I'm getting about +5% with Zen3

Interestingly, I can squeeze up to 64 AES rounds at the end of the loop before hashrate starts decreasing. The current draft only has 16 rounds. TBD if the impact on soft AES systems is acceptable.

64 rounds when running single threads, or all threads on all cores?

1 thread

Need to test 2 threads running on the same core - this will slow down AES rounds, but L3 latency will stay the same

I don't have enough L3 to test with all threads.

I don't think you can squeeze more than 32 rounds

I think 16 rounds are find, this already doubles the amount of AES per hash.

fine*

yes

and it is in line with how we already use AES (4 rounds for program buffer, for example)

you can set thread affinity and run 2 threads on the same core to test how many rounds is possible

the code is available, anyone can try it

16 -> 32 rounds, -0.7% hashrate (2 threads on one core)

32 -> 64 rounds, -3.25% (2 threads on one core)

so even 32 rounds is noticeable

and xor -> 16 rounds?

less than 1 h/s difference

so < 0.1%

command line was "--mine --jit --largePages --threads 2 --affinity 3 --init 16"

so 16 rounds seems to be a good choice

yes

0.7% drop with 32 rounds is a lot. I consider 0.1% speedup to be significant these days, when I optimize XMRig

but I'm testing on Ryzen 7 1700 (Zen 1)

this CPU has the fastest L3

Zen 3 / Zen 4 will have smaller drop on 32 rounds

still, 16 rounds should be optimal

because it will be unnoticeable on all CPUs

*with hardware AES

btw, to measure the CFROUND effect, you can change line 125 in common.hpp to using JitCompiler = JitCompilerX86<RANDOMX_FLAG_DEFAULT>; and rebuild

I measured about 5% with 1 thread on Ryzen 5850U

926.4 h/s -> 1004.2 h/s on Ryzen 7 1700 (2 threads on 1 core)

so 8% speedup

and it was 1004.2 -> 1003.8 h/s when I changed xor to 16 aes rounds

pretty significant

yes, all Zen CPUs implement mxcsr instructions in microcode

Intel is much faster with it

but this change shouldn't make Intel CPUs worse, it will just give smaller boost (0-1%)