tag:blogger.com,1999:blog-2963790426029213933.post6280151156647513884..comments2024-02-29T00:33:07.382-08:00Comments on Alfred Chen's Blog: PDS 0.98g releaseAlfred Chenhttp://www.blogger.com/profile/03164306846702841944noreply@blogger.comBlogger16125tag:blogger.com,1999:blog-2963790426029213933.post-18931382615723501082017-12-22T10:00:18.618-08:002017-12-22T10:00:18.618-08:00Thank you both for your replies! Just wanted to be...Thank you both for your replies! Just wanted to be informed until 2018 faces.<br />BR, Manuel KrauseAnonymousnoreply@blogger.comtag:blogger.com,1999:blog-2963790426029213933.post-54135955551379237562017-12-21T20:53:46.803-08:002017-12-21T20:53:46.803-08:00@all
Sorry for late reply. It has been busy these ...@all<br />Sorry for late reply. It has been busy these days.<br />I am working on an improvement to reduce re-balance overhead to 3/4 or 3/8, hopefully there will be debug patch for testing in Jan 2018.Alfred Chenhttps://www.blogger.com/profile/03164306846702841944noreply@blogger.comtag:blogger.com,1999:blog-2963790426029213933.post-22616258653430284732017-12-21T20:48:22.120-08:002017-12-21T20:48:22.120-08:00@pf
It is ok. I will include this fix in next rele...@pf<br />It is ok. I will include this fix in next release. And add more fix in if there is still an issue reported by the relevant user.Alfred Chenhttps://www.blogger.com/profile/03164306846702841944noreply@blogger.comtag:blogger.com,1999:blog-2963790426029213933.post-23074469366950862292017-12-21T02:55:47.078-08:002017-12-21T02:55:47.078-08:00This issue was reported not by me, and I do not ha...This issue was reported not by me, and I do not have a possibility to test the fix, thus still waiting for the response from the relevant user.Oleksandr Natalenkohttps://www.blogger.com/profile/12098091624630953604noreply@blogger.comtag:blogger.com,1999:blog-2963790426029213933.post-80743318547144587612017-12-20T09:28:19.855-08:002017-12-20T09:28:19.855-08:00@Oleksandr Natalenko/ @post-factum/ @pf:
An applic...@Oleksandr Natalenko/ @post-factum/ @pf:<br />An applicable patch of this uploaded here by me for reference: <br />https://pastebin.com/HSkZrzB1<br /><br />Have you had a chance to (let) test this one?<br /><br />BTW, Oleksandr, again many thanks for your work on your patchset, especially for keeping track of all the BFQ I/O improvements.<br /><br />BR, Manuel KrauseAnonymousnoreply@blogger.comtag:blogger.com,1999:blog-2963790426029213933.post-27485466358699698392017-12-13T11:41:26.392-08:002017-12-13T11:41:26.392-08:00@Alfred:
I'm really confident with the current...@Alfred:<br />I'm really confident with the current state! :-)<br />But please, also let us all others know of possible improvements to be tested. Speaking for me, I'm too curious for the new, and the others don't need to take a risk.<br /><br />BR, Manuel KrauseAnonymousnoreply@blogger.comtag:blogger.com,1999:blog-2963790426029213933.post-47110557906585984792017-12-11T21:51:30.617-08:002017-12-11T21:51:30.617-08:00Hi,
This will be interesting, Alfred! So, You'...Hi,<br /><br />This will be interesting, Alfred! So, You'll improve already exising SMT sensitive scheduling? If You need alpha/beta tester w/ Ryzen, shoot me an e-mail. Since Ryzen have those ccx, it behaves behaves a bit differently.<br /><br />BR, Eduardo<br /><br />P.S. I already tried muqss builds with both smt and mc, results are mixed as Con predicted. Diablo3, as with vrq back in the day, still had interesting scheduler behaviour :)Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-2963790426029213933.post-79060637091157653022017-12-11T16:22:42.997-08:002017-12-11T16:22:42.997-08:00Thanks for this third party tests.
The most import...Thanks for this third party tests.<br />The most important featrue of this year would be the smt group rebalance, it helps the performance when using number of physical cores(make -j4 in the tests).<br />PDS will focus on reducing overhead next year.Alfred Chenhttps://www.blogger.com/profile/03164306846702841944noreply@blogger.comtag:blogger.com,1999:blog-2963790426029213933.post-30216931368862584332017-12-11T14:13:05.935-08:002017-12-11T14:13:05.935-08:00Great work!
It would be interesting to also compa...Great work!<br /><br />It would be interesting to also compare the MuQSS with int0 and RQSHARE_MC=y.Alexandre Fradehttps://www.blogger.com/profile/15964702307004146298noreply@blogger.comtag:blogger.com,1999:blog-2963790426029213933.post-77264885532717867412017-12-11T08:15:45.553-08:002017-12-11T08:15:45.553-08:00Thanks Alfred.
I've done the usual throughput ...Thanks Alfred.<br />I've done the usual throughput tests with PDS 0.98g.<br /><br />https://docs.google.com/spreadsheets/d/163U3H-gnVeGopMrHiJLeEY1b7XlvND2yoceKbOvQRm4/edit?usp=sharing<br /><br />As you already noted, there is a slight regression comparing to linux 4.13 + PDS 0.98d,<br />but the results are still within +- 1% of CFS.<br /><br />PedroAnonymousnoreply@blogger.comtag:blogger.com,1999:blog-2963790426029213933.post-66302938909711024372017-12-10T22:57:01.523-08:002017-12-10T22:57:01.523-08:00@pf
Sorry for the late reply. Please try the patch...@pf<br />Sorry for the late reply. Please try the patch below, if it works, I will include it in next release.<br /><br />diff --git a/kernel/sched/pds.c b/kernel/sched/pds.c<br />index 722f357a782b..fcc7ee0c8091 100644<br />--- a/kernel/sched/pds.c<br />+++ b/kernel/sched/pds.c<br />@@ -6155,7 +6155,14 @@ static void sched_init_topology_cpumask(void)<br /> per_cpu(sd_llc_id, cpu) =<br /> cpumask_first(cpu_coregroup_mask(cpu));<br /><br />+#ifdef CONFIG_SCHED_MC<br /> cpumask_complement(&tmp, cpu_coregroup_mask(cpu));<br />+#elif CONFIG_SCHED_SMT<br />+ cpumask_complement(&tmp, topology_sibling_cpumask(cpu));<br />+#else<br />+ cpumask_setall(&tmp);<br />+ cpumask_clear_cpu(cpu, &tmp);<br />+#endif<br /> if (cpumask_and(&tmp, &tmp, topology_core_cpumask(cpu))) {<br /> printk(KERN_INFO "pds: sched_cpu_affinity_chk_masks[%d] core 0x%08lx",<br /> cpu, tmp.bits[0]);<br />Alfred Chenhttps://www.blogger.com/profile/03164306846702841944noreply@blogger.comtag:blogger.com,1999:blog-2963790426029213933.post-21922650359167884922017-12-06T11:28:59.963-08:002017-12-06T11:28:59.963-08:00Alfred,
could you please check build log for Powe...Alfred,<br /><br />could you please check build log for PowerPC [1]? It fails on cpu_coregroup_mask() because it is not defined for PowerPC architecture. I think the respective code should be wrapped into CONFIG_SCHED_MC ifdef, but wanted to check with you first.<br /><br />[1] https://kojipkgs.fedoraproject.org//work/tasks/9440/23559440/build.logOleksandr Natalenkohttps://www.blogger.com/profile/12098091624630953604noreply@blogger.comtag:blogger.com,1999:blog-2963790426029213933.post-55796090931762556722017-12-04T07:59:58.204-08:002017-12-04T07:59:58.204-08:00Btw, I did build i686 UP last week; pds+bfq w/ Min...Btw, I did build i686 UP last week; pds+bfq w/ Ming's I/O Perf, Bart's suspend/resume, 4.15 BFQ, some other patches. I tested it on my netbook over the weekend, including suspend/resume; working fine. Same for my x64 PC too; thanks!jwh7https://www.blogger.com/profile/09659185315567537391noreply@blogger.comtag:blogger.com,1999:blog-2963790426029213933.post-12432860793685721402017-11-27T16:06:59.096-08:002017-11-27T16:06:59.096-08:00@jwh7
Yes, that would be the fix.
I will include i...@jwh7<br />Yes, that would be the fix.<br />I will include it in next release. Thanks for the reporting.Alfred Chenhttps://www.blogger.com/profile/03164306846702841944noreply@blogger.comtag:blogger.com,1999:blog-2963790426029213933.post-11738802817935875482017-11-27T13:39:49.079-08:002017-11-27T13:39:49.079-08:00Do I just need a "#ifdef CONFIG_SMP" wra...Do I just need a "#ifdef CONFIG_SMP" wrapped around that line? I will try it tonight...jwh7https://www.blogger.com/profile/09659185315567537391noreply@blogger.comtag:blogger.com,1999:blog-2963790426029213933.post-27428463569698971892017-11-26T17:43:22.885-08:002017-11-26T17:43:22.885-08:00Hey Alfred; I'm running this on x64 now. Than...Hey Alfred; I'm running this on x64 now. Thanks as always! I was away for a while testing 4.14-rc's and then linux-next, but have now synced back to 4.14. Everything is fine so far, but when I build x86 UP, I get this:<br />kernel/sched/pds.c: In function ‘sched_init’:<br />kernel/sched/pds.c:6274:11: error: ‘struct rq’ has no member named ‘online’<br /> cpu_rq(0)->online = true;<br />[The `->` is highlighted.]jwh7https://www.blogger.com/profile/09659185315567537391noreply@blogger.com