ikawrakow/ik_llama.cpp
1
0
Fork 0
mirror of https://github.com/ikawrakow/ik_llama.cpp.git synced 2026-02-25 15:44:10 +00:00
Code Issues Packages Projects Releases Wiki Activity

Fix build when iqk_mul_mat is disabled

Browse Source 
Ref #29
This commit is contained in:
Iwan Kawrakow
2024-08-31 08:52:08 +03:00
parent c7e99c88a2
commit eb046ae5f3
2 changed files with 16 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
Makefile
View File

@@ -574,6 +574,7 @@ endif # GGML_NVPL
OBJ_GGML += ggml/src/iqk/iqk_quantize.o
ifndef GGML_NO_IQKMULMAT
MK_CFLAGS += -DGGML_USE_IQK_MULMAT
MK_CPPFLAGS += -DGGML_USE_IQK_MULMAT
OBJ_GGML += ggml/src/iqk/iqk_mul_mat.o
endif
@@ -1228,6 +1229,7 @@ clean:
rm -rvf ggml/*.so
rm -vrf ggml/src/*.o
rm -rvf ggml/src/llamafile/*.o
rm -rvf ggml/src/iqk/*.o
rm -rvf common/build-info.cpp
rm -vrf ggml/src/ggml-metal-embed.metal
rm -vrf ggml/src/ggml-cuda/*.o

14
ggml/src/iqk/iqk_quantize.cpp
View File

@@ -291,9 +291,11 @@ void ggml_vec_dot_iq2_bn_q8_K64(int n, float * s, size_t bs, const void * vx, si
static_assert(QK_IQ1BN == 64, "This dot product implementation for iq2_bn requires a block size of 64");
#if GGML_USE_IQK_MULMAT
if (iqk_mul_mat(1, 1, n, GGML_TYPE_IQ2_BN, vx, 0, GGML_TYPE_Q8_K64, vy, 0, s, 0, 0, 1)) {
return;
}
#endif
constexpr int Nj = QK_IQ1BN/4;
@@ -622,9 +624,11 @@ void vec_dot_iq2_k_q8_k(int n, float * GGML_RESTRICT s, size_t bs, const void *
GGML_UNUSED(by);
GGML_UNUSED(bs);
#if GGML_USE_IQK_MULMAT
if (iqk_mul_mat(1, 1, n, GGML_TYPE_IQ2_K, vx, 0, GGML_TYPE_Q8_K, vy, 0, s, 0, 0, 1)) {
return;
}
#endif
const int nb = n / QK_K;
@@ -893,9 +897,11 @@ void vec_dot_iq3_k_q8_k(int n, float * GGML_RESTRICT s, size_t bs, const void *
GGML_UNUSED(by);
GGML_UNUSED(bs);
#if GGML_USE_IQK_MULMAT
if (iqk_mul_mat(1, 1, n, GGML_TYPE_IQ3_K, vx, 0, GGML_TYPE_Q8_K, vy, 0, s, 0, 0, 1)) {
return;
}
#endif
const int nb = n / QK_K;
@@ -943,9 +949,11 @@ void vec_dot_iq4_k_q8_k(int n, float * s, size_t bs, const void * vx, size_t bx,
GGML_UNUSED(by);
GGML_UNUSED(bs);
#if GGML_USE_IQK_MULMAT
if (iqk_mul_mat(1, 1, n, GGML_TYPE_IQ4_K, vx, 0, GGML_TYPE_Q8_K, vy, 0, s, 0, 0, 1)) {
return;
}
#endif
const int nb = n / QK_K;
@@ -1241,9 +1249,11 @@ void vec_dot_iq5_k_q8_k(int n, float * s, size_t bs, const void * vx, size_t bx,
GGML_UNUSED(by);
GGML_UNUSED(bs);
#if GGML_USE_IQK_MULMAT
if (iqk_mul_mat(1, 1, n, GGML_TYPE_IQ5_K, vx, 0, GGML_TYPE_Q8_K, vy, 0, s, 0, 0, 1)) {
return;
}
#endif
const int nb = n / QK_K;
@@ -1580,9 +1590,11 @@ void vec_dot_iq6_k_q8_k(int n, float * s, size_t bs, const void * vx, size_t bx,
GGML_UNUSED(by);
GGML_UNUSED(bs);
#if GGML_USE_IQK_MULMAT
if (iqk_mul_mat(1, 1, n, GGML_TYPE_IQ6_K, vx, 0, GGML_TYPE_Q8_K, vy, 0, s, 0, 0, 1)) {
return;
}
#endif
// TODO
//const int nb = n / QK_K;
@@ -1950,9 +1962,11 @@ void dequantize_row_iq2_tn(const block_iq2_tn * x, float * y, int64_t k) {
}
void vec_dot_iq2_tn_q8_k(int n, float * s, size_t bs, const void * vx, size_t bx, const void * vy, size_t by, int nrc) {
#if GGML_USE_IQK_MULMAT
if (iqk_mul_mat(1, 1, n, GGML_TYPE_IQ2_TN, vx, 0, GGML_TYPE_Q8_K, vy, 0, s, 0, 0, 1)) {
return;
}
#endif
const int nb = n / QK_K;