When they won't cost any different... why argue against it though? Also, crystalmark is a really crappy benchmark which none of the serious ssd reviewers even use anymore. Pretty much any other benchmark shows results starting at 90MB/s with nvme drives on 4K random starting at QD2 depending on the drive for the consumer tier stuff with some able to do as much as 200MB/s. Enterprise stuff needs a larger QD but can hit up into the 700MB/s plus. On sequential we're looking at 1GB/s speeds.
That intel drive result you selected is also one of the worst performers available. But I can only assume you did that intentionally for your argument.
Do most people need that kind of performance right now? Surely not... but if it's not going to cost much more than current sata ahci drives there is no reason not to step up. It's going to be the standard sooner than later.
Yeah intel 9 series has nvme natively with m.2 support, though not all boards implemented it fully. Otherwise you need a pci-e ssd (or pci-e m.2 adapter card) and UEFI motherboard. Many of them don't work properly as boot drive on old bios systems, they don't even show up in the boot selection.
It most definitely WILL COST MORE..
If it didn't cost more.. you'd not see any argument from me..
How does upgrading a platform not cost money
How does buying new drives not cost money
Does that New drive translate to improved performance FOR regular people vs their existing SSD drive or cheaper drives without NVME?
Drive speeds at low queue depths is not bottlenecked by ahci..
Drive speeds at high queue depths is..
Since we don't actually use high queue depth in the desktop environments..
ANYONE with current platforms should NOT upgrade FOR THAT FEATURE alone..
I don't disagree that it's an uptick of a feature.. my point is that, that-uptick, is not something which Applies to ANYONE outside of Enterprise.