OSchip
/
OpenBLAS
Not watched
1
0
Fork 0
Code
Releases 66 Wiki evaluate Activity
Issues 0 Pull Requests 0 Datasets Model Cloudbrain HPC
Browse Source

Fix xtheadvector compilation

pull/5393/head
xctan 6 months ago
parent
4a94ef57e0
commit
6144004e92
3 changed files with 64 additions and 70 deletions
Split View
Diff Options
Show Stats Download Patch File Download Diff File
  1. +1
    -1
      common_macro.h
  2. +0
    -6
      common_riscv64.h
  3. +63
    -63
      kernel/riscv64/dsdot_vector.c

+ 1
- 1
common_macro.h View File

@@ -2709,7 +2709,7 @@
#ifndef ASSEMBLER
#if !defined(DYNAMIC_ARCH) \
&& (defined(ARCH_X86) || defined(ARCH_X86_64) || defined(ARCH_IA64) || defined(ARCH_MIPS64) || defined(ARCH_ARM64) \
|| defined(ARCH_LOONGARCH64) || defined(ARCH_E2K) || defined(ARCH_ALPHA))
|| defined(ARCH_LOONGARCH64) || defined(ARCH_E2K) || defined(ARCH_ALPHA) || defined(ARCH_RISCV64))
extern BLASLONG gemm_offset_a;
extern BLASLONG gemm_offset_b;
extern BLASLONG bgemm_p;


+ 0
- 6
common_riscv64.h View File

@@ -93,13 +93,7 @@ static inline int blas_quickdivide(blasint x, blasint y){
# include <riscv_vector.h>
#endif

#if defined( __riscv_xtheadc ) && defined( __riscv_v ) && ( __riscv_v <= 7000 )
// t-head toolchain uses obsolete rvv intrinsics, can't build for C910V without this
#define RISCV_0p10_INTRINSICS
#define RISCV_RVV(x) x
#else
#define RISCV_RVV(x) __riscv_ ## x
#endif

#if defined(C910V) || defined(RISCV64_ZVL256B)
# if !defined(DOUBLE)


+ 63
- 63
kernel/riscv64/dsdot_vector.c View File

@@ -37,114 +37,114 @@ double CNAME(BLASLONG n, FLOAT *x, BLASLONG inc_x, FLOAT *y, BLASLONG inc_y)
vfloat32m2_t vx, vy;
unsigned int gvl = 0;
vfloat64m1_t v_res, v_z0;
gvl = vsetvlmax_e64m1();
v_res = vfmv_v_f_f64m1(0, gvl);
v_z0 = vfmv_v_f_f64m1(0, gvl);
gvl = __riscv_vsetvlmax_e64m1();
v_res = __riscv_vfmv_v_f_f64m1(0, gvl);
v_z0 = __riscv_vfmv_v_f_f64m1(0, gvl);

if(inc_x == 1 && inc_y == 1){
gvl = vsetvl_e64m4(n);
vr = vfmv_v_f_f64m4(0, gvl);
gvl = __riscv_vsetvl_e64m4(n);
vr = __riscv_vfmv_v_f_f64m4(0, gvl);
for(i=0,j=0; i<n/gvl; i++){
vx = vle32_v_f32m2(&x[j], gvl);
vy = vle32_v_f32m2(&y[j], gvl);
vr = vfwmacc_vv_f64m4(vr, vx, vy, gvl);
vx = __riscv_vle32_v_f32m2(&x[j], gvl);
vy = __riscv_vle32_v_f32m2(&y[j], gvl);
vr = __riscv_vfwmacc_vv_f64m4(vr, vx, vy, gvl);
j += gvl;
}
if(j > 0){
v_res = vfredusum_vs_f64m4_f64m1(v_res, vr, v_z0, gvl);
dot += (double)vfmv_f_s_f64m1_f64(v_res);
v_res = __riscv_vfredusum_vs_f64m4_f64m1(vr, v_z0, gvl);
dot += (double)__riscv_vfmv_f_s_f64m1_f64(v_res);
}
//tail
if(j < n){
gvl = vsetvl_e64m4(n-j);
vx = vle32_v_f32m2(&x[j], gvl);
vy = vle32_v_f32m2(&y[j], gvl);
vfloat64m4_t vz = vfmv_v_f_f64m4(0, gvl);
//vr = vfdot_vv_f32m2(vx, vy, gvl);
vr = vfwmacc_vv_f64m4(vz, vx, vy, gvl);
v_res = vfredusum_vs_f64m4_f64m1(v_res, vr, v_z0, gvl);
dot += (double)vfmv_f_s_f64m1_f64(v_res);
gvl = __riscv_vsetvl_e64m4(n-j);
vx = __riscv_vle32_v_f32m2(&x[j], gvl);
vy = __riscv_vle32_v_f32m2(&y[j], gvl);
vfloat64m4_t vz = __riscv_vfmv_v_f_f64m4(0, gvl);
//vr = __riscv_vfdot_vv_f32m2(vx, vy, gvl);
vr = __riscv_vfwmacc_vv_f64m4(vz, vx, vy, gvl);
v_res = __riscv_vfredusum_vs_f64m4_f64m1(vr, v_z0, gvl);
dot += (double)__riscv_vfmv_f_s_f64m1_f64(v_res);
}
}else if(inc_y == 1){
gvl = vsetvl_e64m4(n);
vr = vfmv_v_f_f64m4(0, gvl);
gvl = __riscv_vsetvl_e64m4(n);
vr = __riscv_vfmv_v_f_f64m4(0, gvl);
int stride_x = inc_x * sizeof(FLOAT);
for(i=0,j=0; i<n/gvl; i++){
vx = vlse32_v_f32m2(&x[j*inc_x], stride_x, gvl);
vy = vle32_v_f32m2(&y[j], gvl);
vr = vfwmacc_vv_f64m4(vr, vx, vy, gvl);
vx = __riscv_vlse32_v_f32m2(&x[j*inc_x], stride_x, gvl);
vy = __riscv_vle32_v_f32m2(&y[j], gvl);
vr = __riscv_vfwmacc_vv_f64m4(vr, vx, vy, gvl);
j += gvl;
}
if(j > 0){
v_res = vfredusum_vs_f64m4_f64m1(v_res, vr, v_z0, gvl);
dot += (double)vfmv_f_s_f64m1_f64(v_res);
v_res = __riscv_vfredusum_vs_f64m4_f64m1(vr, v_z0, gvl);
dot += (double)__riscv_vfmv_f_s_f64m1_f64(v_res);

}
//tail
if(j < n){
gvl = vsetvl_e64m4(n-j);
vx = vlse32_v_f32m2(&x[j*inc_x], stride_x, gvl);
vy = vle32_v_f32m2(&y[j], gvl);
vfloat64m4_t vz = vfmv_v_f_f64m4(0, gvl);
//vr = vfdot_vv_f32m2(vx, vy, gvl);
vr = vfwmacc_vv_f64m4(vz, vx, vy, gvl);
v_res = vfredusum_vs_f64m4_f64m1(v_res, vr, v_z0, gvl);
dot += (double)vfmv_f_s_f64m1_f64(v_res);
gvl = __riscv_vsetvl_e64m4(n-j);
vx = __riscv_vlse32_v_f32m2(&x[j*inc_x], stride_x, gvl);
vy = __riscv_vle32_v_f32m2(&y[j], gvl);
vfloat64m4_t vz = __riscv_vfmv_v_f_f64m4(0, gvl);
//vr = __riscv_vfdot_vv_f32m2(vx, vy, gvl);
vr = __riscv_vfwmacc_vv_f64m4(vz, vx, vy, gvl);
v_res = __riscv_vfredusum_vs_f64m4_f64m1(vr, v_z0, gvl);
dot += (double)__riscv_vfmv_f_s_f64m1_f64(v_res);

}
}else if(inc_x == 1){
gvl = vsetvl_e64m4(n);
vr = vfmv_v_f_f64m4(0, gvl);
gvl = __riscv_vsetvl_e64m4(n);
vr = __riscv_vfmv_v_f_f64m4(0, gvl);
int stride_y = inc_y * sizeof(FLOAT);
for(i=0,j=0; i<n/gvl; i++){
vx = vle32_v_f32m2(&x[j], gvl);
vy = vlse32_v_f32m2(&y[j*inc_y], stride_y, gvl);
vr = vfwmacc_vv_f64m4(vr, vx, vy, gvl);
vx = __riscv_vle32_v_f32m2(&x[j], gvl);
vy = __riscv_vlse32_v_f32m2(&y[j*inc_y], stride_y, gvl);
vr = __riscv_vfwmacc_vv_f64m4(vr, vx, vy, gvl);
j += gvl;
}
if(j > 0){
v_res = vfredusum_vs_f64m4_f64m1(v_res, vr, v_z0, gvl);
dot += (double)vfmv_f_s_f64m1_f64(v_res);
v_res = __riscv_vfredusum_vs_f64m4_f64m1(vr, v_z0, gvl);
dot += (double)__riscv_vfmv_f_s_f64m1_f64(v_res);

}
//tail
if(j < n){
gvl = vsetvl_e64m4(n-j);
vx = vle32_v_f32m2(&x[j], gvl);
vy = vlse32_v_f32m2(&y[j*inc_y], stride_y, gvl);
vfloat64m4_t vz = vfmv_v_f_f64m4(0, gvl);
//vr = vfdot_vv_f32m2(vx, vy, gvl);
vr = vfwmacc_vv_f64m4(vz, vx, vy, gvl);
v_res = vfredusum_vs_f64m4_f64m1(v_res, vr, v_z0, gvl);
dot += (double)vfmv_f_s_f64m1_f64(v_res);
gvl = __riscv_vsetvl_e64m4(n-j);
vx = __riscv_vle32_v_f32m2(&x[j], gvl);
vy = __riscv_vlse32_v_f32m2(&y[j*inc_y], stride_y, gvl);
vfloat64m4_t vz = __riscv_vfmv_v_f_f64m4(0, gvl);
//vr = __riscv_vfdot_vv_f32m2(vx, vy, gvl);
vr = __riscv_vfwmacc_vv_f64m4(vz, vx, vy, gvl);
v_res = __riscv_vfredusum_vs_f64m4_f64m1(vr, v_z0, gvl);
dot += (double)__riscv_vfmv_f_s_f64m1_f64(v_res);

}
}else{
gvl = vsetvl_e64m4(n);
vr = vfmv_v_f_f64m4(0, gvl);
gvl = __riscv_vsetvl_e64m4(n);
vr = __riscv_vfmv_v_f_f64m4(0, gvl);
int stride_x = inc_x * sizeof(FLOAT);
int stride_y = inc_y * sizeof(FLOAT);
for(i=0,j=0; i<n/gvl; i++){
vx = vlse32_v_f32m2(&x[j*inc_x], stride_x, gvl);
vy = vlse32_v_f32m2(&y[j*inc_y], stride_y, gvl);
vr = vfwmacc_vv_f64m4(vr, vx, vy, gvl);
vx = __riscv_vlse32_v_f32m2(&x[j*inc_x], stride_x, gvl);
vy = __riscv_vlse32_v_f32m2(&y[j*inc_y], stride_y, gvl);
vr = __riscv_vfwmacc_vv_f64m4(vr, vx, vy, gvl);
j += gvl;
}
if(j > 0){
v_res = vfredusum_vs_f64m4_f64m1(v_res, vr, v_z0, gvl);
dot += (double)vfmv_f_s_f64m1_f64(v_res);
v_res = __riscv_vfredusum_vs_f64m4_f64m1(vr, v_z0, gvl);
dot += (double)__riscv_vfmv_f_s_f64m1_f64(v_res);

}
//tail
if(j < n){
gvl = vsetvl_e64m4(n-j);
vx = vlse32_v_f32m2(&x[j*inc_x], stride_x, gvl);
vy = vlse32_v_f32m2(&y[j*inc_y], stride_y, gvl);
vfloat64m4_t vz = vfmv_v_f_f64m4(0, gvl);
//vr = vfdot_vv_f32m2(vx, vy, gvl);
vr = vfwmacc_vv_f64m4(vz, vx, vy, gvl);
v_res = vfredusum_vs_f64m4_f64m1(v_res, vr, v_z0, gvl);
dot += (double)vfmv_f_s_f64m1_f64(v_res);
gvl = __riscv_vsetvl_e64m4(n-j);
vx = __riscv_vlse32_v_f32m2(&x[j*inc_x], stride_x, gvl);
vy = __riscv_vlse32_v_f32m2(&y[j*inc_y], stride_y, gvl);
vfloat64m4_t vz = __riscv_vfmv_v_f_f64m4(0, gvl);
//vr = __riscv_vfdot_vv_f32m2(vx, vy, gvl);
vr = __riscv_vfwmacc_vv_f64m4(vz, vx, vy, gvl);
v_res = __riscv_vfredusum_vs_f64m4_f64m1(vr, v_z0, gvl);
dot += (double)__riscv_vfmv_f_s_f64m1_f64(v_res);

}
}


Write Preview
Loading…
Cancel
Save