linux-kernel - Re: [PATCH v7 0/8] Tunable sched_mc_power

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:	Tue, 6 Jan 2009 20:24:45 +0530
From:	Vaidyanathan Srinivasan <svaidy@...ux.vnet.ibm.com>
To:	Mike Galbraith <efault@....de>
Cc:	Ingo Molnar <mingo@...e.hu>,
	Balbir Singh <balbir@...ux.vnet.ibm.com>,
	linux-kernel@...r.kernel.org,
	Peter Zijlstra <a.p.zijlstra@...llo.nl>,
	Andrew Morton <akpm@...ux-foundation.org>
Subject: Re: [PATCH v7 0/8] Tunable sched_mc_power_savings=n

* Mike Galbraith <efault@....de> [2009-01-05 05:40:16]:

> On Mon, 2009-01-05 at 08:50 +0530, Vaidyanathan Srinivasan wrote: 
> > When CONFIG_SCHED_DEBUG is enabled, the sched domain tree is dumped
> > (dmesg)
> 
> Oh, that.  I'm dense <thwack>
> 
> [    0.476050] CPU0 attaching sched-domain:
> [    0.476052]  domain 0: span 0-1 level MC
> [    0.476054]   groups: 0 1
> [    0.476057]   domain 1: span 0-3 level CPU
> [    0.476058]    groups: 0-1 2-3
> [    0.476062] CPU1 attaching sched-domain:
> [    0.476064]  domain 0: span 0-1 level MC
> [    0.476065]   groups: 1 0
> [    0.476067]   domain 1: span 0-3 level CPU
> [    0.476069]    groups: 0-1 2-3
> [    0.476072] CPU2 attaching sched-domain:
> [    0.476073]  domain 0: span 2-3 level MC
> [    0.476075]   groups: 2 3
> [    0.476077]   domain 1: span 0-3 level CPU
> [    0.476078]    groups: 2-3 0-1
> [    0.476081] CPU3 attaching sched-domain:
> [    0.476083]  domain 0: span 2-3 level MC
> [    0.476084]   groups: 3 2
> [    0.476086]   domain 1: span 0-3 level CPU
> [    0.476088]    groups: 2-3 0-1

Hi Mike,

This seems to be correct for the configuration.  Hope this would be
same for shced_mc=1 and sched_mc=2 since you would have hacked
mc_capable.

By default, all 4 cores will be 1 group at CPU at sched_mc={1,2} so
that packages are clearly identified in the CPU level sched groups.

 
> 2.6.26.8
> [    0.524043] CPU0 attaching sched-domain:
> [    0.524045]  domain 0: span 0-1
> [    0.524046]   groups: 0 1
> [    0.524049]   domain 1: span 0-3
> [    0.524051]    groups: 0-1 2-3
> [    0.524054] CPU1 attaching sched-domain:
> [    0.524055]  domain 0: span 0-1
> [    0.524056]   groups: 1 0
> [    0.524059]   domain 1: span 0-3
> [    0.524060]    groups: 0-1 2-3
> [    0.524063] CPU2 attaching sched-domain:
> [    0.524064]  domain 0: span 2-3
> [    0.524065]   groups: 2 3
> [    0.524068]   domain 1: span 0-3
> [    0.524069]    groups: 2-3 0-1
> [    0.524072] CPU3 attaching sched-domain:
> [    0.524073]  domain 0: span 2-3
> [    0.524075]   groups: 3 2
> [    0.524077]   domain 1: span 0-3
> [    0.524078]    groups: 2-3 0-1
> 
> > I was actually asking about software threads specified in the sysbench
> > benchmark.  Your have run almost 256 clients on a 4 core box, does
> > that mean sysbench had 256 worker threads?
> 
> Yes.

Let me try similar experiments on my dual socket quad core system.
I was limiting the threads to 8 assuming that the system will max-out
by then.

Thanks for the updates.

--Vaidy

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/