On May 4, 7:26 pm, William Stein <wst...@gmail.com> wrote:
> On Mon, May 4, 2009 at 6:50 PM, mabshoff <mabsh...@googlemail.com> wrote:
>
> > On May 4, 11:39 am, Tom Boothby <tomas.boot...@gmail.com> wrote:
> >> Recently, Sun donated a new machine to the Sage community. It's a Sun
> >> T5240 with two SPARC T2 processors (8 cores total) and 32GB of RAM,
>
> > Well, it has 128 "threads".
>
> Unfortunately, it seems that this does *NOT* mean that if you write a
> little C program, spawn 128 threads, and watch them run, then you can
> do 128 times what you would do with 1 thread. You still can only do 8
> times as much as with 1 thread. For raw computation, I don't think
> that processor is any better than 8 single cores. I still have no
> idea how to leverage these 128 "threads". If anybody knows, please
> tell me.
Yes, for FP or INT bound computations you won't get more than a 8 or
16 time speedup when going from 1 to 128 "threads", the box is really
meant for different things than HPC stuff. But it has some crypto
units for example and several build in 10GBit interfaces which can
actually be saturated. The box is a killer for databse and Java stuff
though.
> >> and runs Solaris 5.10. As with the other machines in the network, you
> >> have access to your home directory from there; and provided you
> >> haven't changed your password this week, your login should be the same
> >> as it is at sage.math. Find it at t2.math.washington.edu, or from
> >> sage.math, simply t2.
>
> >> Thanks to Michael Abshoff, there's a custom-built toolchain sufficient
> >> to build Sage.
>
> > No, unfortunately it is broken on the t2 since it dies with an
> > internal compiler error when building gfortran.
>
> Yep. But if you're building C code (not Fortran), then you can use
> the toolchain. If you login via bash you get this toolchain (type
> which gcc to confirm this).
Yeah, I just wanted to point out that building Sage is a problem. I am
working on a gcc 4.4.0 based toolchain and I am hoping the problem
with gfortran will go away.
> > Give the seepd the t2
> > compiles with it is recommended to use the Sparc binaries I post
> > anyway since that is much faster and actually works.
>
> That binary is installed system-wide. Note that the last one I found
> from you was Sage 3.2.3.
Check out
http://www.sagemath.org/bin/solaris/
for 3.4.1 binaries. There should be 3.4.2 binaries shortly. I found
one more critical, but trivial to fix issue in 3.4.2+prime_pi()
patch.
Cheers,
Michael
> --
> William Stein
> Associate Professor of Mathematics
> University of Washingtonhttp://wstein.org
--~--~---------~--~----~------------~-------~--~----~
To post to this group, send email to sage-devel@googlegroups.com
To unsubscribe from this group, send email to
sage-devel-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URLs: http://www.sagemath.org
-~----------~----~----~----~------~----~------~--~---
