root/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggerganov/llama.cpp.git synced 2024-12-26 11:24:35 +00:00
Code Issues Actions 2 Packages Projects Releases Wiki Activity

ggml : prefix lookup tables with ggml_

Browse Source 
ggml-ci
This commit is contained in:
Georgi Gerganov 2023-10-30 18:38:11 +02:00
parent 1039a16ce2
commit d70917f4b2
No known key found for this signature in database
GPG Key ID: 449E073F9DC10735
2 changed files with 25 additions and 24 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
ggml-impl.h
View File

@ -207,7 +207,8 @@ static inline ggml_fp16_t ggml_compute_fp32_to_fp16(float f) {
#endif // __ARM_NEON #endif // __ARM_NEON
// precomputed f32 table for f16 (256 KB) // precomputed f32 table for f16 (256 KB)
extern float table_f32_f16[1 << 16]; // defined in ggml.c, initialized in ggml_init()
extern float ggml_table_f32_f16[1 << 16];
// On ARM NEON, it's quicker to directly convert x -> x instead of calling into ggml_lookup_fp16_to_fp32, // On ARM NEON, it's quicker to directly convert x -> x instead of calling into ggml_lookup_fp16_to_fp32,
// so we define GGML_FP16_TO_FP32 and GGML_FP32_TO_FP16 elsewhere for NEON. // so we define GGML_FP16_TO_FP32 and GGML_FP32_TO_FP16 elsewhere for NEON.
@ -217,7 +218,7 @@ extern float table_f32_f16[1 << 16];
inline static float ggml_lookup_fp16_to_fp32(ggml_fp16_t f) { inline static float ggml_lookup_fp16_to_fp32(ggml_fp16_t f) {
uint16_t s; uint16_t s;
memcpy(&s, &f, sizeof(uint16_t)); memcpy(&s, &f, sizeof(uint16_t));
return table_f32_f16[s]; return ggml_table_f32_f16[s];
} }
#define GGML_FP16_TO_FP32(x) ggml_lookup_fp16_to_fp32(x) #define GGML_FP16_TO_FP32(x) ggml_lookup_fp16_to_fp32(x)

44
ggml.c
View File

@ -232,19 +232,19 @@ typedef double ggml_float;
// //
// precomputed gelu table for f16 (128 KB) // precomputed gelu table for f16 (128 KB)
static ggml_fp16_t table_gelu_f16[1 << 16]; static ggml_fp16_t ggml_table_gelu_f16[1 << 16];
// precomputed quick gelu table for f16 (128 KB) // precomputed quick gelu table for f16 (128 KB)
static ggml_fp16_t table_gelu_quick_f16[1 << 16]; static ggml_fp16_t ggml_table_gelu_quick_f16[1 << 16];
// precomputed silu table for f16 (128 KB) // precomputed silu table for f16 (128 KB)
static ggml_fp16_t table_silu_f16[1 << 16]; static ggml_fp16_t ggml_table_silu_f16[1 << 16];
// precomputed exp table for f16 (128 KB) // precomputed exp table for f16 (128 KB)
static ggml_fp16_t table_exp_f16[1 << 16]; static ggml_fp16_t ggml_table_exp_f16[1 << 16];
// precomputed f32 table for f16 (256 KB) (ggml-impl.h) // precomputed f32 table for f16 (256 KB) (ggml-impl.h)
float table_f32_f16[1 << 16]; float ggml_table_f32_f16[1 << 16];
// note: do not use these inside ggml.c // note: do not use these inside ggml.c
// these are meant to be used via the ggml.h API // these are meant to be used via the ggml.h API
@ -1363,7 +1363,7 @@ inline static float ggml_gelu_f32(float x) {
inline static void ggml_vec_gelu_f16(const int n, ggml_fp16_t * y, const ggml_fp16_t * x) { inline static void ggml_vec_gelu_f16(const int n, ggml_fp16_t * y, const ggml_fp16_t * x) {
const uint16_t * i16 = (const uint16_t *) x; const uint16_t * i16 = (const uint16_t *) x;
for (int i = 0; i < n; ++i) { for (int i = 0; i < n; ++i) {
y[i] = table_gelu_f16[i16[i]]; y[i] = ggml_table_gelu_f16[i16[i]];
} }
} }
@ -1373,7 +1373,7 @@ inline static void ggml_vec_gelu_f32(const int n, float * y, const float * x) {
for (int i = 0; i < n; ++i) { for (int i = 0; i < n; ++i) {
ggml_fp16_t fp16 = GGML_FP32_TO_FP16(x[i]); ggml_fp16_t fp16 = GGML_FP32_TO_FP16(x[i]);
memcpy(&t, &fp16, sizeof(uint16_t)); memcpy(&t, &fp16, sizeof(uint16_t));
y[i] = GGML_FP16_TO_FP32(table_gelu_f16[t]); y[i] = GGML_FP16_TO_FP32(ggml_table_gelu_f16[t]);
} }
} }
#else #else
@ -1391,7 +1391,7 @@ inline static float ggml_gelu_quick_f32(float x) {
//inline static void ggml_vec_gelu_quick_f16(const int n, ggml_fp16_t * y, const ggml_fp16_t * x) { //inline static void ggml_vec_gelu_quick_f16(const int n, ggml_fp16_t * y, const ggml_fp16_t * x) {
// const uint16_t * i16 = (const uint16_t *) x; // const uint16_t * i16 = (const uint16_t *) x;
// for (int i = 0; i < n; ++i) { // for (int i = 0; i < n; ++i) {
// y[i] = table_gelu_quick_f16[i16[i]]; // y[i] = ggml_table_gelu_quick_f16[i16[i]];
// } // }
//} //}
@ -1401,7 +1401,7 @@ inline static void ggml_vec_gelu_quick_f32(const int n, float * y, const float *
for (int i = 0; i < n; ++i) { for (int i = 0; i < n; ++i) {
ggml_fp16_t fp16 = GGML_FP32_TO_FP16(x[i]); ggml_fp16_t fp16 = GGML_FP32_TO_FP16(x[i]);
memcpy(&t, &fp16, sizeof(uint16_t)); memcpy(&t, &fp16, sizeof(uint16_t));
y[i] = GGML_FP16_TO_FP32(table_gelu_quick_f16[t]); y[i] = GGML_FP16_TO_FP32(ggml_table_gelu_quick_f16[t]);
} }
} }
#else #else
@ -1420,7 +1420,7 @@ inline static float ggml_silu_f32(float x) {
//inline static void ggml_vec_silu_f16(const int n, ggml_fp16_t * y, const ggml_fp16_t * x) { //inline static void ggml_vec_silu_f16(const int n, ggml_fp16_t * y, const ggml_fp16_t * x) {
// const uint16_t * i16 = (const uint16_t *) x; // const uint16_t * i16 = (const uint16_t *) x;
// for (int i = 0; i < n; ++i) { // for (int i = 0; i < n; ++i) {
// y[i] = table_silu_f16[i16[i]]; // y[i] = ggml_table_silu_f16[i16[i]];
// } // }
//} //}
@ -1430,7 +1430,7 @@ inline static void ggml_vec_silu_f32(const int n, float * y, const float * x) {
for (int i = 0; i < n; ++i) { for (int i = 0; i < n; ++i) {
ggml_fp16_t fp16 = GGML_FP32_TO_FP16(x[i]); ggml_fp16_t fp16 = GGML_FP32_TO_FP16(x[i]);
memcpy(&t, &fp16, sizeof(uint16_t)); memcpy(&t, &fp16, sizeof(uint16_t));
y[i] = GGML_FP16_TO_FP32(table_silu_f16[t]); y[i] = GGML_FP16_TO_FP32(ggml_table_silu_f16[t]);
} }
} }
#else #else
@ -2146,11 +2146,11 @@ struct ggml_context * ggml_init(struct ggml_init_params params) {
for (int i = 0; i < (1 << 16); ++i) { for (int i = 0; i < (1 << 16); ++i) {
uint16_t ui = i; uint16_t ui = i;
memcpy(&ii, &ui, sizeof(ii)); memcpy(&ii, &ui, sizeof(ii));
const float f = table_f32_f16[i] = GGML_COMPUTE_FP16_TO_FP32(ii); const float f = ggml_table_f32_f16[i] = GGML_COMPUTE_FP16_TO_FP32(ii);
table_gelu_f16[i] = GGML_FP32_TO_FP16(ggml_gelu_f32(f)); ggml_table_gelu_f16[i] = GGML_FP32_TO_FP16(ggml_gelu_f32(f));
table_gelu_quick_f16[i] = GGML_FP32_TO_FP16(ggml_gelu_quick_f32(f)); ggml_table_gelu_quick_f16[i] = GGML_FP32_TO_FP16(ggml_gelu_quick_f32(f));
table_silu_f16[i] = GGML_FP32_TO_FP16(ggml_silu_f32(f)); ggml_table_silu_f16[i] = GGML_FP32_TO_FP16(ggml_silu_f32(f));
table_exp_f16[i] = GGML_FP32_TO_FP16(expf(f)); ggml_table_exp_f16[i] = GGML_FP32_TO_FP16(expf(f));
} }
const uint64_t t_end = ggml_time_us(); UNUSED(t_end); const uint64_t t_end = ggml_time_us(); UNUSED(t_end);
@ -10513,7 +10513,7 @@ static void ggml_compute_forward_soft_max_f32(
// const float val = (sp[i] == -INFINITY) ? 0.0 : exp(sp[i] - max); // const float val = (sp[i] == -INFINITY) ? 0.0 : exp(sp[i] - max);
ggml_fp16_t s = GGML_FP32_TO_FP16(sp[i] - max); ggml_fp16_t s = GGML_FP32_TO_FP16(sp[i] - max);
memcpy(&scvt, &s, sizeof(scvt)); memcpy(&scvt, &s, sizeof(scvt));
const float val = GGML_FP16_TO_FP32(table_exp_f16[scvt]); const float val = GGML_FP16_TO_FP32(ggml_table_exp_f16[scvt]);
sum += (ggml_float)val; sum += (ggml_float)val;
dp[i] = val; dp[i] = val;
} }
@ -12802,7 +12802,7 @@ static void ggml_compute_forward_flash_attn_f32(
#else #else
ggml_fp16_t s = GGML_FP32_TO_FP16(SS[j] - max); ggml_fp16_t s = GGML_FP32_TO_FP16(SS[j] - max);
memcpy(&scvt[j], &s, sizeof(uint16_t)); memcpy(&scvt[j], &s, sizeof(uint16_t));
const float val = GGML_FP16_TO_FP32(table_exp_f16[scvt[j]]); const float val = GGML_FP16_TO_FP32(ggml_table_exp_f16[scvt[j]]);
#endif #endif
sump[j] += (ggml_float)val; sump[j] += (ggml_float)val;
SS[j] = val; SS[j] = val;
@ -13004,7 +13004,7 @@ static void ggml_compute_forward_flash_attn_f16(
} else { } else {
ggml_fp16_t s = GGML_FP32_TO_FP16(SS[j] - max); ggml_fp16_t s = GGML_FP32_TO_FP16(SS[j] - max);
memcpy(&scvt[j], &s, sizeof(uint16_t)); memcpy(&scvt[j], &s, sizeof(uint16_t));
const float val = GGML_FP16_TO_FP32(table_exp_f16[scvt[j]]); const float val = GGML_FP16_TO_FP32(ggml_table_exp_f16[scvt[j]]);
sump[j] += (ggml_float)val; sump[j] += (ggml_float)val;
SS[j] = val; SS[j] = val;
} }
@ -13455,7 +13455,7 @@ static void ggml_compute_forward_flash_attn_back_f32(
#else #else
ggml_fp16_t s = GGML_FP32_TO_FP16(SR[j] - max); ggml_fp16_t s = GGML_FP32_TO_FP16(SR[j] - max);
memcpy(&scvt[j], &s, sizeof(uint16_t)); memcpy(&scvt[j], &s, sizeof(uint16_t));
const float val = GGML_FP16_TO_FP32(table_exp_f16[scvt[j]]); const float val = GGML_FP16_TO_FP32(ggml_table_exp_f16[scvt[j]]);
#endif #endif
sump[j] += (ggml_float)val; sump[j] += (ggml_float)val;
SW[j] = val; SW[j] = val;
@ -14205,7 +14205,7 @@ static void ggml_compute_forward_cross_entropy_loss_f32(
#else #else
ggml_fp16_t s = GGML_FP32_TO_FP16(s0[i] - max); ggml_fp16_t s = GGML_FP32_TO_FP16(s0[i] - max);
memcpy(&scvt, &s, sizeof(scvt)); memcpy(&scvt, &s, sizeof(scvt));
const float val = GGML_FP16_TO_FP32(table_exp_f16[scvt]); const float val = GGML_FP16_TO_FP32(ggml_table_exp_f16[scvt]);
#endif #endif
sum += (ggml_float)val; sum += (ggml_float)val;
st[i] = val; st[i] = val;
@ -14319,7 +14319,7 @@ static void ggml_compute_forward_cross_entropy_loss_back_f32(
#else #else
ggml_fp16_t s = GGML_FP32_TO_FP16(s0[i] - max); ggml_fp16_t s = GGML_FP32_TO_FP16(s0[i] - max);
memcpy(&scvt, &s, sizeof(scvt)); memcpy(&scvt, &s, sizeof(scvt));
const float val = GGML_FP16_TO_FP32(table_exp_f16[scvt]); const float val = GGML_FP16_TO_FP32(ggml_table_exp_f16[scvt]);
#endif #endif
sum += (ggml_float)val; sum += (ggml_float)val;
ds0[i] = val; ds0[i] = val;