This gentleman is associated with Native Instruments; he coded some of Kontakt 7 and has been prominent on the NI Forums for years. The comment I have quoted above (from VI-Control.net) was made in reference to questions concerning SATA SSD speeds and NI's Kontakt. I thought it interesting and informative enough to warrant posting here.Most any SSD will perform 4KQD1 test like 100x better than a rust drive, which should be plenty to draw over 1k stereo voices from one drive at any point in time... your CPU should end up being saturated sooner than SSD throughput. Even at QD1!
- EvilDragon
For a little more on the "4KQD1" term, this comes from a reddit discussion...
While our super-fast NVMe drives will certainly speed project loading, once a project is up and running, there are unlikely to be benefits while running our DAWs that result from that speed other than quicker saves and the ability to run Kontakt and other VIs at their minimum Sample Pre-Load values, thus saving RAM.QD means "Queue Depth". In practice, it means how many IO-operations the operating system is able to give to said device before waiting for a response to any of those operations.
Normal applications usually read something, wait for the result and then proceed to read something else after the first result has been processed. This is QD1 because only one read operation is executed in parallel and the results are needed before another request can be done. Pretty much all non-database applications belong to this class. (All video works usually belong to this class except that the IO size is much bigger/easier than 4K.)
A server system usually runs multiple QD1-processes in parallel and if the server is running say 32-64 parallel QD1-processes the overall load for the device could be around QD32.
If some SSD device spec simply says "Random 4K Read" or "Random 4K Write" you can be absolutely sure that it means QD32 (or even high queue depth!) because QD1 spec is always much lower and manufactures do not like low numbers in the spec sheets. (Some manufacturers also like to cheat by displaying "QD1 T32" numbers where the "T32" means 32 paraller threads were doing parallel QD1 read requests each. Logically this is same as QD32 but the actual behavior could be slightly worse or better than plain old QD32 depending on the operating system. If the spec sheet only mentions this result, you can be pretty sure that manufacturer found this method provides the most marketable spec values. If spec sheet doesn't mention QD or T values, you can assume that the manufacturer used whatever combination that resulted in best spec result and you cannot compare that to any other similarly underspecified spec result from another device.)
What you really want is QD1 T1 random read 4kB measurement. That's the only number that really matters for most generic use and that guarantees the minimum performance you're going to get