|
»
»
»
results from tuning LAPACK
|
View:
New views
5 Messages
—
Rating Filter:
Alert me
|
 |
results from tuning LAPACK
by Clint Whaley-2
::
Rate this Message:
Guys,
My paper on the benefits of empirically tuning LAPACK's ILAENV-controlled NB has been accepted for CANA'08. You can see a draft of it at: http://www.cs.utsa.edu/~whaley/papers/lanb.pdf It shows the advantage you get from tuning NB for several different BLAS an a couple of platforms. Depending on problem size and architecture, you can be leaving as much as 30% (or 75%, if on the Itanium :) on the table if you use stock lapack's ILAENV. It also highlights two new tools available in ATLAS (lapack timer/tuner and a tool that takes ATLAS timer output, does some simple statistics, and prints it out in a spreadsheet friendly format). We can apply this tuning to pretty much any blocked LAPACK routine. ATLAS presently handles tuning the factorizations. My guess is that the next most important routines are the eigencodes. Does anyone want to nominate particular routines that they use a lot? Cheers, Clint ************************************************************************** ** R. Clint Whaley, PhD ** Assist Prof, UTSA ** www.cs.utsa.edu/~whaley ** ************************************************************************** ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/ _______________________________________________ Math-atlas-devel mailing list Math-atlas-devel@... https://lists.sourceforge.net/lists/listinfo/math-atlas-devel |
|
|
Re: results from tuning LAPACK
by M. Edward (Ed) Borasky
::
Rate this Message:
On Sun, 2008-08-10 at 07:50 -0500, Clint Whaley wrote:
> Guys, > > My paper on the benefits of empirically tuning LAPACK's ILAENV-controlled NB > has been accepted for CANA'08. You can see a draft of it at: > http://www.cs.utsa.edu/~whaley/papers/lanb.pdf > > It shows the advantage you get from tuning NB for several different BLAS > an a couple of platforms. Depending on problem size and architecture, > you can be leaving as much as 30% (or 75%, if on the Itanium :) on the table > if you use stock lapack's ILAENV. It also highlights two new tools available > in ATLAS (lapack timer/tuner and a tool that takes ATLAS timer output, > does some simple statistics, and prints it out in a spreadsheet friendly > format). > > We can apply this tuning to pretty much any blocked LAPACK routine. ATLAS > presently handles tuning the factorizations. My guess is that the next > most important routines are the eigencodes. Does anyone want to nominate > particular routines that they use a lot? Real symmetric eigensolvers are a big workhorse in lots of areas ... even undergraduate chemistry. :) Singular Value Decomposition is another workhorse -- gotta have it. Question -- is anybody putting this much effort into tuning *sparse* linear algebra? Something approaching ATLAS performance on huge sparse matrix-vector multiplies would give significant improvements in many areas, including my personal favorite, Markov chains. > > Cheers, > Clint > > ************************************************************************** > ** R. Clint Whaley, PhD ** Assist Prof, UTSA ** www.cs.utsa.edu/~whaley ** > ************************************************************************** > > ------------------------------------------------------------------------- > This SF.Net email is sponsored by the Moblin Your Move Developer's challenge > Build the coolest Linux based applications with Moblin SDK & win great prizes > Grand prize is a trip for two to an Open Source event anywhere in the world > http://moblin-contest.org/redirect.php?banner_id=100&url=/ > _______________________________________________ > Math-atlas-devel mailing list > Math-atlas-devel@... > https://lists.sourceforge.net/lists/listinfo/math-atlas-devel M. Edward (Ed) Borasky ruby-perspectives.blogspot.com "A mathematician is a machine for turning coffee into theorems." -- Alfréd Rényi via Paul Erdős ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/ _______________________________________________ Math-atlas-devel mailing list Math-atlas-devel@... https://lists.sourceforge.net/lists/listinfo/math-atlas-devel |
|
|
Re: results from tuning LAPACK
by Dmitri A. Sergatskov
::
Rate this Message:
On Sun, Aug 10, 2008 at 10:58 AM, M. Edward (Ed) Borasky
<znmeb@...> wrote: > > Question -- is anybody putting this much effort into tuning *sparse* > linear algebra? Something approaching ATLAS performance on huge sparse > matrix-vector multiplies would give significant improvements in many > areas, including my personal favorite, Markov chains. Are you aware of Tim Davis' research: http://www.cise.ufl.edu/~davis/ ? In particular: http://www.cise.ufl.edu/research/sparse/ Dmitri. -- ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/ _______________________________________________ Math-atlas-devel mailing list Math-atlas-devel@... https://lists.sourceforge.net/lists/listinfo/math-atlas-devel |
|
|
Re: results from tuning LAPACK
by Clint Whaley-2
::
Rate this Message:
Ed,
>Question -- is anybody putting this much effort into tuning *sparse* >linear algebra? Something approaching ATLAS performance on huge sparse >matrix-vector multiplies would give significant improvements in many >areas, including my personal favorite, Markov chains. The main empirical tuning folks in this area that I know about are: http://bebop.cs.berkeley.edu/oski/ I haven't used it, so can't comment on how good it is . . . Obviously, your raw performance cannot be close, due to the storage and extra indirection costs (the exception being some of the blocked storage schemes, which can actually use the dense blas). So, I think that OSKI tries to be the ATLAS of sparse, with maybe PETSc http://www-unix.mcs.anl.gov/petsc/petsc-as/ being the LAPACK equivilant (though obviously there are a huge number of higher-level sparse packages) . . . Cheers, Clint ************************************************************************** ** R. Clint Whaley, PhD ** Assist Prof, UTSA ** www.cs.utsa.edu/~whaley ** ************************************************************************** ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/ _______________________________________________ Math-atlas-devel mailing list Math-atlas-devel@... https://lists.sourceforge.net/lists/listinfo/math-atlas-devel |
|
|
Re: results from tuning LAPACK
by Mikhail Kuzminsky
::
Rate this Message:
In message from "M. Edward (Ed) Borasky" <znmeb@...> (Sun, 10
Aug 2008 08:58:36 -0700): >On Sun, 2008-08-10 at 07:50 -0500, Clint Whaley wrote: >> Guys, >> >> My paper on the benefits of empirically tuning LAPACK's >>ILAENV-controlled NB >> has been accepted for CANA'08. You can see a draft of it at: >> http://www.cs.utsa.edu/~whaley/papers/lanb.pdf >> >> It shows the advantage you get from tuning NB for several different >>BLAS >> an a couple of platforms. Depending on problem size and >>architecture, >> you can be leaving as much as 30% (or 75%, if on the Itanium :) on >>the table >> if you use stock lapack's ILAENV. It also highlights two new tools >>available >> in ATLAS (lapack timer/tuner and a tool that takes ATLAS timer >>output, >> does some simple statistics, and prints it out in a spreadsheet >>friendly >> format). >> >> We can apply this tuning to pretty much any blocked LAPACK routine. >> ATLAS >> presently handles tuning the factorizations. My guess is that the >>next >> most important routines are the eigencodes. Does anyone want to >>nominate >> particular routines that they use a lot? > >Real symmetric eigensolvers are a big workhorse in lots of areas ... >even undergraduate chemistry. :) I want to support this proposal - about real (double precision :-)) symmetric matrix diagonalization. To be more exact for quantum chemistry, they solve often also generalized eigenvectors problem, i.e. Ax=lambda*Bx, where B is positive defined simmetric matrix. This task may be reduced to usual diagonalization. What is about sparse matrixes, in quantum chemistry, there is special LINEAR scaling algorithms which avoids usual diagonalization. Mikhail > Singular Value Decomposition is >another >workhorse -- gotta have it. > >Question -- is anybody putting this much effort into tuning *sparse* >linear algebra? Something approaching ATLAS performance on huge >sparse >matrix-vector multiplies would give significant improvements in many >areas, including my personal favorite, Markov chains. >> >> Cheers, >> Clint >> >> ************************************************************************** >> ** R. Clint Whaley, PhD ** Assist Prof, UTSA ** >>www.cs.utsa.edu/~whaley ** >> ************************************************************************** >> >> ------------------------------------------------------------------------- >> This SF.Net email is sponsored by the Moblin Your Move Developer's >>challenge >> Build the coolest Linux based applications with Moblin SDK & win >>great prizes >> Grand prize is a trip for two to an Open Source event anywhere in >>the world >> http://moblin-contest.org/redirect.php?banner_id=100&url=/ >> _______________________________________________ >> Math-atlas-devel mailing list >> Math-atlas-devel@... >> https://lists.sourceforge.net/lists/listinfo/math-atlas-devel >-- >M. Edward (Ed) Borasky >ruby-perspectives.blogspot.com > >"A mathematician is a machine for turning coffee into theorems." -- >Alfréd Rényi via Paul Erdős > > >------------------------------------------------------------------------- >This SF.Net email is sponsored by the Moblin Your Move Developer's >challenge >Build the coolest Linux based applications with Moblin SDK & win >great prizes >Grand prize is a trip for two to an Open Source event anywhere in the >world >http://moblin-contest.org/redirect.php?banner_id=100&url=/ >_______________________________________________ >Math-atlas-devel mailing list >Math-atlas-devel@... >https://lists.sourceforge.net/lists/listinfo/math-atlas-devel ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/ _______________________________________________ Math-atlas-devel mailing list Math-atlas-devel@... https://lists.sourceforge.net/lists/listinfo/math-atlas-devel |
| Free embeddable forum powered by Nabble | Forum Help |
