News for hackers. Stuff that matters.

A comment on Matrix multiplication inches closer to mythic goal. Parent comment

Re:Matrix multiplication inches closer to mythic goal by DesiLurker on Tuesday March 23, 2021 @01:45PM (#26557860)
Can the fortran implementations do auto-vectorization on simd hardware? or map to gpu using cuda-like library?

not rhetorical, genuine question.

[ Parent | on HN ]
Re:Matrix multiplication inches closer to mythic goal by cygx on Tuesday March 23, 2021 @01:53PM (#26557965)
Yes. See e.g.

https://gcc.gnu.org/projects/tree-ssa/vectorization.html#exa...

https://developer.nvidia.com/cuda-fortran

[ Parent | on HN ]
Re:Matrix multiplication inches closer to mythic goal by TheRealKing on Wednesday March 24, 2021 @12:18PM (#26569227)
Fortran is a fully-vectorized and parallelized language at all levels, from instructions to shared and distributed parallelism (corresponding to one-sided MPI communications). This has been the case for at least 3 decades if not more. There are mature GPU support for Fortran. NVIDIA and PGI are currently implementing the standard parallel features of Fortran via GPUs.
[ Parent | on HN ]
Re:Matrix multiplication inches closer to mythic goal by jhayward on Tuesday March 23, 2021 @05:33PM (#26560674)
Yes, in fact all the original research for those things was done in Fortran compilers.
[ Parent | on HN ]