p4 v itinium

Mark Hahn hahn at physics.mcmaster.ca
Tue May 14 14:12:29 PDT 2002

> > can a p4 access more than 4GB memory - some of the codes have a 
> > requirement to access large memory

depends on what you mean by "access".  if you mean "faster than disk",
then yes, that's possible.  you'll wind up using some form of OS-supported
bank-switching, which considering that Linux syscalls are <1us, can
be fairly acceptable.  but a single ia32 process can never
(not even with icky 16:32 segmentation) directly access >4G, since the 
address-mapping hardware goes through a strictly 32b stage.

> A single process on any x86 can only access 2-3 GB of memory under
> Linux without resorting to heroic measures, as far as I know.

3G (-128M) is trivial.  there's a simple non-heroic patch for 3.5G,
but past that is probably less heroic than plain old "risky" ;)

> > if p4 can access greater than 4GB memory are there any 
> > preformance gains in going for the more expensive itanium option
> For floating-point intensive codes, I've found an Itanium 733MHz is
> roughly as fast as a 1.4GHz Athlon; I wouldn't expect a P4 to be much
> faster than that, based on the benchmarks I ran last summer.

I know nothing about ia64, but for other modern processors,
gcc 3.1 is a fairly dramatic step forward.  afaik, its fortran
offering is still merely at the 77 level...

More information about the Beowulf mailing list