[Beowulf] General thoughts on Xeon 56xx versus E5 series?

Christopher Samuel samuel at unimelb.edu.au
Thu Sep 13 16:44:47 PDT 2012


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 14/09/12 00:09, holway at th.physik.uni-frankfurt.de wrote:

> It turned out that the kernel did not yet support AVX or FMA4 or
> some other new fancy bit of register blah blah so it reverted to
> some compatibility support for floating point operations. (I am not
> a kernel dev :)

Hmm, I'd be surprised if the kernel affected FP performance much, I'd
have thought that'd be all in user space and the kernel would be doing
its best to stay out of your way at that point (until you try and do I/O).

Things that might affect it may be poor hardware support in terms of
NUMA detection and locality as well as process scheduling.

I'm not a kernel dev either, I just follow the git repo (and the btrfs
mailing list). :-)

> If your going to play with newer processors you have to make sure
> your kernel supports all the functions of that CPU.

Agreed.

> Kernel development can be a little way behind on mainstream
> releases.

Actually the mainline kernel repo usually has support for CPUs from
Intel/AMD before they hit the streets, it's the "enterprise" distros
where they generally fall short.

cheers,
Chris
- -- 
 Christopher Samuel        Senior Systems Administrator
 VLSCI - Victorian Life Sciences Computation Initiative
 Email: samuel at unimelb.edu.au Phone: +61 (0)3 903 55545
 http://www.vlsci.org.au/      http://twitter.com/vlsci

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://www.enigmail.net/

iEYEARECAAYFAlBSb+8ACgkQO2KABBYQAh9ZVQCfTM93VBuTmayyybN7dOw2oRdY
tYEAn18deGpqEnrTdxRIftn5lWy2tpNK
=65Bh
-----END PGP SIGNATURE-----



More information about the Beowulf mailing list