Performance Comparisons – Client QD Weighted
These results attempt to simplify things by focusing on what really matters – the Queue Depths that folks actually see when using these products. A dimension is eliminated from the previous charts by applying a weighted average to those results. The weights were derived from trace recordings of moderate to heavy workloads, which still ended up running closer to QD=1-2 even on a slower SATA SSD. The intent here is to distill the results into something for those wanting 'just the facts' to grab and go when making their purchasing decisions. Don't be alarmed by the low figures. Remember, these are low queue depths – the place where these SSDs actually operate when in use by those not just running benchmarks all day!
Since client workloads lean heavily towards reads, keep a focus on the blue bars above. The random read results for most NVMe SSDs results in a relatively homogenous grouping around the 20k mark. Note that a good SATA SSD (860 EVO – bottom) comes rather close as well since even a PCIe NVMe interface can't magically make flash memory respond faster than is physically possible.
Shifting to random writes (orange bars), we note that the SX8200 performs admirably. Also, note how very similar the SX8200 results are to the Intel SSD 760p (similar hardware).
That excellent sequential read result we saw a few pages back now shows its real-world payoff, as the SX8200 manages to beat the 970 EVO! Not so much for write speeds, but that was also to be expected. I'm not sure what tuning enabled ADATA to do so much better than the 760p here, but I'm sure not going to complain about it!
This % read sweep chart shows the random QD weighted results between full reads and full writes. We look for a smooth curve here and all SSDs behave as expected with no real surprises.