Commit d69be17
committed
Convert 2X LMUL1 instructions to 1X LMUL2. Improved FP64 GEMM edges - up to more than 3X faster.
1 parent 8fc0004 commit d69be17
2 files changed
Lines changed: 2215 additions & 1249 deletions
1 parent 8fc0004 commit d69be17
2 files changed
0 commit comments