Recently we discovered that a slew of Silicon Power NVMe drives we had, all with the same P/N, being the P34A80 drives had various arrangements of controllers and NAND.
So far we have discovered at least three controllers on this same P/N Model. Those being:
We found that the Silicon Motion 2262 was an adequately performing controller but the other two are terribly under powered. The SM 2263XT is only a 4 channel controller whereas the other two are 8 channel. The Phison E12 is about 1/2 the speed of the SM2262.
SM2262:
sequential-fill: (groupid=0, jobs=1): err= 0: pid=3733: Tue Jul 12 16:32:25 2022
Description : [Sequential fill phase]
write: IOPS=1039, BW=1039MiB/s (1089MB/s)(609GiB/600013msec); 0 zone resets
random-write-steady: (groupid=1, jobs=4): err= 0: pid=4163: Tue Jul 12 16:32:25 2022
Description : [Random write steady state phase]
write: IOPS=198k, BW=774MiB/s (812MB/s)(454GiB/600001msec); 0 zone resets
Phison E12:
sequential-fill: (groupid=0, jobs=1): err= 0: pid=3775: Tue Jul 12 16:32:29 2022
Description : [Sequential fill phase]
write: IOPS=762, BW=762MiB/s (799MB/s)(447GiB/600040msec); 0 zone resets
random-write-steady: (groupid=1, jobs=4): err= 0: pid=4168: Tue Jul 12 16:32:29 2022
Description : [Random write steady state phase]
write: IOPS=55.9k, BW=219MiB/s (229MB/s)(128GiB/600002msec); 0 zone resets