SSE & compilers
Robert G. Brown
rgb at phy.duke.edu
Wed Aug 22 05:45:33 PDT 2001
On Tue, 21 Aug 2001, Dan Kirkpatrick wrote:
> 1. What do you know about SSE ? Apparently the p III and P 4 have extra
> hardware for fp work which is ignored by current compilers but can buy big
> factors in speed...
> we can code in assembler for it pretty easily if the PIII prcocessors have
> it. Comments?
SSE support is coming to gcc. Indeed, it really is already in gcc but
only if you get a very current snapshot and build it yourself -- the gcc
distributed through e.g. RH 7.1 is still 2.96 and doesn't support it.
It looks like gcc 3.x will be in RH 7.2 (3-3.0-6 appears to be in
rawhide at the moment).
I haven't built the new gcc myself but I plan to as I really need to
test out SSE performance on e.g. Athlons, but there are a number of
issues (including matching kernel stuff that is still under development)
that make this a daunting proposition. For example, RH 7.2 still
appears to ALSO distribute 2.96, presumably to build the kernel itself?
I keep waiting in hope that it will all be delivered to me neatly
packaged in 7.2 or an early update thereof (along with ext3 and more).
If anyone on the list HAS built 3.0.x and tested the new SSE stuff
(especially on Tbirds or Palominos) I'd love to hear about it.
> 2. Apparently there are several optimizing compilers out there (like
> portland) which do better than gcc. Any suggestions? Information on costs?
Portland has a pretty decent suite, and you can easily visit their
website to take a look at their products. Their University prices
aren't bad and you ought to be able to wheedle a loaner/test/temporary
license out of them to play with for a month or so. PGI is also a
longtime linux and beowulf supporter -- they often are a presence at
linux shows and help sponsor e.g. extreme linux events. Not bad folks
to give business to.
> Dan Kirkpatrick dkirk at physics.syr.edu
> Computer Systems Manager
> Department of Physics
> Syracuse University, Syracuse, NY
> http://www.physics.syr.edu/help/ Fax:(315) 443-9103
> Beowulf mailing list, Beowulf at beowulf.org
> To change your subscription (digest mode or unsubscribe) visit http://www.beowulf.org/mailman/listinfo/beowulf
Robert G. Brown http://www.phy.duke.edu/~rgb/
Duke University Dept. of Physics, Box 90305
Durham, N.C. 27708-0305
Phone: 1-919-660-2567 Fax: 919-660-2525 email:rgb at phy.duke.edu
More information about the Beowulf