Hi,
Samsung SSD's have the same feature.
I noticed, yep, that some manufacturers add memory reorganizing features to give speed to the end-user. Unfortunately, not every manufacturer is making those options available. Don't know why - buying cheaper parts? Don't know how to write software/program memory allocation?
Nevertheless, a lesson, I've learned when choosing cheap class NVMe, is: Always check those "momentum cache", or whatever manufacturers fancy to name them.
And it's vital to understand that some manufacturers add RAM cache options, other add/replace that with changing QLC memory into SLC. AS much as I understand, SLC memory in those QLC SSD's is much faster. I think in P1, Crucial is reorganizing up to 4GB QLC into the SLC memory cluster. It's a very good solution, as average Joe is working with small files anyway and won't notice speed drops. Unless he's copying files, larger than 4GB. But then, even with falling to "slow" 2GB/s it's quite a quick process to copy files. And if we consider everyday practice, when those files are usually, copied to much slower drives, it doesn't make any sense to get expensive NVMe.
As an owner of P1 NVMe drive, I have to tell, that there is one particular problem with Momentum Cache on Crucial drives. It doesn't survive Windows update (distro update). Windows won't give a crap of Crucial's Drivers and will use its own ones instead. Thus, momentum cache will not be working at all. The only solution, for now, is to wipe windows and install as a fresh one. I don't think Crucial will fix it, as it's their lowest price consumer product, therefore it's unreasonable to spend money on the fairly minor public problem. (Not many users are even aware of those drive features.)
I guess, when Win files system will be made particularly with NVMe in mind, then we will see a great benefit from those SSDs and NVMe, as they are able to access really deep queues.I'd like to see PS5-like hardware to specifically process file reading/writing.
I think it's important to add those real-life use-cases to tests, as end-user will see how fast system he will get, when installing drivers and turning all speed features. Specifically, I mean when drives memory is changed. Idon't think that testing raw performance of video card without installing appropriate drivers is great idea in the first place. As we all know, using a video card without drivers will give a crappy performance in the first place.
Notebook integrated cards are borrowing RAM as their own RAM to work and we see a lot of tests done. Nobody is crying that integrated cards shall not be using the laptop's RAM to get the advantage of speed. Or that is "unfair and doesn't represent actual video card speed".
I think it will be fair to add to existing test also one, with all features enabled. Yes, I completely understand that those speeds will be similar and high. But that what end user will actually work with! And ones, like the user TheLostSwede, can always turn off all features and uninstall drivers (not install them in the first place) to get "Real" performance.
