From: Alessandro Chessa Newsgroups: comp.parallel.mpi,comp.parallel.pvm Subject: Re: Dual- vs. single processor motherboards Date: Fri, 19 Mar 1999 19:27:27 +0100 Organization: Physics Department - University of Cagliari Message-Id: <36F2970F.1AFF3223@dsf.unica.it> References: <36EAA3B0.47FC0EB9@gci.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Xref: ukc comp.parallel.mpi:4784 comp.parallel.pvm:8163 Curt Timmerman wrote: > > I would definitely go with the dual Pentium motherboards. My testing was with > the PVM version of POVRAY on Pentium II's (350 MHz) and RedHat (upgraded to > linux 2.2.2). > > With a single process I could get 99%+ utilization on the CPU. Running 2 process > on the same node, both CPU's would also reach 99%+ utilization. The tests I ran > probably allowed the processors to run mostly from L2 cache. As data > requirements increases, main memory access would lower the utilization percents. and, what about dual port memory? We have, at physics department in cagliari (italy), a beowulf initiative named kalix (http://kalix.dsf.unica.it/). Now we are upgrading to a 16 bi-processor machine with a mixed network topology: 3 nics at each node, one versus a switch and two for a point-to-point connection on a ring (we have some physics problems in mind...). We were just considering to use one processor only for the communocations with the neighbors, while the other processor is working on the bulk calculations. The problem is: can both processors access at the some time the memory if the cache is not enough? Maybe be this is a trivial probelm for all of you... alessandro chessa