KitGuru KitGuru.net – Tech News | Hardware News | Hardware Reviews | IOS | Mobile | Gaming | Graphics Cards
CrystalDiskMark is a useful benchmark to measure theoretical performance levels of hard drives and SSD’s. We are using v7.0 to test NVMe drives.
Looking at the two CrystalDiskMark result screens it appears that the new E18 controller is much more efficient when reading compressible 4K data at certain queue depths. We can also confirm the official Sequential read/write performance of 7,000MB/s and 5,300MB/s respectively using CrystalDiskMark.
CrystalDiskMark, version 7, includes a couple of profiles that can be used for testing – Peak Performance and Real World. The result screens for these two profiles not only display MB/s results but also IOPS and latency.
Looking at the results for Sequential read/write performance we could confirm both official Sequential performance figures. The drive is rated up to 7,000MB/s for reads, with the tested drive producing a best of 7,401MB/s and up to 5,300MB/s for writes, with the best test result of 5,303MB/s.
When it comes to random performance the drive is rated as up to 350,000 IOPS for reads and up to 700,000 IOPS for writes. As you can see from the Peak Performance results we could confirm the random read figure with the drive producing a test result figure of 362, 551 IOPS but sadly the best write test result was way of the official maximum figure at 343,685 IOPS.
We also used CrystalDiskMark 7 to test the random performance of the drive at lower queue depths (QD1 – QD8) where most of the everyday workloads occur) using 1 to 4 threads. The Sabrent Rocket 4 Plus read performance climbs smoothly throughout the tested queue depths and threads with a peak performance figure of 265,073 IOPS (1,085MB/s) at QD8.
In the write tests, the best performance was a pretty close run thing between using 3 threads (351,320 IOPS) and 4 threads (354,528 IOPS) at QD8. Whereas the four threaded test climbed relatively smoothly through the queue depths, the three threaded test plateaued out between QD2 and QD4 before accelerating between QD4 and QD8.
