Benchmarking The Updated Intel CPU Microcode For SRBDS / CrossTalk Mitigation

With the Skylake Xeon E3 system, the outcome was similar. RdRand was at less than 3% the original performance...

But in the real-world workloads, across dozens of different benchmarks the results were flat overall. It was just in a few cases that this updated Intel CPU microcode was hurting the performance.

When it came to evaluating different desktop workloads on the Core i5 8400, the RdRand performance was impacted as expected but given that RdRand / RDSEED is not used prolifically especially by desktop workloads, for the other benchmarks the performance was basically unchanged.

That's the metrics so far over benchmarking CrossTalk / Special Register Buffer Data Sampling over the past day. The Phoronix Test Suite, of course, has hundreds of test profiles available and I am still running additional workloads on the Xeon boxes in looking for any other notable real-world performance changes but at least as of now most workloads are unaffected with just some server tests showing small hits.