Specifications:
The specs list a 5,000 IOPS rating, which seems way low when compared to an Intel X25-M G2's rating of
35,000 IOPS at the same 4KB random reads. I was a bit skeptical at their stated spec being as low as it was, so I did a few quick checks before the suite:
IOMeter shows slightly under 5,000 IOPS for 4KB random reads.
ATTO shows 76 MB/sec for 4K random, which works out to slightly over 4,000 IOPS. The fact that we're still at the same 5,000ish at ATTO's default Queue Depth of 4 suggests this drive might not scale with higher Queue Depths like Intel, Sandforce, and Marvell is capable of. Before we got into the benches, the Silicon Systems controller wasn't looking to be as spectacular as I'd hoped. As a matter of fact, this ATTO run looks more like lesser performing controllers such as Samsung.
An ATTO run of a Samsung-controlled OCZ Summit for comparison.
Well, we're not sure what to expect, so lets get through the full suite and see what happens.