Re: Hyperthreading on Core i7s: To use or not to use?

[peng huang <huangpeng.linux@gmail.com>]

Hello,

I have been done some benchmark things on the processors with HT,
if you run lots of processes/threads that more than the physical 
cores,may be you should enable the HT to get more benefits.
#unless your processes cause some cache competitions. 

I will send you some test data later. 

-huang

2010-01-26 (火) の 10:56 +0000 に Daniel J Blueman さんは書きました:
> On Jan 26, 10:10 am, Justin Piszcz <jpiszcz@lucidpixels.com> wrote:
> > Hello,
> >
> > Should the 'correct' kernel [CPU] configuration for a core i7 860/870..?
> >
> > - Multi-core support
> > - Cores: 8
> > - SMT: Enabled/ON
> >
> > From CONFIG_SCHED_SMT:
> >
> >    . SMT scheduler support improves the CPU scheduler's decision making      .
> >    . when dealing with Intel Pentium 4 chips with HyperThreading at a        .
> >    . cost of slightly increased overhead in some places. If unsure say       .
> >    . N here.                                                                 .
> >
> > Does this also 'help' and/or 'apply' as much when dealing with Core i7s?
> >
> > --
> >
> > Quick little benchmark (pbzip2 -9 linux kernel source), the benchmark is
> > really within the noise (8 on/off)
> > - Multicore(8)/HT(Off) = 73.72user 0.33system 0:09.50elapsed 779%CPU (0avgtext+0avgdata 458528maxresident
> > - Multicore(8)/HT(On) = 74.28user 0.40system 0:09.67elapsed 772%CPU (0avgtext+0avgdata 428304maxresident
> > - Multicore(4)/HT(On) = 68.76user 0.30system 0:17.44elapsed 396%CPU (0avgtext+0avgdata 213616maxresident)k
> >
> > --
> >
> > Has anyone done any in-depth benchmarking for the core i7s that have multiple
> > cores and HT disabled/enabled?
> 
> With my Dell Studio 15 (model 1557) laptop, there is no option to
> disable HT in the current BIOS, so booting with maxcpus=4 (since the
> kernel enumerates non-sibling cores first) gave me a 5-15% speedup on
> some large image processing (convolution, FFTs, conversion) on all
> available cores, presumably due to better cache efficiency.
> 
> Booting with maxcpus=4 prevents any of the cores sitting in C6, needed
> for turbo-boost and a lower thermal profile, though I did find
> scheduling latency and responsiveness better under load booting with
> maxcpus=4, so favour this when plugged in.
> 
> Clearly, having the BIOS option allows benefit to certain applications
> - Dell should give their users the choice!
> 
> Perhaps the 'noht' boot option should be reintroduced to initialise
> all cores, but only expose non-sibling cores to the OS (thus allowing
> C6)?
> 
> Daniel
> 
> tip: modprobe msr and use turbostat to monitor turbo-boost and C-state
> residency: http://bugzilla.kernel.org/attachment.cgi?id=24673


-- 
peng huang <huangpeng.linux@gmail.com>