#
✕ nick:kdrag0n[m]
#asahi · 2021-10-03 23:08
kdrag0n[m]btw is tracing pmgr supposed to work? macOS panics within a few seconds every time I add the tracer
#asahi · 2021-10-03 22:56
kdrag0n[m]new commit based on the new data that includes 3.2 GHz (very little difference): https://github.com/kdrag0n/linux-m1/commit/05c296604a42
#asahi · 2021-10-03 22:53
kdrag0n[m]IPC values look reasonable compared to Snapdragon 888: 10.9 C/MHz (Firestorm) vs 9.3 (Cortex-X1), 7.6 (Icestorm) vs 3.7 (A55) which seems reaosnable considering how old the A55 is
#asahi · 2021-10-03 22:51
kdrag0n[m]and if that reference is right, then the e-cluster should be too
#asahi · 2021-10-03 22:51
kdrag0n[m]sounds about right given what people have said about the M1
#asahi · 2021-10-03 22:50
kdrag0n[m]I ran the benchmarks again with only one active P-core (thanks for the tip) and according to CoreMark, 3.2 GHz is slightly faster than my Zen 2 desktop at max freq
#asahi · 2021-10-03 22:49
kdrag0n[m]marcan: I looked into the capacity numbers a bit more, and it seems more like Dhrystone is the one that's not very realistic
#asahi · 2021-10-03 05:24
kdrag0n[m]not sure how to deal with that from a scheduler standpoint
#asahi · 2021-10-03 05:24
kdrag0n[m]I don't think the scheduler checks it, but it's a semantic difference at least
#asahi · 2021-10-03 05:23
kdrag0n[m]those freqs should probably be marked with the boost flag then
#asahi · 2021-10-03 05:23
kdrag0n[m]marcan: ah that explains the freq behavior
#asahi · 2021-10-03 03:37
kdrag0n[m]not sure how to pin freqs on macOS for testing, but those freqs will skew scheduler capacities if left that way
#asahi · 2021-10-03 03:35
kdrag0n[m]it also looks your cpufreq driver has the same issue as Corellium's, where freqs > 2988 MHz are identical to 2988 in terms of performance: https://github.com/kdrag0n/freqbench/blob/master/results/t8103/asahi-cpufreq/run.log#L46-L49
#asahi · 2021-10-03 03:30
kdrag0n[m]explanation is in the commit message
#asahi · 2021-10-03 03:30
kdrag0n[m]I'm getting ~713 with both your new cpufreq driver and Corellium's: https://github.com/kdrag0n/linux-m1/commit/83b4bfbb156929c89fe1212b98bff27b74233667
#asahi · 2021-10-03 03:29
kdrag0n[m]marcan: how'd you get capacity-dmips-mhz 326 for the little cluster?
#asahi · 2021-09-29 18:04
kdrag0n[m]j_ey: ah ok
#asahi · 2021-09-29 17:35
kdrag0n[m]asking because I ran some benchmarks for capacity-aware scheduling, but frequencies above 2.988 GHz don't seem to do anything with Corellium's cpufreq driver
#asahi · 2021-09-29 17:35
kdrag0n[m]I saw pmgr but nothing related to cpufreq
#asahi · 2021-09-29 17:34
kdrag0n[m]hi, the drivers mentioned in https://twitter.com/marcan42/status/1442871737419321347?s=19 haven't been pushed anywhere yet, right?