Re: [eigen] Cannot compile benchmark

[ Thread Index | Date Index | More lists.tuxfamily.org/eigen Archives ]


Hi!

Some time ago I sent an email

    http://sourceforge.net/mailarchive/message.php?msg_id=28711667

to the ATLAS and GotoBLAS/OpenBLAS lists regarding the Eigen benchmark
results.  I am kind of surprised that this topic did not lead to a longer
discussion.  The main issue seems to be that the benchmarks are not taken
seriously.  But what Clint replied sounds like a really good idea to me:

"... then I would say the thing to do is to install them and ATLAS, and
use ATLAS's benchmarking tools to compare the two on as many platforms as
interest you.  As long as they provide a standard BLAS and LAPACK API, you
can tell ATLAS to build a full timing report comparing the two
implementations directly."

A benchmark test that merely requires that library provide a BLAS interface
is much more transparent (and therefore more trustworthy).

Cheers,

Michael


Am 09.06.2012 um 14:30 schrieb Ilja Honkonen:

This has been fixed yesterday in changeset 73aa01cd5245.

Thanks, I ran the benchmarks with eigen 4646:e9efce779ca4, pdfs attached. GCC: 4.5.2, CPU: AMD Phenom(tm) II X6 1075T 3 GHz, GotoBLAS: 2-1.13. In some of my results Goto is significantly faster than on the benchmark page http://eigen.tuxfamily.org/index.php?title=Benchmark
Eigen results are about the same.

Ilja
<complete_lu_decomp.pdf><rot.pdf><ger.pdf><syr2.pdf><symv.pdf><hessenberg.pdf><tridiagonalization.pdf><partial_lu_decomp.pdf><cholesky.pdf><trisolve_matrix.pdf><trisolve_vector.pdf><trmm.pdf><aat.pdf><matrix_matrix.pdf><atv.pdf><matrix_vector.pdf><axpby.pdf><axpy.pdf>



Mail converted by MHonArc 2.6.19+ http://listengine.tuxfamily.org/