#libxsmm — Public Fediverse posts
Live and recent posts from across the Fediverse tagged #libxsmm, aggregated by home.social.
-
Amazing work from Sarah El-Kazdadi. #LibXSMM has become standard for applications needing small, dense matrix multiply/tensor contraction. It uses JIT, which was widely believed to be necessary to achieve high performance in this domain. Sarah's new library, #nanogemm, is competitive or better without JIT (modulo a caveat about padding). #Rust #HPC #GEMM
-
Amazing work from Sarah El-Kazdadi. #LibXSMM has become standard for applications needing small, dense matrix multiply/tensor contraction. It uses JIT, which was widely believed to be necessary to achieve high performance in this domain. Sarah's new library, #nanogemm, is competitive or better without JIT (modulo a caveat about padding). #Rust #HPC #GEMM
-
Amazing work from Sarah El-Kazdadi. #LibXSMM has become standard for applications needing small, dense matrix multiply/tensor contraction. It uses JIT, which was widely believed to be necessary to achieve high performance in this domain. Sarah's new library, #nanogemm, is competitive or better without JIT (modulo a caveat about padding). #Rust #HPC #GEMM
-
Amazing work from Sarah El-Kazdadi. #LibXSMM has become standard for applications needing small, dense matrix multiply/tensor contraction. It uses JIT, which was widely believed to be necessary to achieve high performance in this domain. Sarah's new library, #nanogemm, is competitive or better without JIT (modulo a caveat about padding). #Rust #HPC #GEMM
-
Amazing work from Sarah El-Kazdadi. #LibXSMM has become standard for applications needing small, dense matrix multiply/tensor contraction. It uses JIT, which was widely believed to be necessary to achieve high performance in this domain. Sarah's new library, #nanogemm, is competitive or better without JIT (modulo a caveat about padding). #Rust #HPC #GEMM