floraachy/OpenBLAS
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Fix BLAS and LAPACK tests for C910V and RISCV64_ZVL256B targets

Browse Source 
* Fixed bugs in dgemm, [a]min\max, asum kernels
* Added zero checks for BLAS kernels
* Added dsdot implementation for RVV 0.7.1
* Fixed bugs in _vector files for C910V and RISCV64_ZVL256B targets
* Added additional definitions for RISCV64_ZVL256B target
This commit is contained in:
kseniyazaytseva
2023-04-07 11:13:23 +03:00
committed by Andrey Sokolov
parent 88e994116c
commit e1afb23811
19 changed files with 205 additions and 38 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
kernel/riscv64/dgemm_kernel_8x4_c910v.c
View File

@@ -196,7 +196,7 @@ int CNAME(BLASLONG bm,BLASLONG bn,BLASLONG bk,FLOAT alpha,FLOAT* ba,FLOAT* bb,FL
asm volatile(
"vsetvli zero, zero, e64,m1 \n\t"
"fmv.w.x ft11, zero \n\t"
"fmv.d.x ft11, zero \n\t"
"mv t0, %[BK] \n\t"
"vfmv.v.f v16, ft11 \n\t"