[Pw_forum] multi core Vs Single core

Paul M. Grant w2agz at pacbell.net
Sat Feb 9 21:23:29 CET 2008


For what it's worth, here're some anecdotal experiences I've had over the
last several days on an AMD64x2 HP Pavilion box I got real cheap.  I've
haven't installed Kubuntu yet, but running some of my benchmarks under
Cygwin/X (Win OS is Vista 32 bit), it seems "cpu intensive" tasks, like
gfortran compile and pwscf nscf run only in one core...(~50% cpu as shown by
Task Manager) with certain Vista "services" occasionally launching,
presumably in the other core, ramping the total to 100%.  BTW, the max
commitment of RAM is only 50% of the that available, and there is very
little disk activity.

It will be interesting to see how Kubuntu handles this.

Paul M. Grant, PhD
Principal, W2AGZ Technologies
Visiting Scholar, Applied Physics, Stanford University
EPRI Science Fellow (Retired)
IBM Research Staff Member Emeritus
w2agz at pacbell.net
http://www.w2agz.com
 
 


-----Original Message-----
From: pw_forum-bounces at pwscf.org [mailto:pw_forum-bounces at pwscf.org] On
Behalf Of Axel Kohlmeyer
Sent: Friday, February 08, 2008 9:34 AM
To: PWSCF Forum
Subject: Re: [Pw_forum] multi core Vs Single core

On Feb 8, 2008 12:10 PM, Marcel Mohr <marcel at physik.tu-berlin.de> wrote:
> There are some posts about this topic some time ago.
>
> Planewave codes use a lot of memory bandwith, so an Intel
> quadcore is not much faster than a dual core on the same mainboard,
> because the memory bus becomes the bottleneck of the system, and all core
> share the same bus.

for some additional numbers on multi-core. have a look at:
http://groups.google.com/group/cp2k/msg/aa1ce54b56598aab?dmode=source

FYI, CP2K/QuickStep is even more demanding in terms of memory bandwidth than
QE.

however, since some vendors are offering quad-core nodes for about the
same price
as dual-core nodes, it is actually a better deal to get quad-core and
then use only
half of the cores (and have twice the cache) if you get processor
affinity set up right.

a.
>
> Cheers
> Marcel
>
>
> On Fri, 8 Feb 2008, Axel Kohlmeyer wrote:
>
> > On Fri, 8 Feb 2008, ambavale sagar wrote:
> >
> >
> > sagar,
> >
> > AS> Is multi (quad) core processor more efficient than single core for
our
> >
> > please clarify what you mean by "efficient".
> > more throughput for the same money, faster execution,
> > better utilization of the cpu, less room needed for
> > a given computational capability?
> >
> > AS> abinitio calculations? How efficiently multi-threading is done while
> > AS> working with serial version of code on multicore machine?
> >
> > it is a common misconceptions that you need to use multi-threading
> > to take advantage of multicore cpus. the vendors of those cpus
> > are somewhat promoting it, because it is simpler to add multi-threading
> > to a code than to parallelize it completely, and also because
> > their new cpus don't have any advantage over the older ones without
> > paralleization. fortunately, the methods used in quantum espresso
> > are well parallelizable and the code offers multiple levels of MPI
> > parallelism which i found to be almost always more efficient than
> > using OpenMP (and thus multi-threading). even more so, QE does not
> > support OpenMP (only indirectly via threaded BLAS/LAPACK libraries).
> >
> > depending on your hardware, problem set size and compiler/library
> > combination, you can get a speedup of roughly 2.5x to 3.5x from
> > a quad core cpu over a corresponding single core. when using multiple
> > nodes via a network, however, the speed and the latency of the
> > network (and the options of parallization, i.e. if you can parallelize
> > over k-points or not) start to matter and can make it all very
> > confusing.
> >
> > in the end, apart from a few general observations, there is no other
> > way to get a definite answer than running tests with representative
> > input files.
> >
> > cheers,
> >   axel.
> >
> > p.s.: this brings up the question again, that we should compile a
> > list of such typical inputs and run them on mutually available
> > machines so people can see what to expect....
> >
> > AS>
> > AS> Thanx.
> > AS>
> > AS>
> > AS>
> > AS> Sagar K. Ambavale
> > AS>
> > AS> PhD student,
> > AS>
> > AS> The M.S. University of Baroda,
> > AS>
> > AS> India
> > AS>
> > AS>
> > AS>
____________________________________________________________________________
____
> > AS> 5, 50, 500, 5000 - Store N number of mails in your inbox. Click
here.
> > AS>
> >
> > --
> > =======================================================================
> > Axel Kohlmeyer   akohlmey at cmm.chem.upenn.edu   http://www.cmm.upenn.edu
> >   Center for Molecular Modeling   --   University of Pennsylvania
> > Department of Chemistry, 231 S.34th Street, Philadelphia, PA 19104-6323
> > tel: 1-215-898-1582,  fax: 1-215-573-6233,  office-tel: 1-215-898-5425
> > =======================================================================
> > If you make something idiot-proof, the universe creates a better idiot.
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://www.democritos.it/mailman/listinfo/pw_forum
>
>



-- 
=======================================================================
Axel Kohlmeyer   akohlmey at cmm.chem.upenn.edu   http://www.cmm.upenn.edu
  Center for Molecular Modeling   --   University of Pennsylvania
Department of Chemistry, 231 S.34th Street, Philadelphia, PA 19104-6323
tel: 1-215-898-1582,  fax: 1-215-573-6233,  office-tel: 1-215-898-5425
=======================================================================
If you make something idiot-proof, the universe creates a better idiot.
_______________________________________________
Pw_forum mailing list
Pw_forum at pwscf.org
http://www.democritos.it/mailman/listinfo/pw_forum




More information about the users mailing list