[Beowulf] Definition of HPC
Craig Tierney - NOAA Affiliate
craig.tierney at noaa.gov
Thu Apr 18 15:40:16 PDT 2013
Only for benchmarking? We have done this for years on our production
clusters (and SGI provides a tool this and more to clean up nodes). We
have this in our epilogue so that we can clean out memory on our diskless
nodes so there is nothing stale sitting around that can impact the next
users job.
Craig
On Thu, Apr 18, 2013 at 2:36 PM, Max R. Dechantsreiter <
max at performancejones.com> wrote:
>
>
>
> On Thu, 18 Apr 2013, Mark Hahn wrote:
>
> >> What problems?
> >
> > performance, of course. drop_caches is really only sane for
> benchmarking,
> > where you want to control for hot/cold caches.
>
> Indeed.
>
> I thought you might know of harmful instances of which I was unaware.
>
> > otherwise, you're almost certainly better off either letting the kernel
> > optimize global caching, and/or fix your application
> > to avoid polluting the cache (O_DIRECT, madvise, etc).
> >
> > chip vendors could provide a drop_caches for CPUs, too, and it would
> also be
> > "non-destructive". afaik, such instructions do exist, and are always
> > privileged, for basically DOS-based the same reason.
> >
> > regards, mark hahn.
> >
> _______________________________________________
> Beowulf mailing list, Beowulf at beowulf.org sponsored by Penguin Computing
> To change your subscription (digest mode or unsubscribe) visit
> http://www.beowulf.org/mailman/listinfo/beowulf
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.beowulf.org/pipermail/beowulf/attachments/20130418/91103320/attachment.html>
More information about the Beowulf
mailing list