Skip to content

feat: turbo3/turbo2 mixed KV cache type support (CUDA) #29

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

Closed
seanrasch wants to merge 1 commit into from
Closed

feat: turbo3/turbo2 mixed KV cache type support (CUDA) #29

Show file tree
Hide file tree
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
4 changes: 3 additions & 1 deletion ggml/src/ggml-cuda/CMakeLists.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -126,7 +126,9 @@ if (CUDAToolkit_FOUND)
template-instances/fattn-vec-instance-q8_0-turbo3_0.cu
template-instances/fattn-vec-instance-turbo2_0-turbo2_0.cu
template-instances/fattn-vec-instance-turbo2_0-q8_0.cu
template-instances/fattn-vec-instance-q8_0-turbo2_0.cu)
template-instances/fattn-vec-instance-q8_0-turbo2_0.cu
template-instances/fattn-vec-instance-turbo3_0-turbo2_0.cu
template-instances/fattn-vec-instance-turbo2_0-turbo3_0.cu)
endif()

ggml_add_backend_library(ggml-cuda
Expand Down
9 changes: 9 additions & 0 deletions ggml/src/ggml-cuda/fattn-vec.cuh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -659,3 +659,12 @@ extern DECL_FATTN_VEC_CASE(256, GGML_TYPE_TURBO2_0, GGML_TYPE_Q8_0);
extern DECL_FATTN_VEC_CASE( 64, GGML_TYPE_Q8_0, GGML_TYPE_TURBO2_0);
extern DECL_FATTN_VEC_CASE(128, GGML_TYPE_Q8_0, GGML_TYPE_TURBO2_0);
extern DECL_FATTN_VEC_CASE(256, GGML_TYPE_Q8_0, GGML_TYPE_TURBO2_0);

// Mixed turbo3/turbo2 KV cache types
extern DECL_FATTN_VEC_CASE( 64, GGML_TYPE_TURBO3_0, GGML_TYPE_TURBO2_0);
extern DECL_FATTN_VEC_CASE(128, GGML_TYPE_TURBO3_0, GGML_TYPE_TURBO2_0);
extern DECL_FATTN_VEC_CASE(256, GGML_TYPE_TURBO3_0, GGML_TYPE_TURBO2_0);

extern DECL_FATTN_VEC_CASE( 64, GGML_TYPE_TURBO2_0, GGML_TYPE_TURBO3_0);
extern DECL_FATTN_VEC_CASE(128, GGML_TYPE_TURBO2_0, GGML_TYPE_TURBO3_0);
extern DECL_FATTN_VEC_CASE(256, GGML_TYPE_TURBO2_0, GGML_TYPE_TURBO3_0);
18 changes: 12 additions & 6 deletions ggml/src/ggml-cuda/fattn.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -294,6 +294,10 @@ static void ggml_cuda_flash_attn_ext_vec(ggml_backend_cuda_context & ctx, ggml_t
FATTN_VEC_CASES_ALL_D(GGML_TYPE_TURBO2_0, GGML_TYPE_Q8_0)
FATTN_VEC_CASES_ALL_D(GGML_TYPE_Q8_0, GGML_TYPE_TURBO2_0)

// Mixed turbo3/turbo2 KV cache types
FATTN_VEC_CASES_ALL_D(GGML_TYPE_TURBO3_0, GGML_TYPE_TURBO2_0)
FATTN_VEC_CASES_ALL_D(GGML_TYPE_TURBO2_0, GGML_TYPE_TURBO3_0)

GGML_ABORT("fatal error");
}

Expand Down Expand Up @@ -368,12 +372,14 @@ static best_fattn_kernel ggml_cuda_get_best_fattn_kernel(const int device, const

#ifndef GGML_CUDA_FA_ALL_QUANTS
if (K->type != V->type) {
// Allow mixed turbo/q8_0 KV types
const bool turbo_q8_mix = (K->type == GGML_TYPE_TURBO3_0 && V->type == GGML_TYPE_Q8_0) ||
(K->type == GGML_TYPE_Q8_0 && V->type == GGML_TYPE_TURBO3_0) ||
(K->type == GGML_TYPE_TURBO2_0 && V->type == GGML_TYPE_Q8_0) ||
(K->type == GGML_TYPE_Q8_0 && V->type == GGML_TYPE_TURBO2_0);
if (!turbo_q8_mix) {
// Allow mixed turbo KV types
const bool turbo_mix = (K->type == GGML_TYPE_TURBO3_0 && V->type == GGML_TYPE_Q8_0) ||
(K->type == GGML_TYPE_Q8_0 && V->type == GGML_TYPE_TURBO3_0) ||
(K->type == GGML_TYPE_TURBO2_0 && V->type == GGML_TYPE_Q8_0) ||
(K->type == GGML_TYPE_Q8_0 && V->type == GGML_TYPE_TURBO2_0) ||
(K->type == GGML_TYPE_TURBO3_0 && V->type == GGML_TYPE_TURBO2_0) ||
(K->type == GGML_TYPE_TURBO2_0 && V->type == GGML_TYPE_TURBO3_0);
if (!turbo_mix) {
return BEST_FATTN_KERNEL_NONE;
}
}
Expand Down
7 changes: 7 additions & 0 deletions ggml/src/ggml-cuda/template-instances/fattn-vec-instance-turbo2_0-turbo3_0.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,7 @@
// Mixed KV: turbo2 K + turbo3 V

#include "../fattn-vec.cuh"

DECL_FATTN_VEC_CASE( 64, GGML_TYPE_TURBO2_0, GGML_TYPE_TURBO3_0);
DECL_FATTN_VEC_CASE(128, GGML_TYPE_TURBO2_0, GGML_TYPE_TURBO3_0);
DECL_FATTN_VEC_CASE(256, GGML_TYPE_TURBO2_0, GGML_TYPE_TURBO3_0);
7 changes: 7 additions & 0 deletions ggml/src/ggml-cuda/template-instances/fattn-vec-instance-turbo3_0-turbo2_0.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,7 @@
// Mixed KV: turbo3 K + turbo2 V

#include "../fattn-vec.cuh"

DECL_FATTN_VEC_CASE( 64, GGML_TYPE_TURBO3_0, GGML_TYPE_TURBO2_0);
DECL_FATTN_VEC_CASE(128, GGML_TYPE_TURBO3_0, GGML_TYPE_TURBO2_0);
DECL_FATTN_VEC_CASE(256, GGML_TYPE_TURBO3_0, GGML_TYPE_TURBO2_0);
Loading