue on Sun, 16 February 2020, 23:35:27
For anyone looking to upgrade their SSD and Ram, DO IT NOW!!!!
...
Don't be surprised if ram prices jump 30-50% by April and to a lesser degree SSDs. Don't expect prices to get back to where we are until we get towards the end of summer, possibly early Fall.

Sounds like I'm due a CPU/RAM failure right around May, got a spare mobo so that will survive for a change :rolleyes:

Threadrippppa
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: suicidal_orange on Mon, 17 February 2020, 04:49:19

Threadrippppa

NUC would be more appropriate for my use :p

More likely 6 core Ryzen and much moaning about paying for unnecessary performance upgrade, like usual.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: tp4tissue on Mon, 17 February 2020, 15:52:27
NUC would be more appropriate for my use :p

More likely 6 core Ryzen and much moaning about paying for unnecessary performance upgrade, like usual.

Threadripppa NUC... !!
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Maledicted on Mon, 17 February 2020, 15:55:15
NUC would be more appropriate for my use :p

More likely 6 core Ryzen and much moaning about paying for unnecessary performance upgrade, like usual.

Threadripppa NUC... !!


This immediately reminded me of those rednecks that jam as huge of v8 engines as they can into lawnmowers.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: tp4tissue on Mon, 17 February 2020, 17:36:23
This immediately reminded me of those rednecks that jam as huge of v8 engines as they can into lawnmowers.

There's bound to be an itx board eventually.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Leslieann on Mon, 17 February 2020, 17:54:50
People have rendered it, and while possible and at least one company is interested, it's unlikely.

The socket itself takes up almost 25% of the board, by the time you add IO there's no room for full size ram, much less multiple channels of it which are needed to take full advantage of Threadripper. Most ideas so far have centered around laptop ram on the back of the board and still comes up short for wifi and m.2 space. Even MATX has been tough for board makers, I think only Asrock has done it.

Here's (https://smallformfactor.net/forum/threads/sff-threadripper-motherboard-petition-thread.2340/) a thread discussing it, including renders and size comparisons, but if it ever does happen, it will be one very expensive motherboard for what you get.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: tp4tissue on Mon, 17 February 2020, 18:25:05
People have rendered it, and while possible and at least one company is interested, it's unlikely.

The socket itself takes up almost 25% of the board, by the time you add IO there's no room for full size ram, much less multiple channels of it which are needed to take full advantage of Threadripper. Most ideas so far have centered around laptop ram on the back of the board and still comes up short for wifi and m.2 space. Even MATX has been tough for board makers, I think only Asrock has done it.

Here's (https://smallformfactor.net/forum/threads/sff-threadripper-motherboard-petition-thread.2340/) a thread discussing it, including renders and size comparisons, but if it ever does happen, it will be one very expensive motherboard for what you get.

They could do a small perpendicular breakout board just like the extra vrms.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Maledicted on Mon, 17 February 2020, 18:30:21
Personally, I would rather it be bigger and just be able to work inside of the thing anyway. Not sure why anybody needs something that powerful in that sort of form factor, with upgradeability/customization.

This is also coming from somebody that's worked, almost exclusively, on cheap Chromebooks and laptops that probably belonged in a wood chipper straight from the factory floor, for 5 years straight. Getting into a desktop at all again is like a breath of fresh air.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: suicidal_orange on Mon, 17 February 2020, 18:33:16
... Not sure why anybody needs something that powerful in that sort of form factor ...
Old school LAN party server? :))
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Maledicted on Mon, 17 February 2020, 18:36:13
... Not sure why anybody needs something that powerful in that sort of form factor ...
Old school LAN party server? :))

If it requires a server, is it an old school LAN party?  :eek:

Makes me think more of just routers and/or switches.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: suicidal_orange on Mon, 17 February 2020, 19:08:58
... Not sure why anybody needs something that powerful in that sort of form factor ...
Old school LAN party server? :))

If it requires a server, is it an old school LAN party?  :eek:

Makes me think more of just routers and/or switches.
I think LAN parties as I knew them ('old school') were replaced years ago by fast internet connections and headsets allowing realtime communication worldwide but yes, back in the day you had to have dedicated game server as computers were too slow plus a network switch and wires.  As specs improved the need for a server went away leaving just the switch and wires so I guess you're a bit younger than me :)
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Maledicted on Mon, 17 February 2020, 19:17:42
... Not sure why anybody needs something that powerful in that sort of form factor ...
Old school LAN party server? :))

If it requires a server, is it an old school LAN party?  :eek:

Makes me think more of just routers and/or switches.
I think LAN parties as I knew them ('old school') were replaced years ago by fast internet connections and headsets allowing realtime communication worldwide but yes, back in the day you had to have dedicated game server as computers were too slow plus a network switch and wires.  As specs improved the need for a server went away leaving just the switch and wires so I guess you're a bit younger than me :)

I may be. Are we talking 10/100 "fast"? What old games required local servers for a LAN connection? Are we talking 1980s? Earlier?
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: suicidal_orange on Mon, 17 February 2020, 19:45:44
I may be. Are we talking 10/100 "fast"? What old games required local servers for a LAN connection? Are we talking 1980s? Earlier?
Haha!  I think in the 80s they were writing pong and losing it on every reboot...

Struggling to work out when but definitely remember the original Counter Strike had a standalone server, on searching it seems CS:GO did too and that was only 2015 so maybe somewhere kids are still getting some exercise carrying their systems around.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Leslieann on Mon, 17 February 2020, 19:52:06
hey could do a small perpendicular breakout board just like the extra vrms.
With the amount of heat TR puts out, the size of an air cooler will take up pretty much entire mobo. Many small ITS systems already require low profile ram to clear the heat sinks. If you go water, you still need a place to squeeze the hoses out and all this vertical stuff gets in the way.

It's already an issue just routing wiring in many of these cases without disturbing airflow. Also keep in mind, this is a problem with 95watt cpus, I can't even imagine trying to cool a 270 watt Threadripper in my Ncase.  If you need a larger case just for the cooling, why saddle it with a compromised motherboard.


If you want to see what you are facing here's my adventure with a "95 watt" 8700k in an Ncase.

More
Here is how tight an Ncase M1 with Noctua C14S fits, however you need to replace the 140mm fan with a 120 to even get this. While I didn't use a Noctua fan when I tried this I was uncomfortable with system temps.
[attach=3]


Pic by M1AF on Reddit
https://www.reddit.com/r/sffpc/comments/ahn4e8/ncase_m1_c14s_stock_psu_mounting_it_fits/

You can turn the psu sideways,  but then you have this (most also mount dual 120 or 140s under the GPU and a 90mm at the back of the case, filling the case entirely).
[attach=2]

Shot from the top
[attach=1]

Pics from a user named Zoob, but closely match what my system looked like inside before I changed to a Noctua U9S.
https://hardwarecanucks.com/forum/threads/ncase-m1-v1.79119/


The C14S ran fine when the PSU was sideways the problem is the PSU was sucking massive amounts of hot air. Under modest loads the Silverstone 450 watt PSU was running at or above recommended temps, and under full load, the heat and noise was terrible. Switching to a 650 Corsair helped, but even then I was pushing max temps while gaming.  I switched to a U9S, which had less cooling capability, added an extra fan, and while it let the cpu heat up a little more under load the rest of the system, particularly the PSU, were kept cooler.

Understand, this is one of only about 3 coolers capable of handling an 8700k or 9900k in such a small box and those are 90-145 watt CPUs, Threadripper can reach 270 watts. Basically once you pass about 200 watts between CPU and GPU things get really complicated in terms of cooling. And if you think these cases are stuffed, the Ncase isn't even considered all that small these days.

Edit, missed a pic
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Leslieann on Mon, 17 February 2020, 19:56:12
I think LAN parties as I knew them ('old school') were replaced years ago by fast internet connections
Dreamhack is still a thing.
https://dreamhack.com/
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: tp4tissue on Mon, 17 February 2020, 21:11:07
Problem with lanparty these days is, I've outgrown most games, nostalgia from time to time, but they're hard to dive into for long.

Ya build the bunker, send dudez to the other base,  meanwhile australia is burning down afk, and the back of your mind is thinking, TP4, Australia is burning, you're wasting humanity's resources playing vidya. 
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: noisyturtle on Tue, 18 February 2020, 01:17:40
Deadmau5 has a sick ass LAN party room with a dedicated T1 internet connection.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Leslieann on Tue, 18 February 2020, 02:18:48
Deadmau5 has a sick ass LAN party room with a dedicated T1 internet connection.
These days a common cable connection is waaay faster than a T1 (a T1 is 1.54Mbit/s up/down).

The only benefit a dedicated T1 has is you can legally run a server on it, has a dedicated IP, and comes with better (you hope) tech support. You are better off with a business class cable connection, even with some lag due to neighbors it smokes a T1.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: noisyturtle on Tue, 18 February 2020, 02:50:45
Deadmau5 has a sick ass LAN party room with a dedicated T1 internet connection.
These days a common cable connection is waaay faster than a T1 (a T1 is 1.54Mbit/s up/down).

The only benefit a dedicated T1 has is you can legally run a server on it, has a dedicated IP, and comes with better (you hope) tech support. You are better off with a business class cable connection, even with some lag due to neighbors it smokes a T1.

Probably that's what he's got, idk. I just say T1 as like a coverall for a dedicated optic line, like sayin I want a Coke when what I really want is a Fanta. I should probably stop doin that.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: suicidal_orange on Tue, 18 February 2020, 03:36:17
I think LAN parties as I knew them ('old school') were replaced years ago by fast internet connections
Dreamhack is still a thing.
https://dreamhack.com/
True, but you would never have taken your server there.  Back in the day 5 friends would pack into a living room, now they can just put on headsets and stay home.

Problem with lanparty these days is, I've outgrown most games, nostalgia from time to time, but they're hard to dive into for long.

Ya build the bunker, send dudez to the other base,  meanwhile australia is burning down afk, and the back of your mind is thinking, TP4, Australia is burning, you're wasting humanity's resources playing vidya. 


Vidya makes you guilty about afk problems but whether you're playing or not your basement datacenter is running 24/7 and that's fine?  I get the thinking behind it but as with anything reconciling personal actions against a global problem seems pointless - I turn off lights when not in use but most offices in New York are lit 24/7.

Still no cure for Coronavirus, the world has a chance :thumb:
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: absyrd on Tue, 18 February 2020, 03:54:37
hey could do a small perpendicular breakout board just like the extra vrms.
With the amount of heat TR puts out, the size of an air cooler will take up pretty much entire mobo. Many small ITS systems already require low profile ram to clear the heat sinks. If you go water, you still need a place to squeeze the hoses out and all this vertical stuff gets in the way.

It's already an issue just routing wiring in many of these cases without disturbing airflow. Also keep in mind, this is a problem with 95watt cpus, I can't even imagine trying to cool a 270 watt Threadripper in my Ncase.  If you need a larger case just for the cooling, why saddle it with a compromised motherboard.


If you want to see what you are facing here's my adventure with a "95 watt" 8700k in an Ncase.

More
Here is how tight an Ncase M1 with Noctua C14S fits, however you need to replace the 140mm fan with a 120 to even get this. While I didn't use a Noctua fan when I tried this I was uncomfortable with system temps.
(Attachment Link)


Pic by M1AF on Reddit
https://www.reddit.com/r/sffpc/comments/ahn4e8/ncase_m1_c14s_stock_psu_mounting_it_fits/

You can turn the psu sideways,  but then you have this (most also mount dual 120 or 140s under the GPU and a 90mm at the back of the case, filling the case entirely).
(Attachment Link)

Shot from the top
(Attachment Link)

Pics from a user named Zoob, but closely match what my system looked like inside before I changed to a Noctua U9S.
https://hardwarecanucks.com/forum/threads/ncase-m1-v1.79119/


The C14S ran fine when the PSU was sideways the problem is the PSU was sucking massive amounts of hot air. Under modest loads the Silverstone 450 watt PSU was running at or above recommended temps, and under full load, the heat and noise was terrible. Switching to a 650 Corsair helped, but even then I was pushing max temps while gaming.  I switched to a U9S, which had less cooling capability, added an extra fan, and while it let the cpu heat up a little more under load the rest of the system, particularly the PSU, were kept cooler.

Understand, this is one of only about 3 coolers capable of handling an 8700k or 9900k in such a small box and those are 90-145 watt CPUs, Threadripper can reach 270 watts. Basically once you pass about 200 watts between CPU and GPU things get really complicated in terms of cooling. And if you think these cases are stuffed, the Ncase isn't even considered all that small these days.

Edit, missed a pic

Nice! Love the ncase.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Sintpinty on Tue, 18 February 2020, 06:04:25
hey could do a small perpendicular breakout board just like the extra vrms.
With the amount of heat TR puts out, the size of an air cooler will take up pretty much entire mobo. Many small ITS systems already require low profile ram to clear the heat sinks. If you go water, you still need a place to squeeze the hoses out and all this vertical stuff gets in the way.

It's already an issue just routing wiring in many of these cases without disturbing airflow. Also keep in mind, this is a problem with 95watt cpus, I can't even imagine trying to cool a 270 watt Threadripper in my Ncase.  If you need a larger case just for the cooling, why saddle it with a compromised motherboard.


If you want to see what you are facing here's my adventure with a "95 watt" 8700k in an Ncase.

More
Here is how tight an Ncase M1 with Noctua C14S fits, however you need to replace the 140mm fan with a 120 to even get this. While I didn't use a Noctua fan when I tried this I was uncomfortable with system temps.
(Attachment Link)


Pic by M1AF on Reddit
https://www.reddit.com/r/sffpc/comments/ahn4e8/ncase_m1_c14s_stock_psu_mounting_it_fits/

You can turn the psu sideways,  but then you have this (most also mount dual 120 or 140s under the GPU and a 90mm at the back of the case, filling the case entirely).
(Attachment Link)

Shot from the top
(Attachment Link)

Pics from a user named Zoob, but closely match what my system looked like inside before I changed to a Noctua U9S.
https://hardwarecanucks.com/forum/threads/ncase-m1-v1.79119/


The C14S ran fine when the PSU was sideways the problem is the PSU was sucking massive amounts of hot air. Under modest loads the Silverstone 450 watt PSU was running at or above recommended temps, and under full load, the heat and noise was terrible. Switching to a 650 Corsair helped, but even then I was pushing max temps while gaming.  I switched to a U9S, which had less cooling capability, added an extra fan, and while it let the cpu heat up a little more under load the rest of the system, particularly the PSU, were kept cooler.

Understand, this is one of only about 3 coolers capable of handling an 8700k or 9900k in such a small box and those are 90-145 watt CPUs, Threadripper can reach 270 watts. Basically once you pass about 200 watts between CPU and GPU things get really complicated in terms of cooling. And if you think these cases are stuffed, the Ncase isn't even considered all that small these days.

Edit, missed a pic

Nice sff build
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Maledicted on Tue, 18 February 2020, 07:43:26
I may be. Are we talking 10/100 "fast"? What old games required local servers for a LAN connection? Are we talking 1980s? Earlier?
Haha!  I think in the 80s they were writing pong and losing it on every reboot...

Struggling to work out when but definitely remember the original Counter Strike had a standalone server, on searching it seems CS:GO did too and that was only 2015 so maybe somewhere kids are still getting some exercise carrying their systems around.

Did you need a dedicated server to play Counter Strike over lan? I would think that that would be odd. I know people were playing that game online, which would make a server make more sense. It looks like maybe Doom needed one? It sounds like it didn't even need to be dedicated hardware though. I know I saw a few old 386 (or so) systems all playing together at the Midwest Gaming Classic in Milwaukee a year or two ago. We didn't have a computer that could run Counter Strike in those days. I just remember going over to a friend's house and him showing me some Half-Life mods, including that.

Still no cure for Coronavirus, the world has a chance :thumb:

I read at least a few days ago that some lab in California had already made a vaccine within 9 hour or so of receiving a sample. Last I saw, the numbers didn't look too terrible either as far as infections vs fatalities.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Charlesxvi on Tue, 18 February 2020, 11:59:20
I primarily game on my PC but I also do school work. I got my new one because it struggled with more modern games. That and my power supply died and it fried some components. I'm not 100% sure which parts are busted but I didn't need to test it really anyways. I got all new components and I couldn't be happier, I did have partial bad timing with my gpu and cpu though. I have an 8700k & 2070 but not long after the super cards came out and the 9700k dropped in price.

Sent from my Pixel 3 XL using Tapatalk

Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: absyrd on Tue, 18 February 2020, 12:54:05
I primarily game on my PC but I also do school work. I got my new one because it struggled with more modern games. That and my power supply died and it fried some components. I'm not 100% sure which parts are busted but I didn't need to test it really anyways. I got all new components and I couldn't be happier, I did have partial bad timing with my gpu and cpu though. I have an 8700k & 2070 but not long after the super cards came out and the 9700k dropped in price.

Sent from my Pixel 3 XL using Tapatalk



It is always "bad timing". The next thing is already here by the time **** ships to you. That is the tech game.

Anyhow, RMA'd by 5700xt. Constant crashing with numerous drivers, underclocking/volting, etc. **** that.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Maledicted on Tue, 18 February 2020, 13:25:06
I primarily game on my PC but I also do school work. I got my new one because it struggled with more modern games. That and my power supply died and it fried some components. I'm not 100% sure which parts are busted but I didn't need to test it really anyways. I got all new components and I couldn't be happier, I did have partial bad timing with my gpu and cpu though. I have an 8700k & 2070 but not long after the super cards came out and the 9700k dropped in price.

Sent from my Pixel 3 XL using Tapatalk




It is always "bad timing". The next thing is already here by the time **** ships to you. That is the tech game.

Anyhow, RMA'd by 5700xt. Constant crashing with numerous drivers, underclocking/volting, etc. **** that.

Sounds like good old ATI all over again.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Charlesxvi on Tue, 18 February 2020, 13:55:53
I primarily game on my PC but I also do school work. I got my new one because it struggled with more modern games. That and my power supply died and it fried some components. I'm not 100% sure which parts are busted but I didn't need to test it really anyways. I got all new components and I couldn't be happier, I did have partial bad timing with my gpu and cpu though. I have an 8700k & 2070 but not long after the super cards came out and the 9700k dropped in price.

Sent from my Pixel 3 XL using Tapatalk



Sounds like good old ATI all over again.

It is always "bad timing". The next thing is already here by the time **** ships to you. That is the tech game.

Anyhow, RMA'd by 5700xt. Constant crashing with numerous drivers, underclocking/volting, etc. **** that.
Yikes, that's gotta suck. With something that expensive and with how many issues people have been having they really should've just taken a little more time and done it right.

Sent from my Pixel 3 XL using Tapatalk

Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: absyrd on Tue, 18 February 2020, 15:26:22
Yeah. I'm back on my 570 8gb for now. I have a freesync monitor that isn't supported by nvidia (yet), so I'll just wait on a 5600 xt / 5700 / 5700 xt when the drivers get sorted, I guess.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: tp4tissue on Tue, 18 February 2020, 16:07:28
Probably that's what he's got, idk. I just say T1 as like a coverall for a dedicated optic line, like sayin I want a Coke when what I really want is a Fanta. I should probably stop doin that.

Pineapple fanta = Best soda.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: noisyturtle on Tue, 18 February 2020, 16:28:55
Probably that's what he's got, idk. I just say T1 as like a coverall for a dedicated optic line, like sayin I want a Coke when what I really want is a Fanta. I should probably stop doin that.

Pineapple fanta = Best soda.


true facts
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: fanpeople on Tue, 18 February 2020, 16:37:46
Absolutely not, pineapple Fanta = trash/10.

Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: tp4tissue on Tue, 18 February 2020, 16:54:42
Absolutely not, pineapple Fanta = trash/10.



/GASP....

which soda is best in straya ?
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: fohat.digs on Tue, 18 February 2020, 16:58:38
I did not know that Fanta was even still a thing. When I was a kid, in the 1950s-60s (yes) my mother loved grape Fanta, which, even then, was nauseating to me and I couldn't even stand the smell of it when I was in the car. On the rare occasion that Fanta was the only choice for a soft drink, I could manage to drink an orange Fanta, but that was at the bottom of the list of choices.

I haven't even looked at soft drinks in years, except on that very rare occasion when there are no alternatives, when I quickly find a Sprite or some other lemon-lime option or walk away thirsty. On the other hand, I do quite like soda water and quinine water, although I now generally avoid quinine water because of the sugar and the atrificially sweetened variant is horrendous.
 
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: fanpeople on Tue, 18 February 2020, 17:07:40
Absolutely not, pineapple Fanta = trash/10.



/GASP....

which soda is best in straya ?


XXXX
Emu Export
Great Northern
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: noisyturtle on Tue, 18 February 2020, 17:47:44
Absolutely not, pineapple Fanta = trash/10.



/GASP....

which soda is best in straya ?


XXXX
Emu Export
Great Northern

Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: fanpeople on Tue, 18 February 2020, 17:51:46
Holy ****, we really are a pack of uncultured ****s arn't we?
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Sintpinty on Wed, 19 February 2020, 05:41:20
I did not know that Fanta was even still a thing. When I was a kid, in the 1950s-60s (yes) my mother loved grape Fanta, which, even then, was nauseating to me and I couldn't even stand the smell of it when I was in the car. On the rare occasion that Fanta was the only choice for a soft drink, I could manage to drink an orange Fanta, but that was at the bottom of the list of choices.

I haven't even looked at soft drinks in years, except on that very rare occasion when there are no alternatives, when I quickly find a Sprite or some other lemon-lime option or walk away thirsty. On the other hand, I do quite like soda water and quinine water, although I now generally avoid quinine water because of the sugar and the atrificially sweetened variant is horrendous.
 
I think we should move all this soda talk to a different thread but whats your review of bübly soda
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Maledicted on Wed, 19 February 2020, 08:59:21
I think we should move all this soda talk to a different thread but whats your review of bübly soda

You mean the sparkling water? I don't know of any Bubly soda. Their watermelon is great, otherwise the rest could be better. Polar is probably the best overall brand in that market. Kroger, surprisingly, has really good dr. pepper and root beer knockoffs.

You should start a new thread on this in the Off Topic category.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Leslieann on Sat, 29 February 2020, 16:15:40
If you are planning an upgrade, DO IT NOW.

Parts are going to be short for months due to the virus going around.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: tp4tissue on Sat, 29 February 2020, 18:32:08
If you are planning an upgrade, DO IT NOW.

Parts are going to be short for months due to the virus going around.


most of those mobos are made in taiwan. buh.... i guess maybe some of the other stuff.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: Leslieann on Sat, 29 February 2020, 23:29:43
Besides the fact that China is not the only one being hit, only the worst and it doesn't matter if it gets made if you can't get it to your doorstep.

Not to mention we already have a GPU shortage and we were heading into an artificial shortage of ram and ssd drives.
Title: Re: How do you decide when it's time to build a new PC vs incrementally upgrading?
Post by: absyrd on Sun, 01 March 2020, 23:39:28
Yeah. GPUs are def low stock everywhere. And that is post-mining super craze, too.