[Beowulf] AMD performance (was 500GB systems)

Christopher Samuel samuel at unimelb.edu.au
Mon Jan 14 18:44:02 PST 2013


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

On 14/01/13 19:20, Stu Midgley wrote:

> As you might guess, we were very happy with how our codes run on
> the Phi and the time/effort required to port.  It is very very
> simple to use and the performance is excellent :)  With no tuning
> (just recompile) we saw a single phi go at about 1.7x faster than
> our current AMD 64 cores nodes.

Is that running natively on the MIC only, or using the compiler/MKL
offload capabilities and thus the host CPU & RAM as well?

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 Thunderbird - http://www.enigmail.net/

iEYEARECAAYFAlD0wnIACgkQO2KABBYQAh+oSgCfUe5JXPM9C/wnkgyCFII8WULc
V5IAn0fHbBQjLwrsAPHreTvyY74sjrku
=f64k
-----END PGP SIGNATURE-----


More information about the Beowulf mailing list