CS348Project/llama.cpp
5
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-11-13 10:57:15 +00:00
Code Issues Packages Projects Releases Wiki Activity

rm __ARM_FEATURE_SVE

Browse Source
This commit is contained in:
Xuan Son Nguyen
2025-07-09 11:56:48 +02:00
parent 563aca0b56
commit 50c678f6da
1 changed files with 16 additions and 46 deletions
Download Patch File Download Diff File Expand all files Collapse all files

62
ggml/src/ggml-cpu/vec.h
View File

@@ -355,57 +355,27 @@ inline static void ggml_vec_mad1_f32(const int n, float * y, const float s, cons
#if defined(GGML_USE_ACCELERATE) #if defined(GGML_USE_ACCELERATE)
vDSP_vsmsa(y, 1, &s, &b, y, 1, n); vDSP_vsmsa(y, 1, &s, &b, y, 1, n);
#elif defined(GGML_SIMD) #elif defined(GGML_SIMD)
#if defined(__ARM_FEATURE_SVE) // TODO: #if defined(__ARM_FEATURE_SVE)
const int sve_register_length = ggml_cpu_get_sve_cnt() * 8; const int np = (n & ~(GGML_F32_STEP - 1));
const int ggml_f32_epr = sve_register_length / 32;//8;//svcntw(); // SVE128:4, SVE256:8, SVE512:16
const int ggml_f32_step = 2 * ggml_f32_epr;
GGML_F32_VEC vs = GGML_F32_VEC_SET1(s); GGML_F32_VEC vs = GGML_F32_VEC_SET1(s);
GGML_F32_VEC vb = GGML_F32_VEC_SET1(b); GGML_F32_VEC vb = GGML_F32_VEC_SET1(b);
const int np = (n & ~(ggml_f32_step - 1)); GGML_F32_VEC ay[GGML_F32_ARR];
svfloat32_t ay1;
svfloat32_t ay2;
for (int i = 0; i < np; i += ggml_f32_step) {
ay1 = GGML_F32_VEC_LOAD(y + i);
ay1 = GGML_F32_VEC_FMA(ay1, vs, vb);
GGML_F32_VEC_STORE(y + i, ay1);
ay2 = GGML_F32_VEC_LOAD(y + i + 1*ggml_f32_epr); for (int i = 0; i < np; i += GGML_F32_STEP) {
ay2 = GGML_F32_VEC_FMA(ay2, vs, vb); for (int j = 0; j < GGML_F32_ARR; j++) {
GGML_F32_VEC_STORE(y + i + 1*ggml_f32_epr, ay2); ay[j] = GGML_F32_VEC_LOAD(y + i + j*GGML_F32_EPR);
ay[j] = GGML_F32_VEC_FMA(ay[j], vs, vb);
GGML_F32_VEC_STORE(y + i + j*GGML_F32_EPR, ay[j]);
} }
// leftovers }
// maximum number of leftover elements will be less that ggml_f32_epr. Apply predicated svmad on available elements only
if (np < n) {
svbool_t pg = svwhilelt_b32(np, n);
ay1 = svld1_f32(pg, y + np);
ay1 = svmul_f32_m(pg, ay1, vs);
ay1 = svadd_f32_m(pg, ay1, vb);
svst1_f32(pg, y + np, ay1);
}
#else
const int np = (n & ~(GGML_F32_STEP - 1));
GGML_F32_VEC vs = GGML_F32_VEC_SET1(s); // leftovers
GGML_F32_VEC vb = GGML_F32_VEC_SET1(b); for (int i = np; i < n; ++i) {
y[i] = y[i]*s + b;
GGML_F32_VEC ay[GGML_F32_ARR]; }
for (int i = 0; i < np; i += GGML_F32_STEP) {
for (int j = 0; j < GGML_F32_ARR; j++) {
ay[j] = GGML_F32_VEC_LOAD(y + i + j*GGML_F32_EPR);
ay[j] = GGML_F32_VEC_FMA(ay[j], vs, vb);
GGML_F32_VEC_STORE(y + i + j*GGML_F32_EPR, ay[j]);
}
}
// leftovers
for (int i = np; i < n; ++i) {
y[i] = y[i]*s + b;
}
#endif
#else #else
// scalar // scalar
for (int i = 0; i < n; ++i) { for (int i = 0; i < n; ++i) {