[SOLVED] Samsung m.2 SATA vs NVMe Benchmark Results

Apr 6, 2020
2
0
10
Hi All

Does this look right?

SSD 850 EVO M.2 500GB SATA Benchmark

Sequential (MB/s)
Read = 5,868
Write = 5,405

Random (IOPS)
Read = 150,146
Write = 85,205



SSD 970 EVO 1TB NVMe Benchmark

Sequential (MB/s)
Read =3,543
Write = 2,501

Random (IOPS)
Read = 348,876
Write = 334,960

I have Rapid Mode turned on for the SATA drive but Samsung Magician tells me this option is not available fro the NVMe drive.

So the Random (IOPS) on the NVMe drive are clearly higher than the SATA drive, but the Sequential Read and Write speeds are better on the SATA drive.

Is this expected?
 
Solution
RAPID mode is using part of your RAM as a cache for the drive.
That number you're reading is basically data coming from the RAM, not the drive.

In actual use, I personally saw zero benefit. So I turned it off.

For a SATA drive, Sequential numbers in the range of 520-550 are typical and expected.
Apr 6, 2020
2
0
10
It appears you have RAPID mode enabled for the 850 EVO. That is giving artificially high numbers.
Turn it off.

I only just turned it on after running the benchmark with it off and seeing how slow the SATA is compared to NVMe :D

I guess my question is then why in rapid mode can the SATA apparently out perform, the NVMe?
 

USAFRet

Titan
Moderator
RAPID mode is using part of your RAM as a cache for the drive.
That number you're reading is basically data coming from the RAM, not the drive.

In actual use, I personally saw zero benefit. So I turned it off.

For a SATA drive, Sequential numbers in the range of 520-550 are typical and expected.
 
Solution

wcndave

Reputable
Mar 16, 2016
12
0
4,510
I think you can only have one drive in rapid mode, and maybe it has to be the first drive? Also, NVMe is so fast, that rapid mode is not needed, and just uses your RAM. So, your NVMe is faster, and your SATA can artificially boost its speed using RAM.