CrystalDiskMark Random 4K Q1T1 write I get:
2TB Samsung 980 Pro in 1st M.2 slot: 251 MB/s
2TB Samsung 980 Pro in 2nd M.2 slot: 150 MB/s
My motherboard is an MSI Tomahawk X670E with two 2TB Samsung 980 Pros installed. Now the question I have is, why is the 2nd M.2 slot in my motherboard much slower at the 4K Q1T1 performance in CrystalDiskMark even though they both are operating at PCI-E 4.0x4 mode? I can understand if it would be about 5-10% slower for that category because the chipset lanes have higher latency than the CPU lanes but 40% slower, I don't get it. All other categories in CrystalDiskMark are within 5% difference in performance between both these M.2 slots. I have installed the latest chipset drivers downloaded from MSI for my specific motherboard which are newer than the ones directly from AMD.
2TB Samsung 980 Pro in 1st M.2 slot: 251 MB/s
2TB Samsung 980 Pro in 2nd M.2 slot: 150 MB/s
My motherboard is an MSI Tomahawk X670E with two 2TB Samsung 980 Pros installed. Now the question I have is, why is the 2nd M.2 slot in my motherboard much slower at the 4K Q1T1 performance in CrystalDiskMark even though they both are operating at PCI-E 4.0x4 mode? I can understand if it would be about 5-10% slower for that category because the chipset lanes have higher latency than the CPU lanes but 40% slower, I don't get it. All other categories in CrystalDiskMark are within 5% difference in performance between both these M.2 slots. I have installed the latest chipset drivers downloaded from MSI for my specific motherboard which are newer than the ones directly from AMD.
Last edited: