On 12/15/18 7:49 AM, Toon Moene wrote:
It is probably helpful to ask this question on the GNU Fortran mailing
list (copied here), as more Fortran developers are listening there ...
Two questions off-hand:
1. Do you have a small test case that shows the problem ?
2. Is this single- or multi-threaded code ?
On 12/13/18 10:31 PM, Harald Anlauf wrote:
Hello,
is there a public gcc benchmark that checks for performance
regressions in particular for sparse-matrix vector multiplication?
I am sure we can just pick a routine out of netlib and create a benchmark.
I've noticed (and confirmed by perf measurements) a significant
performance drop on x86_64/Skylake from gfortran-8 to gfortran-9
(current trunk version) for a code that uses both CSC and CSR
storage. gfortran-8 performance is the same as gfortran-7.
Thanks,
Harald
Can you give us the dimensions of the sparse matrix and its basic
arrangement? Is it randomly filled with zeroes or is it just several
main diagonals?
Same for the vector?
Also, are you using builtin matmul or is it using a custom or roll your
own algorithm?
Jerry
