[WIP] Add ARMv9.2 SME GEMM kernels ported from vlovero's project - #5971

Open

martin-frbg wants to merge 18 commits into

OpenMathLib:developfrom

martin-frbg:issue5841

Commits on Aug 11, 2026

Credit Vincent Lovero for his ARM SME kernel work
martin-frbg
authored
5aa2c41
Add f64f64 extension to VortexM4 options
martin-frbg
authored
2f915c5
Integrate SME GEMM kernels
martin-frbg
authored
f2dc747
Add SME GEMM kernels ported from vlovero's ARMv9.2-GEMM project
martin-frbg
authored
78f0321
Clean up non-OpenMP build and add CBLAS GEMM benchmark
martin-frbg
authored
cf1b7c1
Add ARM64 SME GEMM kernels
martin-frbg
authored
39b56e9
Add sme-f64f64 to ARMV9SME archflags too
martin-frbg
authored
e24e0da

Commits on Aug 12, 2026

Add SME GEMM kernels
martin-frbg
authored
1ed9981
Make the compute kernel static
martin-frbg
authored
4ae369a

Commits on Aug 13, 2026

Add declarations for ARM64 SME GEMM kernels
martin-frbg
authored
9724481
Add ARM64 SME GEMM kernels
martin-frbg
authored
c073f08
Improve clobber lists and interfaces
martin-frbg
authored
3ead57f
Move declarations of the ARM64 SME kernels to the appropriate headers
martin-frbg
authored
81dd859
Add sme-f64f64 capability to VortexM4 and ARMV9SME build flags
martin-frbg
authored
f8830b6
Add +sme-f64f64 to build flags of VortexM4 and ARMV9SME
martin-frbg
authored
305bd67
Merge branch 'develop' into issue5841
martin-frbg
authored
f285f47
Add casts to pacify homebrew-llvm
martin-frbg
authored
9884c48

Commits on Aug 14, 2026

Merge branch 'OpenMathLib:develop' into issue5841
martin-frbg
authored
17d9974