Question I have two identical Hynix M2 drives in my PC but they benchmark at different speeds ?

[mrjenkins44]

I have two SK hynix Platinum P41 2TB PCIe NVMe Gen4 in a Gigabyte Aorus Elite x670 AX motherboard. Both are in M2 slots on the mobo.

There are no other drives.

For some reason, the one with Windows on it clocks in at 2,637 MB/s and the 2nd one at 3,500 MBs. I am using UserBenchmark.

They both have about 1.3 TB free space.

Is there any particular reason once is scoring lower than the other? The slower one reads at 456% to 557% of the faster one. It says it is performing below expectations wile the other is performing as expected.

I tried the test multiple times, on different days, and after rebooting, and it keeps coming back the same.

 

[USAFRet]

Don't use UBM for benchmarking drives.

Show us a screencap of CrystalDiskMark for each drive.

 

[USAFRet]

Also, a drive with the OS on it might benchmark a bit slower than a similar drive with no OS on it.
The system is also running the OS at the same time as the bench.

 

[Misgar]

If I've got hold of the correct user guide, the block diagram for the motherboard shows four M.2 drives.
M2A_CPU and M2B_CPU are driven direct from the AM5 CPU.
M2C_SB is driven from the first chipset
M2D_SB is driven from the second cascaded chipset.

I would expect there to be speed differences across the four M.2 sockets. Which sockets were you using?
At a guess, I'd say the CPU sockets would be faster than the chipset sockets.

 

[Bob.B]

I have two SK hynix Platinum P41 2TB PCIe NVMe Gen4 in a Gigabyte Aorus Elite x670 AX motherboard. Both are in M2 slots on the mobo.

There are no other drives.

For some reason, the one with Windows on it clocks in at 2,637 MB/s and the 2nd one at 3,500 MBs. I am using UserBenchmark.

They both have about 1.3 TB free space.

Is there any particular reason once is scoring lower than the other? The slower one reads at 456% to 557% of the faster one. It says it is performing below expectations wile the other is performing as expected.

I tried the test multiple times, on different days, and after rebooting, and it keeps coming back the same.

Swap slots see if it makes a diff.

 

[mrjenkins44]

So I have them in slots:

M2A
M2B

And here is the CrystalMark scores side by side:

 

[Misgar]

As a final test before returning the slow drive for replacement, you could try swapping the drives over between M2A and M2B, just to prove it's not socket related.

 

[SkyNetRising]

So I have them in slots:
M2A
M2B
And here is the CrystalMark scores side by side:

Windows reads/writes to OS drive all the time.
Benchmark results on OS drive will be lower.

 

[mrjenkins44]

Windows reads/writes to OS drive all the time.
Benchmark results on OS drive will be lower.

It would be THAT much slower?

 

[SkyNetRising]

It would be THAT much slower?

Yes. Easy.

To benchmark it properly, you'd have to install windows on another drive and
run benchmark from there.