Difference between revisions of "User:Everton"

From Eigen
Jump to: navigation, search
(Created page with "* New Power 10 MMA Initial support for Power 10 matrix multiplication assist instructions for float32, float64 real and complex. * Altivec/Power improvements General perfor...")
 
Line 3: Line 3:
  
 
* Altivec/Power improvements  
 
* Altivec/Power improvements  
General performance improvement and bugfixes, better vectorization. Move matrix multiplication to new packet generic rank update functions which simulate Power 10's MMA instructions. New packing procedure.
+
General performance improvement and bugfixes, better vectorization. Changes to the gebp_kernel specific to Altivec, using VSX implementation of the MMA functions to gain speed improvements up to 4x.

Revision as of 18:24, 17 August 2021

  • New Power 10 MMA

Initial support for Power 10 matrix multiplication assist instructions for float32, float64 real and complex.

  • Altivec/Power improvements

General performance improvement and bugfixes, better vectorization. Changes to the gebp_kernel specific to Altivec, using VSX implementation of the MMA functions to gain speed improvements up to 4x.