[geeks] New Itanium machines from SGI

[Joshua D Boyd]

On Wed, Jan 08, 2003 at 02:25:59AM -0500, dave wrote:
> jdboyd at cs.millersville.edu writes:
> 
> >Well, they aren't exactly just sitting around.  My understanding is that
> >the main hope for a good compiler rests on SGI having taken the MIPSpro
> >compiler and stripped out the MIPS backend, then donated it as open
> >source, while both SGI and Intel and paying money to a chinese
> >university to develop this compiler for Itanium.  
> 
> So what percentage of theoretical peak performance are the
> current compilers wringing out of it? I've never seen a
> number for that.

I haven't seen a specific number for theoretical peak versus what it
currently does.  I just know that with GCC, most bundles only have one
instruction in them rather than 3.  However, that doesn't mean that a
better compiler would actually be able to get 3 times the performance.
-- 
Joshua D. Boyd