Our test of random write burst performance is structured similarly to the random read burst test, but each burst is only 4MB and the total test length is 128MB. Looking at the 1TB 860 QVO's random read results compared to all the other SATA drives in the benchmark database, it is clear that the QVO isn't anywhere near the cutting edge for power efficiency or peak performance, but there are worse drives out there. However, the vertical scale is important: the QVO doesn't ever reach even half the performance of the best TLC-based SATA SSDs. The queue depth scaling of the 860 QVO during random reads is fairly typical in shape, with improvements starting to taper off after QD16. The power consumption of the 860 QVO during the random read test is only slightly higher than its TLC-based relatives, but that's plenty to push its efficiency scores into last place, given the poor performance. On the longer random read test, the Toshiba TR200 DRAMless TLC drive is no longer able to stay ahead of the 860 QVO, and even the Intel/Micron QLC drives fall behind most mainstream SATA drives (especially when full).
The individual read operations are again 4kB, and cover a 64GB span of the drive. After each queue depth is tested, the drive is given up to one minute to cool off so that the higher queue depths are unlikely to be affected by accumulated heat build-up.
Each queue depth is tested for one minute or 32GB of data transferred, whichever is shorter. Our sustained random read performance is similar to the random read test from our 2015 test suite: queue depths from 1 to 32 are tested, and the average performance and power efficiency across QD1, QD2 and QD4 are reported as the primary scores. Even in the worst case of the smallest QVO being entirely full, read speeds are still vastly better than a hard drive. The burst random read performance of the Samsung 860 QVO is clearly lower than the 3D TLC competition, while the Intel/Micron QLC NVMe drives have no trouble competing against the field of mainstream SATA TLC drives. Each burst consists of a total of 32MB of 4kB random reads, from a 16GB span of the disk. The drives are given enough idle time between bursts to yield an overall duty cycle of 20%, so thermal throttling is impossible. Our first test of random read performance uses very short bursts of operations issued one at a time with no queuing.
0 kommentar(er)
