Fresh Win 10 install on MSI x570 Ace with 3700x, 2 identical new Adata SX8200PNP 2TB, only differences are: OS is installed on C; C is connected to M.2_1 (assuming this is direct to CPU), D is on M.2_3 (assuming chipset); C is 3% full, D is 12% full.
Tested with Crystal (and got similar results in another program)--I expected a very slight extra latency on D, but nothing like these performance differences. Note: C maxed at 43C, D at 42C, so temperature isn't the issue:
C:
[Read]
Sequential 1MiB (Q= 8, T= 1): 3274.791 MB/s [ 3123.1 IOPS] < 2560.34 us>
Sequential 1MiB (Q= 1, T= 1): 2794.642 MB/s [ 2665.2 IOPS] < 374.97 us>
Random 4KiB (Q= 32, T=16): 1484.068 MB/s [ 362321.3 IOPS] < 1411.95 us>
Random 4KiB (Q= 1, T= 1): 71.817 MB/s [ 17533.4 IOPS] < 56.90 us>
[Write]
Sequential 1MiB (Q= 8, T= 1): 3318.346 MB/s [ 3164.6 IOPS] < 2524.11 us>
Sequential 1MiB (Q= 1, T= 1): 3008.741 MB/s [ 2869.4 IOPS] < 348.27 us>
Random 4KiB (Q= 32, T=16): 1451.095 MB/s [ 354271.2 IOPS] < 1444.38 us>
Random 4KiB (Q= 1, T= 1): 226.385 MB/s [ 55269.8 IOPS] < 17.98 us>
D:
[Read]
Sequential 1MiB (Q= 8, T= 1): 2715.890 MB/s [ 2590.1 IOPS] < 3087.00 us>
Sequential 1MiB (Q= 1, T= 1): 2422.096 MB/s [ 2309.9 IOPS] < 432.68 us>
Random 4KiB (Q= 32, T=16): 1441.400 MB/s [ 351904.3 IOPS] < 1453.53 us>
Random 4KiB (Q= 1, T= 1): 69.483 MB/s [ 16963.6 IOPS] < 58.84 us>
[Write]
Sequential 1MiB (Q= 8, T= 1): 2466.232 MB/s [ 2352.0 IOPS] < 3391.31 us>
Sequential 1MiB (Q= 1, T= 1): 2294.523 MB/s [ 2188.2 IOPS] < 456.67 us>
Random 4KiB (Q= 32, T=16): 1041.318 MB/s [ 254228.0 IOPS] < 2012.52 us>
Random 4KiB (Q= 1, T= 1): 199.663 MB/s [ 48745.8 IOPS] < 20.40 us>
I haven't been able to find any CPU vs Chipset benchmarks with Google, have others experienced the above with x570?
Tested with Crystal (and got similar results in another program)--I expected a very slight extra latency on D, but nothing like these performance differences. Note: C maxed at 43C, D at 42C, so temperature isn't the issue:
C:
[Read]
Sequential 1MiB (Q= 8, T= 1): 3274.791 MB/s [ 3123.1 IOPS] < 2560.34 us>
Sequential 1MiB (Q= 1, T= 1): 2794.642 MB/s [ 2665.2 IOPS] < 374.97 us>
Random 4KiB (Q= 32, T=16): 1484.068 MB/s [ 362321.3 IOPS] < 1411.95 us>
Random 4KiB (Q= 1, T= 1): 71.817 MB/s [ 17533.4 IOPS] < 56.90 us>
[Write]
Sequential 1MiB (Q= 8, T= 1): 3318.346 MB/s [ 3164.6 IOPS] < 2524.11 us>
Sequential 1MiB (Q= 1, T= 1): 3008.741 MB/s [ 2869.4 IOPS] < 348.27 us>
Random 4KiB (Q= 32, T=16): 1451.095 MB/s [ 354271.2 IOPS] < 1444.38 us>
Random 4KiB (Q= 1, T= 1): 226.385 MB/s [ 55269.8 IOPS] < 17.98 us>
D:
[Read]
Sequential 1MiB (Q= 8, T= 1): 2715.890 MB/s [ 2590.1 IOPS] < 3087.00 us>
Sequential 1MiB (Q= 1, T= 1): 2422.096 MB/s [ 2309.9 IOPS] < 432.68 us>
Random 4KiB (Q= 32, T=16): 1441.400 MB/s [ 351904.3 IOPS] < 1453.53 us>
Random 4KiB (Q= 1, T= 1): 69.483 MB/s [ 16963.6 IOPS] < 58.84 us>
[Write]
Sequential 1MiB (Q= 8, T= 1): 2466.232 MB/s [ 2352.0 IOPS] < 3391.31 us>
Sequential 1MiB (Q= 1, T= 1): 2294.523 MB/s [ 2188.2 IOPS] < 456.67 us>
Random 4KiB (Q= 32, T=16): 1041.318 MB/s [ 254228.0 IOPS] < 2012.52 us>
Random 4KiB (Q= 1, T= 1): 199.663 MB/s [ 48745.8 IOPS] < 20.40 us>
I haven't been able to find any CPU vs Chipset benchmarks with Google, have others experienced the above with x570?