Synthetic Storage Benchmarks
PCMark 10
We are using PCMark 10’s Full System Drive Storage Benchmark. It outputs an overall score (higher is better) that is derived from the Bandwidth (higher is better) as well as access time (lower is better) results.
Right off the bat, we find there to be a loss in performance running the exact same SSD off the secondary M.2 socket versus the primary socket, despite both being PCIe 4.0 x4 slots. The drop in PCMark10 is 3%, for a test that is very consistent, this is a decent amount.
The PCMark 10 bandwidth test continues the trend, with the secondary M.2 socket being 3% slower as well in the bandwidth test.
Even the access time is slower on the secondary M.2 socket at 70 versus 68, certainly pointing toward latency increasing.
PassMark PerformanceTEST
We are using PassMark’s PerformanceTEST Disk Mark benchmark only. This test benchmarks Disk Sequential Read, Disk Sequential Write, IOPS 32KQD20, IOPS 4KQD1 tests and outputs an overall score (higher is better.)
The PassMark PerformanceTEST Disk Mark benchmark continues the trend, though it isn’t as extreme here, with less than 1% difference.