икониThe day job uses lots of SSDs as well as disks in various of our products. We rely upon internal testing and external benchmarks (which tend to be poor at best, but a very rough zeroth order test) to select them.
We had a pair of Intel 510 SSD units in for a customer, and they performed … just meh … not all that exceptional. Better than our OS drives, but not as good as the higher end SSDs. Of which they are supposed to be some.
I’ve got to look around for firmware for them, see if there is an update from Intel. Basically I think many of the SSD vendors may be building units which are really well optimized for the benchmark programs, but maybe don’t perform well in the real world. Way way back when, in the early days of the Intel compiler, it was taught to recognize SPEC benchmark and other benchmark code, and emit hand written/optimized assembler, but only when run on an Intel chip. This sort of benchmark gaming was common in those days, and frustrated many users and vendors.
I believe we may be seeing it again, though I don’t have hard evidence of it. Just a hunch. And a whole mess of 16x SSDs that don’t seem to come close to what their external benchmark tests suggest they should. Which suggests either bad batches, bad firmware … or malicious benchmarketing.
Yeah, I know the maxim: “Never attribute to malice that which might otherwise be explained by ignorance.” And that might apply here. If it weren’t for the profit motive, I could cut some slack on this. Not ready to call out anyone on this, just pointing out that this is an annoying possibility I can’t easily discount at this time.
And until then, we aren’t buying any more Intel 510 SSDs for testing/use. The real world performance is about equivalent to a reasonably fast hard disk. Which sort of destroys the utility of getting an SSD …