tag:blogger.com,1999:blog-2963790426029213933.post1682635864127135261..comments2024-02-29T00:33:07.382-08:00Comments on Alfred Chen's Blog: BMQ 0.98 releaseAlfred Chenhttp://www.blogger.com/profile/03164306846702841944noreply@blogger.comBlogger17125tag:blogger.com,1999:blog-2963790426029213933.post-64861560825082278292019-08-13T10:56:37.574-07:002019-08-13T10:56:37.574-07:00I have done testing on Manjaro, there seems to be ...I have done testing on Manjaro, there seems to be a regression. I'll post details to gitlab soon.<br />The thing is and I don't know the reason, on Ubuntu with 5.2+BMQ+nohz_full kernel builds even faster - 11 minutes, the same kernel tree with the same config with the same gcc version was tested in Manjaro and Ubuntu, but I'll discard that result for now until I'm sure why.<br />Other thing is that 5.2 kernel or BMQ hangs on Ryzen and i7 as well, more frequent on Ryzen. I haven't figured out whether that's 5.2 or BMQ.<br /><br />BR, EduardoAnonymousnoreply@blogger.comtag:blogger.com,1999:blog-2963790426029213933.post-31058067799330767012019-08-12T22:49:12.667-07:002019-08-12T22:49:12.667-07:00I have Ryzen 1700 + OC, I'll try to reproduce ...I have Ryzen 1700 + OC, I'll try to reproduce the issue in Manjaro which has full tickless by default, therefore I'm not able to check tickless idle without recompiling stock kernel and my own.<br />Let's see how much time I have to compile everything for tests :)<br /><br />Btw, ryzen differs from intel with so called core complexes (CCX) 4 cores in each CCX, multiple CCX found in one ryzen chip.<br />The benefit could be to keep tasks scheduled in one CCX then there is no penalty, but when multiple CCX communicate, bigger latency is involved.<br /><br />My observations are that CK do not address this issue either, therefore queue count is off when running muqss on ryzen, like 17 on 8 cores or 25 on 4 cores, etc., users report different counts and performance numbers, this somehow doesn't seem right, but I'm just an observer here :)<br /><br />BR, EduardoAnonymousnoreply@blogger.comtag:blogger.com,1999:blog-2963790426029213933.post-29009278218265701902019-08-10T05:23:12.830-07:002019-08-10T05:23:12.830-07:00Since Con Kolivas recommends the whole -ck patchse...Since Con Kolivas recommends the whole -ck patchset, and not only the MuQSS patch + that it is intended to run with CONFIG_NO_HZ_IDLE as per comment: http://ck-hack.blogspot.com/2018/12/linux-420-ck1-muqss-version-0185-for.html?showComment=1549056997845#c6415843931930386538 , i was wondering if you had done the same comparison with those settings?Sveinar Søplerhttps://www.blogger.com/profile/18401720133659243541noreply@blogger.comtag:blogger.com,1999:blog-2963790426029213933.post-16260403660482720962019-08-01T11:57:33.986-07:002019-08-01T11:57:33.986-07:00Thanks to response, yes, it works :)
I noted that ...Thanks to response, yes, it works :)<br />I noted that cpuacct is also disabled on others cpu sched, like muqss, so I discovered that had an reason for it and decided delete my post above xD.<br />I'll continue to use bmq with this patch and see how it will work for next days. ltsdwhttps://www.blogger.com/profile/15407009266806601658noreply@blogger.comtag:blogger.com,1999:blog-2963790426029213933.post-42368391749596845042019-08-01T06:29:57.282-07:002019-08-01T06:29:57.282-07:00Here is the debug patch to enable CGROUP_CPUACCOUN...Here is the debug patch to enable CGROUP_CPUACCOUNTING, pls give it a test, it's not guarantee to work.<br />https://gitlab.com/alfredchen/bmq/blob/master/5.2/enabled_CGROUP_CPUACCOUNTING.patchAlfred Chenhttps://www.blogger.com/profile/03164306846702841944noreply@blogger.comtag:blogger.com,1999:blog-2963790426029213933.post-85207643360815915482019-07-29T22:54:34.175-07:002019-07-29T22:54:34.175-07:00>> I guess you pass "-j12" for the...>> I guess you pass "-j12" for the kernel compilation, right?<br />yes<br />>> Pls help to report a bug at https://gitlab.com/alfredchen/bmq/issues<br />done: https://gitlab.com/alfredchen/bmq/issues/7Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-2963790426029213933.post-78360223656389909562019-07-29T22:01:59.953-07:002019-07-29T22:01:59.953-07:00I guess you pass "-j12" for the kernel c...I guess you pass "-j12" for the kernel compilation, right?<br />All my machines have just two level topology, those are "smt" and "coregroup". It's different form ryzen 5 2600 which has an additional "core" level. My best guess is BMQ failed to sustain cpu load from "core" level cpus, but it's hard to tell from code.<br />Pls help to report a bug at https://gitlab.com/alfredchen/bmq/issues, I'd like to see if other ryzen 5(or similar topology) users report similar issues.<br />Meanwhile, I will try to reproduce it using qemu with proper cpu topology setup.Alfred Chenhttps://www.blogger.com/profile/03164306846702841944noreply@blogger.comtag:blogger.com,1999:blog-2963790426029213933.post-90984578518730613042019-07-29T19:14:44.334-07:002019-07-29T19:14:44.334-07:00# cfs default (not tuned) 500hz NO_HZ_FULL
real 4m...# cfs default (not tuned) 500hz NO_HZ_FULL<br />real 4m1,772s<br />user 36m55,804s<br />sys 4m2,401s<br /><br />hm i didn't think it would be faster than muqss<br />all cpu schedulers tested under the same conditionsAnonymousnoreply@blogger.comtag:blogger.com,1999:blog-2963790426029213933.post-91582252810019368372019-07-29T16:33:46.074-07:002019-07-29T16:33:46.074-07:00Pls check with mainline CFS, I'll double check...Pls check with mainline CFS, I'll double check coregroup scheduling in some cases.Alfred Chenhttps://www.blogger.com/profile/03164306846702841944noreply@blogger.comtag:blogger.com,1999:blog-2963790426029213933.post-23465534446180523232019-07-29T12:14:40.215-07:002019-07-29T12:14:40.215-07:00and with muqss i have bigger cpu usage and more/st...and with muqss i have bigger cpu usage and more/stable fpsAnonymousnoreply@blogger.comtag:blogger.com,1999:blog-2963790426029213933.post-57326301306814493532019-07-29T08:38:15.291-07:002019-07-29T08:38:15.291-07:00[ 0.167088] bmq: BMQ CPU Scheduler 0.98 by Alfr...[ 0.167088] bmq: BMQ CPU Scheduler 0.98 by Alfred Chen.<br />[ 0.461691] bmq: cpu #0 affinity check mask - smt 0x00000040<br />[ 0.461691] bmq: cpu #0 affinity check mask - coregroup 0x000001c6<br />[ 0.461691] bmq: cpu #0 affinity check mask - core 0x00000e38<br />[ 0.461691] bmq: cpu #1 affinity check mask - smt 0x00000080<br />[ 0.461691] bmq: cpu #1 affinity check mask - coregroup 0x000001c5<br />[ 0.461691] bmq: cpu #1 affinity check mask - core 0x00000e38<br />[ 0.461691] bmq: cpu #2 affinity check mask - smt 0x00000100<br />[ 0.461691] bmq: cpu #2 affinity check mask - coregroup 0x000001c3<br />[ 0.461691] bmq: cpu #2 affinity check mask - core 0x00000e38<br />[ 0.461691] bmq: cpu #3 affinity check mask - smt 0x00000200<br />[ 0.461691] bmq: cpu #3 affinity check mask - coregroup 0x00000e30<br />[ 0.461691] bmq: cpu #3 affinity check mask - core 0x000001c7<br />[ 0.461691] bmq: cpu #4 affinity check mask - smt 0x00000400<br />[ 0.461691] bmq: cpu #4 affinity check mask - coregroup 0x00000e28<br />[ 0.461691] bmq: cpu #4 affinity check mask - core 0x000001c7<br />[ 0.461691] bmq: cpu #5 affinity check mask - smt 0x00000800<br />[ 0.461691] bmq: cpu #5 affinity check mask - coregroup 0x00000e18<br />[ 0.461691] bmq: cpu #5 affinity check mask - core 0x000001c7<br />[ 0.461691] bmq: cpu #6 affinity check mask - smt 0x00000001<br />[ 0.461691] bmq: cpu #6 affinity check mask - coregroup 0x00000187<br />[ 0.461691] bmq: cpu #6 affinity check mask - core 0x00000e38<br />[ 0.461691] bmq: cpu #7 affinity check mask - smt 0x00000002<br />[ 0.461691] bmq: cpu #7 affinity check mask - coregroup 0x00000147<br />[ 0.461691] bmq: cpu #7 affinity check mask - core 0x00000e38<br />[ 0.461691] bmq: cpu #8 affinity check mask - smt 0x00000004<br />[ 0.461691] bmq: cpu #8 affinity check mask - coregroup 0x000000c7<br />[ 0.461691] bmq: cpu #8 affinity check mask - core 0x00000e38<br />[ 0.461691] bmq: cpu #9 affinity check mask - smt 0x00000008<br />[ 0.461691] bmq: cpu #9 affinity check mask - coregroup 0x00000c38<br />[ 0.461691] bmq: cpu #9 affinity check mask - core 0x000001c7<br />[ 0.461691] bmq: cpu #10 affinity check mask - smt 0x00000010<br />[ 0.461691] bmq: cpu #10 affinity check mask - coregroup 0x00000a38<br />[ 0.461691] bmq: cpu #10 affinity check mask - core 0x000001c7<br />[ 0.461691] bmq: cpu #11 affinity check mask - smt 0x00000020<br />[ 0.461691] bmq: cpu #11 affinity check mask - coregroup 0x00000638<br />[ 0.461691] bmq: cpu #11 affinity check mask - core 0x000001c7<br />Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-2963790426029213933.post-86030741142468014882019-07-29T03:44:14.052-07:002019-07-29T03:44:14.052-07:00Interesting. On my test platforms, there is no sig...Interesting. On my test platforms, there is no sign of regression during the sanity tests from release to release, also comparing to CFS mainline scheduler.<br /><br />So, does other ryzen user has similar issues?<br />And would you plz post the output of "dmesg | grep -i bmq" so I can check the cpu topology setup for bmq?Alfred Chenhttps://www.blogger.com/profile/03164306846702841944noreply@blogger.comtag:blogger.com,1999:blog-2963790426029213933.post-42814916958225074552019-07-28T16:53:12.092-07:002019-07-28T16:53:12.092-07:00my test with ryzen 5 2600 in kernel compilation
#...my test with ryzen 5 2600 in kernel compilation<br /><br /># bmq0.98 500hz NO_HZ_FULL<br />real 5m14,579s<br />user 33m45,957s<br />sys 3m48,909s<br /><br /># muqss0.193-smt 100hz NO_HZ_FULL<br />real 4m13,836s<br />user 41m25,789s<br />sys 2m29,432s<br /><br /># muqss0.193-mc 100hz NO_HZ_FULL<br />real 4m13,132s<br />user 42m56,970s<br />sys 2m38,041s<br />Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-2963790426029213933.post-38660093283168902102019-07-27T10:27:09.349-07:002019-07-27T10:27:09.349-07:00This comment has been removed by the author.ltsdwhttps://www.blogger.com/profile/15407009266806601658noreply@blogger.comtag:blogger.com,1999:blog-2963790426029213933.post-32473134027881767072019-07-16T18:34:48.427-07:002019-07-16T18:34:48.427-07:00There is just #2 minor change to BMQ itself, it sh...There is just #2 minor change to BMQ itself, it shouldn't be any human noticeable different. :)Alfred Chenhttps://www.blogger.com/profile/03164306846702841944noreply@blogger.comtag:blogger.com,1999:blog-2963790426029213933.post-70501576780201030742019-07-15T14:06:44.315-07:002019-07-15T14:06:44.315-07:00Runs good but I think 0.97 has been faster.Runs good but I think 0.97 has been faster.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-2963790426029213933.post-84566541687845740772019-07-15T11:17:19.054-07:002019-07-15T11:17:19.054-07:00Up and running on multiple machines no problems so...Up and running on multiple machines no problems so far.Anonymousnoreply@blogger.com