4KB RANDOM READ/WRITE
In order to properly test random 4KB performance, we first secure erase to get it in a clean state. Next, the drive is filled by sequentially writing to the RAW NAND capacity twice. We then precondition the drive with 4KB random writes at QD256 until the drive is in a steady state. Finally, we cycle through QD1-256 for 5 minutes each for writes and then reads. All this is scripted to run with no breaks in between. The last hour of preconditioning, the average IOPS, and average latency for each QD is graphed below.
Our first look at 4K performance once again validates the high maximum latency results we saw in our SNIA testing. Our performance data points are scattered around 80K IOPS in each direction from the average of about 180K IOPS. Latency can be as high as 10ms with an average of just under 2ms.
During 4K reads the DCP1000 trades blows with the HGST SN100 we have tested in the past. At QD 1 and two the DCP1000 has the advantage, but up until QD 32, the SN100 takes the lead by a small margin. Beyond that, however, it just completely dominates the test pool by delivering over 850K IOPS at QD 128, which is half of what the Intel DC P3608 needs and delivers 1.1 million IOPS at QD256!
Latency ranges from 0.106ms at QD 1-16 then increases to 0.111ms at QD 32, 0.124ms at QD 64, 0.149ms at QD 128, and maxes out at 0.235ms at QD256.
While the Kingston DCP1000 completely dominated in 4K reads against our comparison drives, it wasn’t able to beat out the write oriented and larger capacity Micron 9100 MAX in 4K writes.
From QD 1-4 the DCP1000 shows to deliver higher IOPS than the rest of the pool, but beyond that, the Micron takes the cake rightfully.