Agner's CPU blog

Call [m] is similar to mov r, [m] call r It is using the address generation units (p237) twice, once for reading m and once for storing the return address on the stack. call instructions and return instructions are usually tested together to avoid messing up the return stack buffer. It may be diffi...

This Reddit post is reporting experiments with how to make sure heavy tasks are running in the P cores https://www.reddit.com/r/XMG_gg/comments/vlqn6d/psa_rendering_tasks_are_moved_to_ecores_when/ . I still think it is unreasonable to expect ordinary computer users to attend to processor-specific pe...

The µop cache can hold 4k µops. The µop cache can deliver a maximum of 6 µops per clock cycle for a single thread

This agrees with my measurements.

We have now made some more tests on the P core after fixing the problem with overheating the CPU. The results are more stable now and basically confirming what Andreas wrote: The decoders can handle up to 6 µops per clock Simple integer instructions have a maximum throughput of 5 instructions per cl...

Andreas wrote:
This can be seen from the UOPS_MITE count that is shown in the output

With 4000 8-byte NOPS I see 32 bytes per clock. With 8000 8-byte NOPS I see 16 bytes per clock.
MITE_UOPS keeps counting (event 0x79, umask=4). DSB_UOPS (umask=8) stop counting.

Thank you for contributing with your measurements. Andreas wrote: The decoders on the P cores can decode 6 instructions per cycle. ..they can handle 32 bytes per clock This is when your code is running out of the µop cache. The µops have already been decoded. The decoder throughput can only be measu...

I have now been able to test the Alder Lake P cores with full access to the AVX512 instructions thanks to a lot of help from Zingaburga . I can confirm that the full AVX512 instruction set is working on the Alder Lake with an early microcode and a certain BIOS when only the P cores are enabled. The ...

I will propose that you post your question on https://stackoverflow.com with details about your code. Your results may in fact reflect cache effects, compiler optimization, or thread effects.

davidbak wrote:

Isn't there also software that would have been pessimized in the other direction

Yes

Thank you for your proposal. The easiest way is if I can test it by remote access (Linux). Please send me an email agner_at-agner.org