Recently I was running some benchmarks to double-check that the performance of AMD EPYCs servers delivered by Company A, match the performance of those delivered by Company B. I was using our benchmark automation tool (about which I may write a bit more in the future). It is currently based on sysbench to hammer MySQL instances. Surprisingly, the results have shown very significant difference: B servers were on average 83% slower…
Skip to content
Change is the only constant... 