Reduced NVME drive performance following OS re-imaging


Limp Gawd
Nov 30, 2018
So, I have a dell XPS 9510 laptop (i7-11800H, 32GB DDR4-3200, RTX 3050, Win10 Enterprise, for reference). From the factory it came with a 512GB Micron 2300 PCIe 3.0x4 SSD. However, the laptop actually supports PCIe 4.0, so I got a Sabrent Rocket 4 Plus 1TB to replace the OEM Micron drive.

I had re-imaged the 9510 from scratch using a flash drive and manually installing all the Dell drivers from the XPS support page prior to the drive swap, so it wasn't running the factory image. Initially, I used the Sabrent Acronis software to simply clone the Micron over to the Sabrent. However, I have a, perhaps unfounded, skepticism of exactly what goes on in a clone operation and figured it wouldn't hurt to start from scratch, and perhaps it would even squeeze a little more performance out of the drive to do a fresh install on it.

Anyway, following the re-imaging of the Sabrent R4P, Crystal DiskMark actually shows significantly worse numbers on a number of benchmarks than it did when it was cloned from the Micron (though some are comparable, and one is actually higher). I am trying to come up with an explanation or troubleshooting process, but I haven't been able to diagnose it. My first thought is drivers, but as far as I know I followed the same process for both imaging events, so the drivers should be the same... but I don't have the old image, so I can't do a line by line check. Still, I had the same list of Dell drivers, so I have a hard time imagining there could be a difference. The only real potential reason for a difference would the drivers that Windows Update pulls down once I give it an internet connection, if any.

Micron 2300 Benchmarks:


Sabrent R4P after Acronis Cloning:


Sabrent R4P after re-imaging:

So, my write speeds on the SEQ tests are consistently slower by up to 12%, and then my RND4K Q32T16 reads are less than 50% of the clone numbers (not even matching the original Micron numbers), while the writes are around 2/3. These numbers are all consistent, so it's not a fluke. I've tried rebooting and ensuring there's not activity in the background. Besides, the SEQ read numbers are not showing any issues. As far as I know, nothing has changed. I didn't mess with the BIOS or anything like that. I just have no good explanation, and it's bugging the crap out of me. If anyone can provide any insight, I'd be very appreciative. Thanks!
Can you tell any difference?

Nope, didn't think so.

Some days you get good benches and some days you get bad.
Can you tell any difference?

Nope, didn't think so.

Some days you get good benches and some days you get bad.
Not by that much. Some thing's obviously wrong but I don't have suggestions on how to fix it.
are you reimaging, cloning or reinstalling? they are not the same. maybe theres a formatting difference between the cloned oem drive and whatever you are meaning by "re-imaged".
Benchmarking an empty (minimal garbage collection going on in the background) SSD which has no other processes in contention will often bench a few percentage points off per run. Doing the test on your active OS drive can yield wildly different numbers.
Try AS-SSD, it will show if your alignment is bad. That will cause lower benchmarks.