More work on Skylake-X target with AVX512 ISA

giaf / blasfeo

Basic linear algebra subroutines for embedded optimization

Other

308 stars 84 forks source link

Closed giaf closed 3 years ago

giaf commented 3 years ago

Some more work on Skylake target with AVX512 ISA:

add panel-major routines needed in HPIPM robust mode (e.g. various kinds of tailored LQ factorizations)
experiment with 24x8 kernels on selected routines
general bug fixes and performance improvements