Thanks to visit codestin.com
Credit goes to github.com

Skip to content

vulkan: Allow up to 4096 elements for mul_mat_id row_ids #13326

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.

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
May 9, 2025
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
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: 2 additions & 2 deletions ggml/src/ggml-vulkan/ggml-vulkan.cpp
Original file line number Diff line number Diff line change
Expand Up @@ -1632,7 +1632,7 @@ static bool ggml_vk_matmul_shmem_support(const vk_device& device, const std::vec
const uint32_t warps = warptile[0] / warptile[10];

const uint32_t load_bufs = (warptile[1] + warptile[2]) * (warptile[3] + bank_conflict_offset) * type_size;
const uint32_t mmid_row_ids = mul_mat_id ? 3072 * sizeof(uint32_t) : 0;
const uint32_t mmid_row_ids = mul_mat_id ? 4096 * sizeof(uint32_t) : 0;
const uint32_t coopmat_stage = device->coopmat_support ? warptile[7] * warptile[8] / warps * sizeof(float) : 0;

const uint32_t total_size = load_bufs + mmid_row_ids + coopmat_stage + lut_size;
Expand Down Expand Up @@ -5260,7 +5260,7 @@ static void ggml_vk_mul_mat_id_q_f16(ggml_backend_vk_context * ctx, vk_context&

const uint64_t nei0 = ids->ne[0];
const uint64_t nei1 = ids->ne[1];
GGML_ASSERT(nei0 * nei1 <= 3072);
GGML_ASSERT(nei0 * nei1 <= 4096);

const uint32_t nbi1 = ids->nb[1];
const uint32_t nbi2 = ids->nb[2];
Expand Down
2 changes: 1 addition & 1 deletion ggml/src/ggml-vulkan/vulkan-shaders/mul_mm.comp
Original file line number Diff line number Diff line change
Expand Up @@ -103,7 +103,7 @@ shared FLOAT_TYPE buf_a[BM * SHMEM_STRIDE];
shared FLOAT_TYPE buf_b[BN * SHMEM_STRIDE];

#ifdef MUL_MAT_ID
shared u16vec2 row_ids[3072];
shared u16vec2 row_ids[4096];
#endif // MUL_MAT_ID

#define NUM_WARPS (BLOCK_SIZE / WARP)
Expand Down
2 changes: 1 addition & 1 deletion ggml/src/ggml-vulkan/vulkan-shaders/mul_mm_cm2.comp
Original file line number Diff line number Diff line change
Expand Up @@ -92,7 +92,7 @@ layout (binding = 2) writeonly buffer D {D_TYPE data_d[];};
#ifdef MUL_MAT_ID
layout (binding = 3) readonly buffer IDS {int data_ids[];};

shared u16vec4 row_ids[3072];
shared u16vec4 row_ids[4096];

layout(buffer_reference, std430, buffer_reference_align = 2) buffer decodeBufB {
B_TYPE b[];
Expand Down
2 changes: 1 addition & 1 deletion ggml/src/ggml-vulkan/vulkan-shaders/mul_mmq.comp
Original file line number Diff line number Diff line change
Expand Up @@ -101,7 +101,7 @@ shared FLOAT_TYPE_VEC2 buf_b_ds[BN];
#define LOAD_VEC_B 4

#ifdef MUL_MAT_ID
shared u16vec2 row_ids[3072];
shared u16vec2 row_ids[4096];
#endif // MUL_MAT_ID

#define NUM_WARPS (BLOCK_SIZE / WARP)
Expand Down