[gmx-users] hyperthreading

[David van der Spoel]

On Thu, 2004-01-29 at 19:31, Anton Feenstra wrote:
> IMHO, hyperthreading sounds like a dirty fix to a stupid problem that any
> decent operating system should be able to solve (and solve it better!).
yes and no,

since the P4 has such long pipelines, a stall due to, e.g., a cache miss
is expensive, if you have another process waiting for the CPU it can use
this stall time. Some benchmarks, like database/web server benchmarks do
show improved performance by 10-15%. Doesn't work for gromacs though.
However, Intel's plan is to put two chips on a chip, which will appear
to the OS as the same thing that is already here and now with
hyperthreading. Other vendors already have such chips AFAIK.
-- 
David.
________________________________________________________________________
David van der Spoel, PhD, Assist. Prof., Molecular Biophysics group,
Dept. of Cell and Molecular Biology, Uppsala University.
Husargatan 3, Box 596,  	75124 Uppsala, Sweden
phone:	46 18 471 4205		fax: 46 18 511 755
spoel at xray.bmc.uu.se	spoel at gromacs.org   http://xray.bmc.uu.se/~spoel
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++