Skip to content

ggml : fix unmerged GGML_FPxx_TO_FPxx refactoring #14443

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
Jun 29, 2025
Merged

ggml : fix unmerged GGML_FPxx_TO_FPxx refactoring #14443

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
18 changes: 9 additions & 9 deletions ggml/src/ggml-cpu/vec.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -913,8 +913,8 @@ inline static void ggml_vec_reglu_f32 (const int n, float * y, const float * x,

inline static void ggml_vec_reglu_f16 (const int n, ggml_fp16_t * y, const ggml_fp16_t * x, const ggml_fp16_t * g) {
for (int i = 0; i < n; ++i) {
float v = GGML_FP16_TO_FP32(x[i]);
y[i] = GGML_FP32_TO_FP16((v > 0.f) ? v * GGML_FP16_TO_FP32(g[i]) : 0.f);
float v = GGML_CPU_FP16_TO_FP32(x[i]);
y[i] = GGML_CPU_FP32_TO_FP16((v > 0.f) ? v * GGML_CPU_FP16_TO_FP32(g[i]) : 0.f);
}
}

Expand All @@ -927,9 +927,9 @@ inline static void ggml_vec_geglu_f32(const int n, float * y, const float * x, c
} else if (x[i] >= 10.0f) {
y[i] = x[i] * g[i];
} else {
ggml_fp16_t fp16 = GGML_FP32_TO_FP16(x[i]);
ggml_fp16_t fp16 = GGML_CPU_FP32_TO_FP16(x[i]);
memcpy(&t, &fp16, sizeof(uint16_t));
y[i] = GGML_FP16_TO_FP32(ggml_table_gelu_f16[t]) * g[i];
y[i] = GGML_CPU_FP16_TO_FP32(ggml_table_gelu_f16[t]) * g[i];
}
}
}
Expand All @@ -944,18 +944,18 @@ inline static void ggml_vec_geglu_f32(const int n, float * y, const float * x, c
inline static void ggml_vec_geglu_f16(const int n, ggml_fp16_t * y, const ggml_fp16_t * x, const ggml_fp16_t * g) {
const uint16_t * i16 = (const uint16_t *) x;
for (int i = 0; i < n; ++i) {
float v = GGML_FP16_TO_FP32(g[i]);
y[i] = GGML_FP32_TO_FP16(GGML_FP16_TO_FP32(ggml_table_gelu_f16[i16[i]]) * v);
float v = GGML_CPU_FP16_TO_FP32(g[i]);
y[i] = GGML_CPU_FP32_TO_FP16(GGML_CPU_FP16_TO_FP32(ggml_table_gelu_f16[i16[i]]) * v);
}
}

void ggml_vec_swiglu_f32(const int n, float * y, const float * x, const float * g);

inline static void ggml_vec_swiglu_f16(const int n, ggml_fp16_t * y, const ggml_fp16_t * x, const ggml_fp16_t * g) {
for (int i = 0; i < n; ++i) {
float v = GGML_FP16_TO_FP32(x[i]);
float w = GGML_FP16_TO_FP32(g[i]);
y[i] = GGML_FP32_TO_FP16((v/(1.0f + expf(-v))) * w);
float v = GGML_CPU_FP16_TO_FP32(x[i]);
float w = GGML_CPU_FP16_TO_FP32(g[i]);
y[i] = GGML_CPU_FP32_TO_FP16((v/(1.0f + expf(-v))) * w);
}
}

Expand Down
Loading