Skip to content

Commit

Permalink
metal : support TQ2_0
Browse files Browse the repository at this point in the history 
Mostly adapted from the IQ2_TN kernels
from ikawrakow/ik_llama.cpp#13
  • Loading branch information
@compilade
compilade committed Aug 22, 2024
1 parent cb6d996 commit 03a138d
Show file tree
Hide file tree
Showing 2 changed files with 173 additions and 2 deletions.
31 changes: 29 additions & 2 deletions ggml/src/ggml-metal.m
View file
Original file line number Diff line number Diff line change
Expand Up @@ -69,6 +69,7 @@
GGML_METAL_KERNEL_TYPE_GET_ROWS_Q4_K,
GGML_METAL_KERNEL_TYPE_GET_ROWS_Q5_K,
GGML_METAL_KERNEL_TYPE_GET_ROWS_Q6_K,
GGML_METAL_KERNEL_TYPE_GET_ROWS_TQ2_0,
GGML_METAL_KERNEL_TYPE_GET_ROWS_IQ2_XXS,
GGML_METAL_KERNEL_TYPE_GET_ROWS_IQ2_XS,
GGML_METAL_KERNEL_TYPE_GET_ROWS_IQ3_XXS,
Expand Down Expand Up @@ -97,6 +98,7 @@
GGML_METAL_KERNEL_TYPE_MUL_MV_Q4_K_F32,
GGML_METAL_KERNEL_TYPE_MUL_MV_Q5_K_F32,
GGML_METAL_KERNEL_TYPE_MUL_MV_Q6_K_F32,
GGML_METAL_KERNEL_TYPE_MUL_MV_TQ2_0_F32,
GGML_METAL_KERNEL_TYPE_MUL_MV_IQ2_XXS_F32,
GGML_METAL_KERNEL_TYPE_MUL_MV_IQ2_XS_F32,
GGML_METAL_KERNEL_TYPE_MUL_MV_IQ3_XXS_F32,
Expand All @@ -121,6 +123,7 @@
GGML_METAL_KERNEL_TYPE_MUL_MV_ID_Q4_K_F32,
GGML_METAL_KERNEL_TYPE_MUL_MV_ID_Q5_K_F32,
GGML_METAL_KERNEL_TYPE_MUL_MV_ID_Q6_K_F32,
GGML_METAL_KERNEL_TYPE_MUL_MV_ID_TQ2_0_F32,
GGML_METAL_KERNEL_TYPE_MUL_MV_ID_IQ2_XXS_F32,
GGML_METAL_KERNEL_TYPE_MUL_MV_ID_IQ2_XS_F32,
GGML_METAL_KERNEL_TYPE_MUL_MV_ID_IQ3_XXS_F32,
Expand All @@ -142,6 +145,7 @@
GGML_METAL_KERNEL_TYPE_MUL_MM_Q4_K_F32,
GGML_METAL_KERNEL_TYPE_MUL_MM_Q5_K_F32,
GGML_METAL_KERNEL_TYPE_MUL_MM_Q6_K_F32,
GGML_METAL_KERNEL_TYPE_MUL_MM_TQ2_0_F32,
GGML_METAL_KERNEL_TYPE_MUL_MM_IQ2_XXS_F32,
GGML_METAL_KERNEL_TYPE_MUL_MM_IQ2_XS_F32,
GGML_METAL_KERNEL_TYPE_MUL_MM_IQ3_XXS_F32,
Expand All @@ -163,6 +167,7 @@
GGML_METAL_KERNEL_TYPE_MUL_MM_ID_Q4_K_F32,
GGML_METAL_KERNEL_TYPE_MUL_MM_ID_Q5_K_F32,
GGML_METAL_KERNEL_TYPE_MUL_MM_ID_Q6_K_F32,
GGML_METAL_KERNEL_TYPE_MUL_MM_ID_TQ2_0_F32,
GGML_METAL_KERNEL_TYPE_MUL_MM_ID_IQ2_XXS_F32,
GGML_METAL_KERNEL_TYPE_MUL_MM_ID_IQ2_XS_F32,
GGML_METAL_KERNEL_TYPE_MUL_MM_ID_IQ3_XXS_F32,
Expand Down Expand Up @@ -529,6 +534,7 @@ static void ggml_metal_log(enum ggml_log_level level, const char * format, ...){
GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_GET_ROWS_Q4_K, get_rows_q4_K, true);
GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_GET_ROWS_Q5_K, get_rows_q5_K, true);
GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_GET_ROWS_Q6_K, get_rows_q6_K, true);
GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_GET_ROWS_TQ2_0, get_rows_tq2_0, true);
GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_GET_ROWS_IQ2_XXS, get_rows_iq2_xxs, true);
GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_GET_ROWS_IQ2_XS, get_rows_iq2_xs, true);
GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_GET_ROWS_IQ3_XXS, get_rows_iq3_xxs, true);
Expand Down Expand Up @@ -557,6 +563,7 @@ static void ggml_metal_log(enum ggml_log_level level, const char * format, ...){
GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_Q4_K_F32, mul_mv_q4_K_f32, ctx->support_simdgroup_reduction);
GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_Q5_K_F32, mul_mv_q5_K_f32, ctx->support_simdgroup_reduction);
GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_Q6_K_F32, mul_mv_q6_K_f32, ctx->support_simdgroup_reduction);
GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_TQ2_0_F32, mul_mv_tq2_0_f32, ctx->support_simdgroup_reduction);
GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_IQ2_XXS_F32, mul_mv_iq2_xxs_f32, ctx->support_simdgroup_reduction);
GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_IQ2_XS_F32, mul_mv_iq2_xs_f32, ctx->support_simdgroup_reduction);
GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_IQ3_XXS_F32, mul_mv_iq3_xxs_f32, ctx->support_simdgroup_reduction);
Expand All @@ -581,6 +588,7 @@ static void ggml_metal_log(enum ggml_log_level level, const char * format, ...){
GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_ID_Q4_K_F32, mul_mv_id_q4_K_f32, ctx->support_simdgroup_reduction);
GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_ID_Q5_K_F32, mul_mv_id_q5_K_f32, ctx->support_simdgroup_reduction);
GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_ID_Q6_K_F32, mul_mv_id_q6_K_f32, ctx->support_simdgroup_reduction);
GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_ID_TQ2_0_F32, mul_mv_id_tq2_0_f32, ctx->support_simdgroup_reduction);
GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_ID_IQ2_XXS_F32, mul_mv_id_iq2_xxs_f32, ctx->support_simdgroup_reduction);
GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_ID_IQ2_XS_F32, mul_mv_id_iq2_xs_f32, ctx->support_simdgroup_reduction);
GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_ID_IQ3_XXS_F32, mul_mv_id_iq3_xxs_f32, ctx->support_simdgroup_reduction);
Expand All @@ -602,6 +610,7 @@ static void ggml_metal_log(enum ggml_log_level level, const char * format, ...){
GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_Q4_K_F32, mul_mm_q4_K_f32, ctx->support_simdgroup_mm);
GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_Q5_K_F32, mul_mm_q5_K_f32, ctx->support_simdgroup_mm);
GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_Q6_K_F32, mul_mm_q6_K_f32, ctx->support_simdgroup_mm);
GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_TQ2_0_F32, mul_mm_tq2_0_f32, ctx->support_simdgroup_mm);
GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_IQ2_XXS_F32, mul_mm_iq2_xxs_f32, ctx->support_simdgroup_mm);
GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_IQ2_XS_F32, mul_mm_iq2_xs_f32, ctx->support_simdgroup_mm);
GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_IQ3_XXS_F32, mul_mm_iq3_xxs_f32, ctx->support_simdgroup_mm);
Expand All @@ -623,6 +632,7 @@ static void ggml_metal_log(enum ggml_log_level level, const char * format, ...){
GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_ID_Q4_K_F32, mul_mm_id_q4_K_f32, ctx->support_simdgroup_mm);
GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_ID_Q5_K_F32, mul_mm_id_q5_K_f32, ctx->support_simdgroup_mm);
GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_ID_Q6_K_F32, mul_mm_id_q6_K_f32, ctx->support_simdgroup_mm);
GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_ID_TQ2_0_F32, mul_mm_id_tq2_0_f32, ctx->support_simdgroup_mm);
GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_ID_IQ2_XXS_F32, mul_mm_id_iq2_xxs_f32, ctx->support_simdgroup_mm);
GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_ID_IQ2_XS_F32, mul_mm_id_iq2_xs_f32, ctx->support_simdgroup_mm);
GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_ID_IQ3_XXS_F32, mul_mm_id_iq3_xxs_f32, ctx->support_simdgroup_mm);
Expand Down Expand Up @@ -1607,6 +1617,7 @@ static enum ggml_status ggml_metal_graph_compute(
case GGML_TYPE_Q4_K: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_Q4_K_F32 ].pipeline; break;
case GGML_TYPE_Q5_K: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_Q5_K_F32 ].pipeline; break;
case GGML_TYPE_Q6_K: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_Q6_K_F32 ].pipeline; break;
case GGML_TYPE_TQ2_0: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_TQ2_0_F32 ].pipeline; break;
case GGML_TYPE_IQ2_XXS: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_IQ2_XXS_F32].pipeline; break;
case GGML_TYPE_IQ2_XS: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_IQ2_XS_F32 ].pipeline; break;
case GGML_TYPE_IQ3_XXS: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_IQ3_XXS_F32].pipeline; break;
Expand Down Expand Up @@ -1732,6 +1743,12 @@ static enum ggml_status ggml_metal_graph_compute(
nth1 = 32;
pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_Q6_K_F32].pipeline;
} break;
case GGML_TYPE_TQ2_0:
{
nth0 = 4;
nth1 = 16;
pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_TQ2_0_F32].pipeline;
} break;
case GGML_TYPE_IQ2_XXS:
{
nth0 = 4;
Expand Down Expand Up @@ -1816,7 +1833,8 @@ static enum ggml_status ggml_metal_graph_compute(

if (src0t == GGML_TYPE_Q4_0 || src0t == GGML_TYPE_Q4_1 || src0t == GGML_TYPE_Q5_0 ||
src0t == GGML_TYPE_Q5_1 || src0t == GGML_TYPE_Q8_0 || src0t == GGML_TYPE_Q2_K ||
src0t == GGML_TYPE_IQ1_S || src0t == GGML_TYPE_IQ1_M || src0t == GGML_TYPE_IQ2_S) {
src0t == GGML_TYPE_IQ1_S || src0t == GGML_TYPE_IQ1_M || src0t == GGML_TYPE_IQ2_S||
src0t == GGML_TYPE_TQ2_0) {
[encoder dispatchThreadgroups:MTLSizeMake((ne01 + 7)/8, ne11, ne12*ne13) threadsPerThreadgroup:MTLSizeMake(nth0, nth1, 1)];
}
else if (src0t == GGML_TYPE_IQ2_XXS || src0t == GGML_TYPE_IQ2_XS) {
Expand Down Expand Up @@ -1909,6 +1927,7 @@ static enum ggml_status ggml_metal_graph_compute(
case GGML_TYPE_Q4_K: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_ID_Q4_K_F32 ].pipeline; break;
case GGML_TYPE_Q5_K: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_ID_Q5_K_F32 ].pipeline; break;
case GGML_TYPE_Q6_K: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_ID_Q6_K_F32 ].pipeline; break;
case GGML_TYPE_TQ2_0: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_ID_TQ2_0_F32 ].pipeline; break;
case GGML_TYPE_IQ2_XXS: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_ID_IQ2_XXS_F32].pipeline; break;
case GGML_TYPE_IQ2_XS: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_ID_IQ2_XS_F32 ].pipeline; break;
case GGML_TYPE_IQ3_XXS: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_ID_IQ3_XXS_F32].pipeline; break;
Expand Down Expand Up @@ -2028,6 +2047,12 @@ static enum ggml_status ggml_metal_graph_compute(
nth1 = 32;
pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_ID_Q6_K_F32].pipeline;
} break;
case GGML_TYPE_TQ2_0:
{
nth0 = 4;
nth1 = 16;
pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_ID_TQ2_0_F32].pipeline;
} break;
case GGML_TYPE_IQ2_XXS:
{
nth0 = 4;
Expand Down Expand Up @@ -2123,7 +2148,8 @@ static enum ggml_status ggml_metal_graph_compute(

if (src0t == GGML_TYPE_Q4_0 || src0t == GGML_TYPE_Q4_1 || src0t == GGML_TYPE_Q5_0 ||
src0t == GGML_TYPE_Q5_1 || src0t == GGML_TYPE_Q8_0 || src0t == GGML_TYPE_Q2_K ||
src0t == GGML_TYPE_IQ1_S || src0t == GGML_TYPE_IQ1_M || src0t == GGML_TYPE_IQ2_S) {
src0t == GGML_TYPE_IQ1_S || src0t == GGML_TYPE_IQ1_M || src0t == GGML_TYPE_IQ2_S||
src0t == GGML_TYPE_TQ2_0) {
[encoder dispatchThreadgroups:MTLSizeMake((ne01 + 7)/8, _ne1, tgz) threadsPerThreadgroup:MTLSizeMake(nth0, nth1, 1)];
}
else if (src0t == GGML_TYPE_IQ2_XXS || src0t == GGML_TYPE_IQ2_XS) {
Expand Down Expand Up @@ -2175,6 +2201,7 @@ static enum ggml_status ggml_metal_graph_compute(
case GGML_TYPE_Q4_K: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_GET_ROWS_Q4_K ].pipeline; break;
case GGML_TYPE_Q5_K: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_GET_ROWS_Q5_K ].pipeline; break;
case GGML_TYPE_Q6_K: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_GET_ROWS_Q6_K ].pipeline; break;
case GGML_TYPE_TQ2_0: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_GET_ROWS_TQ2_0 ].pipeline; break;
case GGML_TYPE_IQ2_XXS: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_GET_ROWS_IQ2_XXS].pipeline; break;
case GGML_TYPE_IQ2_XS: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_GET_ROWS_IQ2_XS ].pipeline; break;
case GGML_TYPE_IQ3_XXS: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_GET_ROWS_IQ3_XXS].pipeline; break;
Expand Down
Loading

0 comments on commit 03a138d

Please sign in to comment.