public_git
/
whisper.cpp
mirror of https://github.com/ggml-org/whisper.cpp.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

ggml-webgpu: add mulmat with overlapping src0/src1 (e.g., for minimax-01) (llama/27321)

This commit is contained in:
Masashi Yoshimura 2026-08-19 16:29:33 +09:00 committed by Georgi Gerganov
parent fa1e2bcaaf
commit 7df5fa8e62
8 changed files with 171 additions and 88 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

26
ggml/src/ggml-webgpu/ggml-webgpu-shader-lib.hpp
View File

@ -954,10 +954,11 @@ struct ggml_webgpu_mul_mat_vec_pipeline_key {
int vectorized; int vectorized;
uint32_t num_cols; uint32_t num_cols;
bool use_mmvq; bool use_mmvq;
bool src_overlap;
bool operator==(const ggml_webgpu_mul_mat_vec_pipeline_key & other) const { bool operator==(const ggml_webgpu_mul_mat_vec_pipeline_key & other) const {
return src0_type == other.src0_type && src1_type == other.src1_type && vectorized == other.vectorized && return src0_type == other.src0_type && src1_type == other.src1_type && vectorized == other.vectorized &&
num_cols == other.num_cols && use_mmvq == other.use_mmvq; num_cols == other.num_cols && use_mmvq == other.use_mmvq && src_overlap == other.src_overlap;
} }
}; };
@ -969,6 +970,7 @@ struct ggml_webgpu_mul_mat_vec_pipeline_key_hash {
ggml_webgpu_hash_combine(seed, key.vectorized); ggml_webgpu_hash_combine(seed, key.vectorized);
ggml_webgpu_hash_combine(seed, key.num_cols); ggml_webgpu_hash_combine(seed, key.num_cols);
ggml_webgpu_hash_combine(seed, key.use_mmvq); ggml_webgpu_hash_combine(seed, key.use_mmvq);
ggml_webgpu_hash_combine(seed, key.src_overlap);
return seed; return seed;
} }
}; };
@ -977,6 +979,7 @@ struct ggml_webgpu_mul_mat_vec_shader_decisions {
uint32_t wg_size; uint32_t wg_size;
uint32_t outputs_per_wg; uint32_t outputs_per_wg;
uint32_t vec_size; uint32_t vec_size;
bool src_overlap = false;
}; };
struct ggml_webgpu_quantize_q8_pipeline_key { struct ggml_webgpu_quantize_q8_pipeline_key {
@ -998,10 +1001,11 @@ struct ggml_webgpu_mul_mat_pipeline_key {
ggml_type src1_type; ggml_type src1_type;
int vectorized; int vectorized;
int use_subgroup_matrix; int use_subgroup_matrix;
bool src_overlap;
bool operator==(const ggml_webgpu_mul_mat_pipeline_key & other) const { bool operator==(const ggml_webgpu_mul_mat_pipeline_key & other) const {
return src0_type == other.src0_type && src1_type == other.src1_type && vectorized == other.vectorized && return src0_type == other.src0_type && src1_type == other.src1_type && vectorized == other.vectorized &&
use_subgroup_matrix == other.use_subgroup_matrix; use_subgroup_matrix == other.use_subgroup_matrix && src_overlap == other.src_overlap;
} }
}; };
@ -1012,6 +1016,7 @@ struct ggml_webgpu_mul_mat_pipeline_key_hash {
ggml_webgpu_hash_combine(seed, key.src1_type); ggml_webgpu_hash_combine(seed, key.src1_type);
ggml_webgpu_hash_combine(seed, key.vectorized); ggml_webgpu_hash_combine(seed, key.vectorized);
ggml_webgpu_hash_combine(seed, key.use_subgroup_matrix); ggml_webgpu_hash_combine(seed, key.use_subgroup_matrix);
ggml_webgpu_hash_combine(seed, key.src_overlap);
return seed; return seed;
} }
}; };
@ -1034,6 +1039,7 @@ struct ggml_webgpu_mul_mat_shader_decisions {
uint32_t subgroup_matrix_n; uint32_t subgroup_matrix_n;
uint32_t mul_mat_wg_size; uint32_t mul_mat_wg_size;
bool src_overlap = false;
}; };
/** MUL_MAT_ID **/ /** MUL_MAT_ID **/
@ -1950,7 +1956,7 @@ class ggml_webgpu_shader_lib {
return quantize_q8_pipelines[key]; return quantize_q8_pipelines[key];
} }
webgpu_pipeline get_mul_mat_vec_pipeline(const ggml_webgpu_shader_lib_context & context) { webgpu_pipeline get_mul_mat_vec_pipeline(const ggml_webgpu_shader_lib_context & context, bool src_overlap) {
ggml_webgpu_mul_mat_vec_pipeline_key key = {}; ggml_webgpu_mul_mat_vec_pipeline_key key = {};
key.src0_type = context.src0->type; key.src0_type = context.src0->type;
key.src1_type = context.src1->type; key.src1_type = context.src1->type;
@ -1961,6 +1967,7 @@ class ggml_webgpu_shader_lib {
key.num_cols = context.dst->ne[1]; key.num_cols = context.dst->ne[1];
key.use_mmvq = key.use_mmvq =
ggml_webgpu_can_use_mmvq(context.src0, context.src1, context.supports_dot_product, context.vendor); ggml_webgpu_can_use_mmvq(context.src0, context.src1, context.supports_dot_product, context.vendor);
key.src_overlap = src_overlap;
auto it = mul_mat_vec_pipelines.find(key); auto it = mul_mat_vec_pipelines.find(key);
if (it != mul_mat_vec_pipelines.end()) { if (it != mul_mat_vec_pipelines.end()) {
@ -2068,6 +2075,11 @@ class ggml_webgpu_shader_lib {
defines.push_back("Q8_1_T"); defines.push_back("Q8_1_T");
} }
if (key.src_overlap) {
defines.push_back("SRC_OVERLAP");
variant += "_src_overlap";
}
defines.push_back(std::string("WG_SIZE=") + std::to_string(wg_size)); defines.push_back(std::string("WG_SIZE=") + std::to_string(wg_size));
defines.push_back(std::string("OUTPUTS_PER_WG=") + std::to_string(outputs_per_wg)); defines.push_back(std::string("OUTPUTS_PER_WG=") + std::to_string(outputs_per_wg));
defines.push_back(context.supports_subgroups ? "USE_SUBGROUP_REDUCTION" : "USE_WORKGROUP_REDUCTION"); defines.push_back(context.supports_subgroups ? "USE_SUBGROUP_REDUCTION" : "USE_WORKGROUP_REDUCTION");
@ -2089,7 +2101,7 @@ class ggml_webgpu_shader_lib {
return mul_mat_vec_pipelines[key]; return mul_mat_vec_pipelines[key];
} }
webgpu_pipeline get_mul_mat_fast_pipeline(const ggml_webgpu_shader_lib_context & context) { webgpu_pipeline get_mul_mat_fast_pipeline(const ggml_webgpu_shader_lib_context & context, bool src_overlap) {
ggml_webgpu_mul_mat_pipeline_key key = {}; ggml_webgpu_mul_mat_pipeline_key key = {};
key.src0_type = context.src0->type; key.src0_type = context.src0->type;
key.src1_type = context.src1->type; key.src1_type = context.src1->type;
@ -2098,6 +2110,7 @@ class ggml_webgpu_shader_lib {
1 : 1 :
0; 0;
key.use_subgroup_matrix = context.supports_subgroup_matrix; key.use_subgroup_matrix = context.supports_subgroup_matrix;
key.src_overlap = src_overlap;
auto it = mul_mat_fast_pipelines.find(key); auto it = mul_mat_fast_pipelines.find(key);
if (it != mul_mat_fast_pipelines.end()) { if (it != mul_mat_fast_pipelines.end()) {
@ -2216,6 +2229,11 @@ class ggml_webgpu_shader_lib {
variant += "_vectorized"; variant += "_vectorized";
} }
if (key.src_overlap) {
defines.push_back("SRC_OVERLAP");
variant += "_src_overlap";
}
if (!key.use_subgroup_matrix) { if (!key.use_subgroup_matrix) {
defines.push_back("WORKGROUP_SIZE_M=" + std::to_string(WEBGPU_MUL_MAT_WG_SIZE_M) + "u"); defines.push_back("WORKGROUP_SIZE_M=" + std::to_string(WEBGPU_MUL_MAT_WG_SIZE_M) + "u");
defines.push_back("WORKGROUP_SIZE_N=" + std::to_string(WEBGPU_MUL_MAT_WG_SIZE_N) + "u"); defines.push_back("WORKGROUP_SIZE_N=" + std::to_string(WEBGPU_MUL_MAT_WG_SIZE_N) + "u");

63
ggml/src/ggml-webgpu/ggml-webgpu.cpp
View File

@ -1628,48 +1628,65 @@ static webgpu_encoded_op ggml_webgpu_mul_mat(webgpu_context & ctx,
// Get or create pipeline // Get or create pipeline
webgpu_pipeline pipeline; webgpu_pipeline pipeline;
std::vector<webgpu_dispatch_desc> dispatches; std::vector<webgpu_dispatch_desc> dispatches;
const bool src_overlap = ggml_webgpu_tensor_binding_overlap(ctx->global_ctx, src0, src1) && !use_mmvq;
if (use_mat_vec) { if (use_mat_vec) {
if (use_mmvq) { if (use_mmvq) {
ggml_webgpu_quantize_q8_dispatch(ctx, src0, src1, dst, dispatches); ggml_webgpu_quantize_q8_dispatch(ctx, src0, src1, dst, dispatches);
} }
pipeline = ctx->shader_lib->get_mul_mat_vec_pipeline(shader_lib_ctx); pipeline = ctx->shader_lib->get_mul_mat_vec_pipeline(shader_lib_ctx, src_overlap);
} else { } else {
pipeline = ctx->shader_lib->get_mul_mat_fast_pipeline(shader_lib_ctx); pipeline = ctx->shader_lib->get_mul_mat_fast_pipeline(shader_lib_ctx, src_overlap);
}
uint32_t offset_src0 = (uint32_t) (ggml_webgpu_tensor_misalignment(ctx, src0) / ggml_type_size(src0->type));
uint32_t offset_src1 = (uint32_t) (ggml_webgpu_tensor_misalignment(ctx, src1) / ggml_type_size(src1->type));
size_t merged_offset = 0;
size_t merged_size = 0;
if (src_overlap) {
const ggml_webgpu_merged_binding_range merged_range =
ggml_webgpu_tensor_merged_binding_range(ctx, { src0, src1 });
merged_offset = merged_range.offset;
merged_size = merged_range.size;
offset_src0 = ggml_webgpu_tensor_merged_element_offset(src0, merged_range);
offset_src1 = ggml_webgpu_tensor_merged_element_offset(src1, merged_range);
} }
// Build params // Build params
std::vector<uint32_t> params = { std::vector<uint32_t> params = { offset_src0,
(uint32_t) (ggml_webgpu_tensor_misalignment(ctx, src0) / ggml_type_size(src0->type)), offset_src1,
(uint32_t) (ggml_webgpu_tensor_misalignment(ctx, src1) / ggml_type_size(src1->type)), (uint32_t) (ggml_webgpu_tensor_misalignment(ctx, dst) / ggml_type_size(dst->type)),
(uint32_t) (ggml_webgpu_tensor_misalignment(ctx, dst) / ggml_type_size(dst->type)), (uint32_t) dst->ne[0],
(uint32_t) dst->ne[0], (uint32_t) dst->ne[1],
(uint32_t) dst->ne[1], (uint32_t) src0->ne[0],
(uint32_t) src0->ne[0], (uint32_t) (src0->nb[1] / ggml_type_size(src0->type)),
(uint32_t) (src0->nb[1] / ggml_type_size(src0->type)), (uint32_t) (src1->nb[1] / ggml_type_size(src1->type)),
(uint32_t) (src1->nb[1] / ggml_type_size(src1->type)), (uint32_t) (src0->nb[2] / ggml_type_size(src0->type)),
(uint32_t) (src0->nb[2] / ggml_type_size(src0->type)), (uint32_t) (src1->nb[2] / ggml_type_size(src1->type)),
(uint32_t) (src1->nb[2] / ggml_type_size(src1->type)), (uint32_t) (src0->nb[3] / ggml_type_size(src0->type)),
(uint32_t) (src0->nb[3] / ggml_type_size(src0->type)), (uint32_t) (src1->nb[3] / ggml_type_size(src1->type)),
(uint32_t) (src1->nb[3] / ggml_type_size(src1->type)), (uint32_t) src0->ne[2],
(uint32_t) src0->ne[2], (uint32_t) src0->ne[3],
(uint32_t) src0->ne[3], (uint32_t) (src1->ne[2] / src0->ne[2]),
(uint32_t) (src1->ne[2] / src0->ne[2]), (uint32_t) (src1->ne[3] / src0->ne[3]) };
(uint32_t) (src1->ne[3] / src0->ne[3])
};
// Build bind group entries // Build bind group entries
std::vector<wgpu::BindGroupEntry> entries = {}; std::vector<wgpu::BindGroupEntry> entries = {};
entries.push_back(ggml_webgpu_make_tensor_bind_group_entry(ctx, 0, src0));
if (use_mmvq) { if (use_mmvq) {
entries.push_back(ggml_webgpu_make_tensor_bind_group_entry(ctx, 0, src0));
auto & mmvq_qq8_entry = dispatches[0].bind_group_entries[1]; auto & mmvq_qq8_entry = dispatches[0].bind_group_entries[1];
entries.push_back(ggml_webgpu_make_bind_group_entry(1, ggml_webgpu_tensor_buf(dst), mmvq_qq8_entry.offset, entries.push_back(ggml_webgpu_make_bind_group_entry(1, ggml_webgpu_tensor_buf(dst), mmvq_qq8_entry.offset,
mmvq_qq8_entry.size)); mmvq_qq8_entry.size));
entries.push_back(ggml_webgpu_make_tensor_bind_group_entry(ctx, 2, dst));
} else if (src_overlap) {
entries.push_back(
ggml_webgpu_make_bind_group_entry(0, ggml_webgpu_tensor_buf(src0), merged_offset, merged_size));
entries.push_back(ggml_webgpu_make_tensor_bind_group_entry(ctx, 1, dst));
} else { } else {
entries.push_back(ggml_webgpu_make_tensor_bind_group_entry(ctx, 0, src0));
entries.push_back(ggml_webgpu_make_tensor_bind_group_entry(ctx, 1, src1)); entries.push_back(ggml_webgpu_make_tensor_bind_group_entry(ctx, 1, src1));
entries.push_back(ggml_webgpu_make_tensor_bind_group_entry(ctx, 2, dst));
} }
entries.push_back(ggml_webgpu_make_tensor_bind_group_entry(ctx, 2, dst));
// Calculate workgroup dimensions // Calculate workgroup dimensions
uint32_t wg_x = 1; uint32_t wg_x = 1;

14
ggml/src/ggml-webgpu/wgsl-shaders/common_decls.tmpl
View File

@ -1,3 +1,7 @@
#ifndef SRC0
#define SRC0 src0
#endif
#ifdef BYTE_HELPERS #ifdef BYTE_HELPERS
fn get_byte(value: u32, index: u32) -> u32 { fn get_byte(value: u32, index: u32) -> u32 {
return (value >> (index * 8)) & 0xFF; return (value >> (index * 8)) & 0xFF;
@ -46,7 +50,7 @@ fn load_f16_as_f32_at_src(byte_offset: u32) -> f32 {
#ifdef DECLARE_BYTE_LOADERS_SRC0 #ifdef DECLARE_BYTE_LOADERS_SRC0
fn load_u16_at_src0(byte_offset: u32) -> u32 { fn load_u16_at_src0(byte_offset: u32) -> u32 {
let word = src0[byte_offset / 4u]; let word = SRC0[byte_offset / 4u];
let shift = (byte_offset & 0x2u) * 8u; let shift = (byte_offset & 0x2u) * 8u;
return (word >> shift) & 0xFFFFu; return (word >> shift) & 0xFFFFu;
} }
@ -55,14 +59,14 @@ fn load_u16_at_src0(byte_offset: u32) -> u32 {
// Caller extracts the 16-bit half it needs via & 0xFFFFu or >> 16u. // Caller extracts the 16-bit half it needs via & 0xFFFFu or >> 16u.
// this is used in k-quants for better performance // this is used in k-quants for better performance
fn load_u32_at_src0_aligned(byte_offset: u32) -> u32 { fn load_u32_at_src0_aligned(byte_offset: u32) -> u32 {
return src0[(byte_offset & ~3u) / 4u]; return SRC0[(byte_offset & ~3u) / 4u];
} }
fn load_u32_at_src0(byte_offset: u32) -> u32 { fn load_u32_at_src0(byte_offset: u32) -> u32 {
let word_idx = byte_offset / 4u; let word_idx = byte_offset / 4u;
let shift = (byte_offset & 0x3u) * 8u; let shift = (byte_offset & 0x3u) * 8u;
let lo = src0[word_idx]; let lo = SRC0[word_idx];
let hi = src0[word_idx + 1u]; let hi = SRC0[word_idx + 1u];
let shifted = (lo >> shift) | (hi << (32u - shift)); let shifted = (lo >> shift) | (hi << (32u - shift));
return select(shifted, lo, shift == 0u); return select(shifted, lo, shift == 0u);
} }
@ -73,7 +77,7 @@ fn load_f16_at_src0(byte_offset: u32) -> f16 {
} }
fn load_f16_as_f32_at_src0(byte_offset: u32) -> f32 { fn load_f16_as_f32_at_src0(byte_offset: u32) -> f32 {
let word = src0[byte_offset / 4u]; let word = SRC0[byte_offset / 4u];
let shift = (byte_offset & 0x2u) * 8u; let shift = (byte_offset & 0x2u) * 8u;
let d_bits = (word >> shift) & 0xFFFFu; let d_bits = (word >> shift) & 0xFFFFu;
return unpack2x16float(d_bits)[0]; return unpack2x16float(d_bits)[0];

11
ggml/src/ggml-webgpu/wgsl-shaders/mul_mat_decls.tmpl
View File

@ -1,3 +1,10 @@
#ifndef SRC0
#define SRC0 src0
#endif
#ifndef SRC1
#define SRC1 src1
#endif
#ifdef VEC #ifdef VEC
#define VEC_SIZE 4 #define VEC_SIZE 4
#define SHMEM_TYPE vec4<f16> #define SHMEM_TYPE vec4<f16>
@ -39,7 +46,7 @@ fn init_shmem_src0(thread_id: u32, batch_offset: u32, offset_m: u32, k_outer: u3
let src0_idx = batch_offset + global_m * params.stride_01 + global_k; let src0_idx = batch_offset + global_m * params.stride_01 + global_k;
let src0_val = select( // taking a slight performance hit to avoid oob let src0_val = select( // taking a slight performance hit to avoid oob
SRC0_TYPE(0.0), SRC0_TYPE(0.0),
src0[src0_idx/VEC_SIZE], SRC0[src0_idx/VEC_SIZE],
global_m < params.m && global_k < params.k); global_m < params.m && global_k < params.k);
store_shmem(SHMEM_TYPE(src0_val), elem_idx); store_shmem(SHMEM_TYPE(src0_val), elem_idx);
} }
@ -57,7 +64,7 @@ fn init_shmem_src1(thread_id: u32, batch_offset: u32, offset_n: u32, k_outer: u3
let src1_idx = batch_offset + global_n * params.stride_11 + global_k; let src1_idx = batch_offset + global_n * params.stride_11 + global_k;
let src1_val = select( let src1_val = select(
SRC1_TYPE(0.0), SRC1_TYPE(0.0),
src1[src1_idx/VEC_SIZE], SRC1[src1_idx/VEC_SIZE],
global_n < params.n && global_k < params.k); global_n < params.n && global_k < params.k);
store_shmem(SHMEM_TYPE(src1_val), TILE_SRC0_SHMEM + elem_idx); store_shmem(SHMEM_TYPE(src1_val), TILE_SRC0_SHMEM + elem_idx);
} }

16
ggml/src/ggml-webgpu/wgsl-shaders/mul_mat_reg_tile.wgsl
View File

@ -1,8 +1,12 @@
enable f16; enable f16;
#define DECLARE_BYTE_LOADERS_SRC0 #define DECLARE_BYTE_LOADERS_SRC0
#include "common_decls.tmpl"
#ifdef SRC_OVERLAP
#define SRC0 merged_src
#define SRC1 merged_src
#endif
#include "common_decls.tmpl"
#include "mul_mat_decls.tmpl" #include "mul_mat_decls.tmpl"
#ifdef VEC #ifdef VEC
@ -36,11 +40,17 @@ struct MulMatParams {
broadcast3: u32 broadcast3: u32
}; };
#ifdef SRC_OVERLAP
@group(0) @binding(0) var<storage, read_write> merged_src: array<SRC0_TYPE>;
#define DST_BINDING 1
#else
@group(0) @binding(0) var<storage, read_write> src0: array<SRC0_TYPE>; // M rows, K columns @group(0) @binding(0) var<storage, read_write> src0: array<SRC0_TYPE>; // M rows, K columns
@group(0) @binding(1) var<storage, read_write> src1: array<SRC1_TYPE>; // K rows, N columns (transposed) @group(0) @binding(1) var<storage, read_write> src1: array<SRC1_TYPE>; // K rows, N columns (transposed)
@group(0) @binding(2) var<storage, read_write> dst: array<DST_TYPE>; // M rows, N columns (transposed) #define DST_BINDING 2
#endif
@group(0) @binding(3) var<uniform> params: MulMatParams; @group(0) @binding(DST_BINDING) var<storage, read_write> dst: array<DST_TYPE>; // M rows, N columns (transposed)
@group(0) @binding(DST_BINDING + 1) var<uniform> params: MulMatParams;
fn get_local_n(thread_id: u32) -> u32 { fn get_local_n(thread_id: u32) -> u32 {
return thread_id / WORKGROUP_SIZE_M; return thread_id / WORKGROUP_SIZE_M;

14
ggml/src/ggml-webgpu/wgsl-shaders/mul_mat_subgroup_matrix.wgsl
View File

@ -4,6 +4,10 @@ enable subgroups;
enable chromium_experimental_subgroup_matrix; enable chromium_experimental_subgroup_matrix;
#define DECLARE_BYTE_LOADERS_SRC0 #define DECLARE_BYTE_LOADERS_SRC0
#ifdef SRC_OVERLAP
#define SRC0 merged_src
#define SRC1 merged_src
#endif
#include "common_decls.tmpl" #include "common_decls.tmpl"
#include "mul_mat_decls.tmpl" #include "mul_mat_decls.tmpl"
@ -48,11 +52,17 @@ struct MulMatParams {
}; };
// SRC0_TYPE and SRC1_TYPE are defined in mul_mat_decls, which is included // SRC0_TYPE and SRC1_TYPE are defined in mul_mat_decls, which is included
#ifdef SRC_OVERLAP
@group(0) @binding(0) var<storage, read_write> merged_src: array<SRC0_TYPE>;
#define DST_BINDING 1
#else
@group(0) @binding(0) var<storage, read_write> src0: array<SRC0_TYPE>; // M rows, K columns @group(0) @binding(0) var<storage, read_write> src0: array<SRC0_TYPE>; // M rows, K columns
@group(0) @binding(1) var<storage, read_write> src1: array<SRC1_TYPE>; // K rows, N columns (transposed) @group(0) @binding(1) var<storage, read_write> src1: array<SRC1_TYPE>; // K rows, N columns (transposed)
@group(0) @binding(2) var<storage, read_write> dst: array<DST_TYPE>; // M rows, N columns (transposed) #define DST_BINDING 2
#endif
@group(0) @binding(3) var<uniform> params: MulMatParams; @group(0) @binding(DST_BINDING) var<storage, read_write> dst: array<DST_TYPE>; // M rows, N columns (transposed)
@group(0) @binding(DST_BINDING + 1) var<uniform> params: MulMatParams;
const WG_M_SG_TILE_SIZE = SUBGROUP_M * SUBGROUP_MATRIX_M * SUBGROUP_MATRIX_M_SIZE; const WG_M_SG_TILE_SIZE = SUBGROUP_M * SUBGROUP_MATRIX_M * SUBGROUP_MATRIX_M_SIZE;
const WG_N_SG_TILE_SIZE = SUBGROUP_N * SUBGROUP_MATRIX_N * SUBGROUP_MATRIX_N_SIZE; const WG_N_SG_TILE_SIZE = SUBGROUP_N * SUBGROUP_MATRIX_N * SUBGROUP_MATRIX_N_SIZE;

18
ggml/src/ggml-webgpu/wgsl-shaders/mul_mat_vec.wgsl
View File

@ -7,6 +7,11 @@ enable f16;
requires packed_4x8_integer_dot_product; requires packed_4x8_integer_dot_product;
#endif #endif
#ifdef SRC_OVERLAP
#define SRC0 merged_src
#define SRC1 merged_src
#endif
#define DECLARE_BYTE_LOADERS_SRC0 #define DECLARE_BYTE_LOADERS_SRC0
#include "common_decls.tmpl" #include "common_decls.tmpl"
@ -35,17 +40,22 @@ struct MulMatParams {
broadcast3: u32 broadcast3: u32
}; };
#if defined(MMVQ)
@group(0) @binding(0) var<storage, read_write> src0: array<SRC0_TYPE>; @group(0) @binding(0) var<storage, read_write> src0: array<SRC0_TYPE>;
#ifdef MMVQ
@group(0) @binding(1) var<storage, read_write> src1q: array<q8_1>; @group(0) @binding(1) var<storage, read_write> src1q: array<q8_1>;
#define DST_BINDING 2
#elif defined(SRC_OVERLAP)
@group(0) @binding(0) var<storage, read_write> merged_src: array<SRC0_TYPE>;
#define DST_BINDING 1
#else #else
@group(0) @binding(0) var<storage, read_write> src0: array<SRC0_TYPE>;
@group(0) @binding(1) var<storage, read_write> src1: array<SRC1_TYPE>; @group(0) @binding(1) var<storage, read_write> src1: array<SRC1_TYPE>;
#define DST_BINDING 2
#endif #endif
@group(0) @binding(2) var<storage, read_write> dst: array<f32>; @group(0) @binding(DST_BINDING) var<storage, read_write> dst: array<f32>;
// "mul_mat_vec_acc.tmpl" requires params.k, params.m, params.stride_01 // "mul_mat_vec_acc.tmpl" requires params.k, params.m, params.stride_01
@group(0) @binding(3) var<uniform> params: MulMatParams; @group(0) @binding(DST_BINDING + 1) var<uniform> params: MulMatParams;
// Flattened as [row][thread] to keep each row's reduction contiguous in memory. // Flattened as [row][thread] to keep each row's reduction contiguous in memory.
var<workgroup> partial_sums: array<f32, OUTPUTS_PER_WG * WG_SIZE>; var<workgroup> partial_sums: array<f32, OUTPUTS_PER_WG * WG_SIZE>;

97
ggml/src/ggml-webgpu/wgsl-shaders/mul_mat_vec_acc.tmpl
View File

@ -1,3 +1,10 @@
#ifndef SRC0
#define SRC0 src0
#endif
#ifndef SRC1
#define SRC1 src1
#endif
#ifdef U32_DEQUANT_HELPERS #ifdef U32_DEQUANT_HELPERS
#define SRC0_TYPE u32 #define SRC0_TYPE u32
@ -43,13 +50,13 @@ fn accumulate_vec_dot(thread_id: u32, row_base: u32, src0_batch_offset: u32, src
for (var k = thread_id; k < k_vec; k += WG_SIZE) { for (var k = thread_id; k < k_vec; k += WG_SIZE) {
var x_vals: array<SRC1_TYPE, NUM_COLS>; var x_vals: array<SRC1_TYPE, NUM_COLS>;
for (var col = 0u;col < NUM_COLS;col += 1) { for (var col = 0u;col < NUM_COLS;col += 1) {
x_vals[col] = src1[src1_idx_base_vec + col * (params.stride_11 / VEC_SIZE) + k]; x_vals[col] = SRC1[src1_idx_base_vec + col * (params.stride_11 / VEC_SIZE) + k];
} }
for (var row = 0u; row < OUTPUTS_PER_WG; row++) { for (var row = 0u; row < OUTPUTS_PER_WG; row++) {
let output_row = row_base + row; let output_row = row_base + row;
if (output_row < params.m) { if (output_row < params.m) {
let src0_idx = (src0_batch_offset + output_row * params.stride_01) / VEC_SIZE + k; let src0_idx = (src0_batch_offset + output_row * params.stride_01) / VEC_SIZE + k;
let w = src0[src0_idx]; let w = SRC0[src0_idx];
for (var col = 0u;col < NUM_COLS;col += 1) { for (var col = 0u;col < NUM_COLS;col += 1) {
acc[col][row] += inner_dot(w, x_vals[col]); acc[col][row] += inner_dot(w, x_vals[col]);
} }
@ -76,7 +83,7 @@ fn accumulate_vec_dot(thread_id: u32, row_base: u32, src0_batch_offset: u32, src
var x_block: array<array<f32, ELEMS_PER_THREAD>, NUM_COLS>; var x_block: array<array<f32, ELEMS_PER_THREAD>, NUM_COLS>;
for (var col = 0u; col < NUM_COLS;col += 1) { for (var col = 0u; col < NUM_COLS;col += 1) {
for (var i = 0u; i < ELEMS_PER_THREAD; i++) { for (var i = 0u; i < ELEMS_PER_THREAD; i++) {
x_block[col][i] = f32(src1[x_base + col * params.stride_11 + i]); x_block[col][i] = f32(SRC1[x_base + col * params.stride_11 + i]);
} }
} }
for (var row = 0u; row < OUTPUTS_PER_WG; row++) { for (var row = 0u; row < OUTPUTS_PER_WG; row++) {
@ -116,8 +123,8 @@ fn accumulate_vec_dot(thread_id: u32, row_base: u32, src0_batch_offset: u32, src
var x_block: array<array<f32, ELEMS_PER_THREAD>, NUM_COLS>; var x_block: array<array<f32, ELEMS_PER_THREAD>, NUM_COLS>;
for (var col = 0u; col < NUM_COLS;col += 1) { for (var col = 0u; col < NUM_COLS;col += 1) {
for (var i = 0u; i < ELEMS_PER_THREAD / 2; i++) { for (var i = 0u; i < ELEMS_PER_THREAD / 2; i++) {
x_block[col][i] = f32(src1[x_base + col * params.stride_11 + i]); x_block[col][i] = f32(SRC1[x_base + col * params.stride_11 + i]);
x_block[col][i + 4] = f32(src1[x_base + col * params.stride_11 + i + 16]); x_block[col][i + 4] = f32(SRC1[x_base + col * params.stride_11 + i + 16]);
} }
} }
for (var row = 0u; row < OUTPUTS_PER_WG; row++) { for (var row = 0u; row < OUTPUTS_PER_WG; row++) {
@ -160,8 +167,8 @@ fn accumulate_vec_dot(thread_id: u32, row_base: u32, src0_batch_offset: u32, src
var x_block: array<array<f32, ELEMS_PER_THREAD>, NUM_COLS>; var x_block: array<array<f32, ELEMS_PER_THREAD>, NUM_COLS>;
for (var col = 0u; col < NUM_COLS;col += 1) { for (var col = 0u; col < NUM_COLS;col += 1) {
for (var i = 0u; i < ELEMS_PER_THREAD / 2; i++) { for (var i = 0u; i < ELEMS_PER_THREAD / 2; i++) {
x_block[col][i] = f32(src1[x_base + col * params.stride_11 + i]); x_block[col][i] = f32(SRC1[x_base + col * params.stride_11 + i]);
x_block[col][i + 4] = f32(src1[x_base + col * params.stride_11 + i + 16]); x_block[col][i + 4] = f32(SRC1[x_base + col * params.stride_11 + i + 16]);
} }
} }
for (var row = 0u; row < OUTPUTS_PER_WG; row++) { for (var row = 0u; row < OUTPUTS_PER_WG; row++) {
@ -205,8 +212,8 @@ fn accumulate_vec_dot(thread_id: u32, row_base: u32, src0_batch_offset: u32, src
var x_block: array<array<f32, ELEMS_PER_THREAD>, NUM_COLS>; var x_block: array<array<f32, ELEMS_PER_THREAD>, NUM_COLS>;
for (var col = 0u; col < NUM_COLS;col += 1) { for (var col = 0u; col < NUM_COLS;col += 1) {
for (var i = 0u; i < ELEMS_PER_THREAD / 2; i++) { for (var i = 0u; i < ELEMS_PER_THREAD / 2; i++) {
x_block[col][i] = f32(src1[x_base + col * params.stride_11 + i]); x_block[col][i] = f32(SRC1[x_base + col * params.stride_11 + i]);
x_block[col][i + 4] = f32(src1[x_base + col * params.stride_11 + i + 16]); x_block[col][i + 4] = f32(SRC1[x_base + col * params.stride_11 + i + 16]);
} }
} }
for (var row = 0u; row < OUTPUTS_PER_WG; row++) { for (var row = 0u; row < OUTPUTS_PER_WG; row++) {
@ -253,8 +260,8 @@ fn accumulate_vec_dot(thread_id: u32, row_base: u32, src0_batch_offset: u32, src
var x_block: array<array<f32, ELEMS_PER_THREAD>, NUM_COLS>; var x_block: array<array<f32, ELEMS_PER_THREAD>, NUM_COLS>;
for (var col = 0u; col < NUM_COLS;col += 1) { for (var col = 0u; col < NUM_COLS;col += 1) {
for (var i = 0u; i < ELEMS_PER_THREAD / 2; i++) { for (var i = 0u; i < ELEMS_PER_THREAD / 2; i++) {
x_block[col][i] = f32(src1[x_base + col * params.stride_11 + i]); x_block[col][i] = f32(SRC1[x_base + col * params.stride_11 + i]);
x_block[col][i + 4] = f32(src1[x_base + col * params.stride_11 + i + 16]); x_block[col][i + 4] = f32(SRC1[x_base + col * params.stride_11 + i + 16]);
} }
} }
for (var row = 0u; row < OUTPUTS_PER_WG; row++) { for (var row = 0u; row < OUTPUTS_PER_WG; row++) {
@ -302,7 +309,7 @@ fn accumulate_vec_dot(thread_id: u32, row_base: u32, src0_batch_offset: u32, src
var x_block: array<array<f32, ELEMS_PER_THREAD>, NUM_COLS>; var x_block: array<array<f32, ELEMS_PER_THREAD>, NUM_COLS>;
for (var col = 0u; col < NUM_COLS;col += 1) { for (var col = 0u; col < NUM_COLS;col += 1) {
for (var i = 0u; i < ELEMS_PER_THREAD; i++) { for (var i = 0u; i < ELEMS_PER_THREAD; i++) {
x_block[col][i] = f32(src1[x_base + col * params.stride_11 + i]); x_block[col][i] = f32(SRC1[x_base + col * params.stride_11 + i]);
} }
} }
for (var row = 0u; row < OUTPUTS_PER_WG; row++) { for (var row = 0u; row < OUTPUTS_PER_WG; row++) {
@ -347,7 +354,7 @@ fn accumulate_vec_dot(thread_id: u32, row_base: u32, src0_batch_offset: u32, src
var x_block: array<array<f32, ELEMS_PER_THREAD>, NUM_COLS>; var x_block: array<array<f32, ELEMS_PER_THREAD>, NUM_COLS>;
for (var col = 0u; col < NUM_COLS;col += 1) { for (var col = 0u; col < NUM_COLS;col += 1) {
for (var i = 0u; i < ELEMS_PER_THREAD; i++) { for (var i = 0u; i < ELEMS_PER_THREAD; i++) {
x_block[col][i] = f32(src1[x_base + col * params.stride_11 + i]); x_block[col][i] = f32(SRC1[x_base + col * params.stride_11 + i]);
} }
} }
for (var row = 0u; row < OUTPUTS_PER_WG; row++) { for (var row = 0u; row < OUTPUTS_PER_WG; row++) {
@ -409,10 +416,10 @@ fn accumulate_vec_dot(thread_id: u32, row_base: u32, src0_batch_offset: u32, src
var x_block: array<array<f32, 16>, NUM_COLS>; var x_block: array<array<f32, 16>, NUM_COLS>;
for (var col = 0u; col < NUM_COLS;col += 1) { for (var col = 0u; col < NUM_COLS;col += 1) {
for (var i = 0u; i < 4u; i++) { for (var i = 0u; i < 4u; i++) {
x_block[col][i] = f32(src1[x_base + col * params.stride_11 + i]); x_block[col][i] = f32(SRC1[x_base + col * params.stride_11 + i]);
x_block[col][i + 4u] = f32(src1[x_base + col * params.stride_11 + 32u + i]); x_block[col][i + 4u] = f32(SRC1[x_base + col * params.stride_11 + 32u + i]);
x_block[col][i + 8u] = f32(src1[x_base + col * params.stride_11 + 64u + i]); x_block[col][i + 8u] = f32(SRC1[x_base + col * params.stride_11 + 64u + i]);
x_block[col][i + 12u] = f32(src1[x_base + col * params.stride_11 + 96u + i]); x_block[col][i + 12u] = f32(SRC1[x_base + col * params.stride_11 + 96u + i]);
} }
} }
for (var row = 0u; row < OUTPUTS_PER_WG; row++) { for (var row = 0u; row < OUTPUTS_PER_WG; row++) {
@ -518,8 +525,8 @@ fn accumulate_vec_dot(thread_id: u32, row_base: u32, src0_batch_offset: u32, src
var x_block: array<array<f32, 16>, NUM_COLS>; var x_block: array<array<f32, 16>, NUM_COLS>;
for (var col = 0u; col < NUM_COLS;col += 1) { for (var col = 0u; col < NUM_COLS;col += 1) {
for (var i = 0u; i < 8u; i++) { for (var i = 0u; i < 8u; i++) {
x_block[col][i] = f32(src1[x_base + col * params.stride_11 + i]); x_block[col][i] = f32(SRC1[x_base + col * params.stride_11 + i]);
x_block[col][i + 8u] = f32(src1[x_base + col * params.stride_11 + 32u + i]); x_block[col][i + 8u] = f32(SRC1[x_base + col * params.stride_11 + 32u + i]);
} }
} }
for (var row = 0u; row < OUTPUTS_PER_WG; row++) { for (var row = 0u; row < OUTPUTS_PER_WG; row++) {
@ -610,10 +617,10 @@ fn accumulate_vec_dot(thread_id: u32, row_base: u32, src0_batch_offset: u32, src
for (var col = 0u; col < NUM_COLS;col += 1) { for (var col = 0u; col < NUM_COLS;col += 1) {
let col_base = x_base + col * params.stride_11; let col_base = x_base + col * params.stride_11;
for (var i = 0u; i < 4u; i++) { for (var i = 0u; i < 4u; i++) {
x_block[col][i] = f32(src1[col_base + i]); x_block[col][i] = f32(SRC1[col_base + i]);
x_block[col][i + 4u] = f32(src1[col_base + 32u + i]); x_block[col][i + 4u] = f32(SRC1[col_base + 32u + i]);
x_block[col][i + 8u] = f32(src1[col_base + 128u + i]); x_block[col][i + 8u] = f32(SRC1[col_base + 128u + i]);
x_block[col][i + 12u] = f32(src1[col_base + 160u + i]); x_block[col][i + 12u] = f32(SRC1[col_base + 160u + i]);
} }
} }
@ -713,10 +720,10 @@ fn accumulate_vec_dot(thread_id: u32, row_base: u32, src0_batch_offset: u32, src
for (var col = 0u; col < NUM_COLS;col += 1) { for (var col = 0u; col < NUM_COLS;col += 1) {
let col_base = x_base + col * params.stride_11; let col_base = x_base + col * params.stride_11;
for (var i = 0u; i < 4u; i++) { for (var i = 0u; i < 4u; i++) {
x_block[col][i] = f32(src1[col_base + i]); x_block[col][i] = f32(SRC1[col_base + i]);
x_block[col][i + 4u] = f32(src1[col_base + 32u + i]); x_block[col][i + 4u] = f32(SRC1[col_base + 32u + i]);
x_block[col][i + 8u] = f32(src1[col_base + 128u + i]); x_block[col][i + 8u] = f32(SRC1[col_base + 128u + i]);
x_block[col][i + 12u] = f32(src1[col_base + 160u + i]); x_block[col][i + 12u] = f32(SRC1[col_base + 160u + i]);
} }
} }
for (var row = 0u; row < OUTPUTS_PER_WG; row++) { for (var row = 0u; row < OUTPUTS_PER_WG; row++) {
@ -823,10 +830,10 @@ fn accumulate_vec_dot(thread_id: u32, row_base: u32, src0_batch_offset: u32, src
for (var col = 0u; col < NUM_COLS;col += 1) { for (var col = 0u; col < NUM_COLS;col += 1) {
let col_base = x_base + col * params.stride_11; let col_base = x_base + col * params.stride_11;
for (var l = 0u; l < 4u; l++) { for (var l = 0u; l < 4u; l++) {
x_block[col][l] = f32(src1[col_base + l]); x_block[col][l] = f32(SRC1[col_base + l]);
x_block[col][l + 4u] = f32(src1[col_base + 32u + l]); x_block[col][l + 4u] = f32(SRC1[col_base + 32u + l]);
x_block[col][l + 8u] = f32(src1[col_base + 64u + l]); x_block[col][l + 8u] = f32(SRC1[col_base + 64u + l]);
x_block[col][l + 12u] = f32(src1[col_base + 96u + l]); x_block[col][l + 12u] = f32(SRC1[col_base + 96u + l]);
} }
} }
for (var row = 0u; row < OUTPUTS_PER_WG; row++) { for (var row = 0u; row < OUTPUTS_PER_WG; row++) {
@ -899,7 +906,7 @@ fn accumulate_vec_dot(thread_id: u32, row_base: u32, src0_batch_offset: u32, src
var x_block: array<array<f32, 16>, NUM_COLS>; var x_block: array<array<f32, 16>, NUM_COLS>;
for (var col = 0u; col < NUM_COLS;col += 1) { for (var col = 0u; col < NUM_COLS;col += 1) {
for (var i = 0u; i < 16u; i++) { for (var i = 0u; i < 16u; i++) {
x_block[col][i] = f32(src1[x_base + col * params.stride_11 + i]); x_block[col][i] = f32(SRC1[x_base + col * params.stride_11 + i]);
} }
} }
for (var row = 0u; row < OUTPUTS_PER_WG; row++) { for (var row = 0u; row < OUTPUTS_PER_WG; row++) {
@ -960,7 +967,7 @@ fn accumulate_vec_dot(thread_id: u32, row_base: u32, src0_batch_offset: u32, src
var x_block: array<array<f32, 16>, NUM_COLS>; var x_block: array<array<f32, 16>, NUM_COLS>;
for (var col = 0u; col < NUM_COLS;col += 1) { for (var col = 0u; col < NUM_COLS;col += 1) {
for (var i = 0u; i < 16u; i++) { for (var i = 0u; i < 16u; i++) {
x_block[col][i] = f32(src1[x_base + col * params.stride_11 + i]); x_block[col][i] = f32(SRC1[x_base + col * params.stride_11 + i]);
} }
} }
for (var row = 0u; row < OUTPUTS_PER_WG; row++) { for (var row = 0u; row < OUTPUTS_PER_WG; row++) {
@ -1039,7 +1046,7 @@ fn accumulate_vec_dot(thread_id: u32, row_base: u32, src0_batch_offset: u32, src
var x_block: array<array<f32, 16>, NUM_COLS>; var x_block: array<array<f32, 16>, NUM_COLS>;
for (var col = 0u; col < NUM_COLS;col += 1) { for (var col = 0u; col < NUM_COLS;col += 1) {
for (var i = 0u; i < 16u; i++) { for (var i = 0u; i < 16u; i++) {
x_block[col][i] = f32(src1[x_base + col * params.stride_11 + i]); x_block[col][i] = f32(SRC1[x_base + col * params.stride_11 + i]);
} }
} }
for (var row = 0u; row < OUTPUTS_PER_WG; row++) { for (var row = 0u; row < OUTPUTS_PER_WG; row++) {
@ -1101,7 +1108,7 @@ fn accumulate_vec_dot(thread_id: u32, row_base: u32, src0_batch_offset: u32, src
var x_block: array<array<f32, 16>, NUM_COLS>; var x_block: array<array<f32, 16>, NUM_COLS>;
for (var col = 0u; col < NUM_COLS;col += 1) { for (var col = 0u; col < NUM_COLS;col += 1) {
for (var i = 0u; i < 16u; i++) { for (var i = 0u; i < 16u; i++) {
x_block[col][i] = f32(src1[x_base + col * params.stride_11 + i]); x_block[col][i] = f32(SRC1[x_base + col * params.stride_11 + i]);
} }
} }
for (var row = 0u; row < OUTPUTS_PER_WG; row++) { for (var row = 0u; row < OUTPUTS_PER_WG; row++) {
@ -1168,7 +1175,7 @@ fn accumulate_vec_dot(thread_id: u32, row_base: u32, src0_batch_offset: u32, src
var x_block: array<array<f32, 16>, NUM_COLS>; var x_block: array<array<f32, 16>, NUM_COLS>;
for (var col = 0u; col < NUM_COLS;col += 1) { for (var col = 0u; col < NUM_COLS;col += 1) {
for (var i = 0u; i < 16u; i++) { for (var i = 0u; i < 16u; i++) {
x_block[col][i] = f32(src1[x_base + col * params.stride_11 + i]); x_block[col][i] = f32(SRC1[x_base + col * params.stride_11 + i]);
} }
} }
for (var row = 0u; row < OUTPUTS_PER_WG; row++) { for (var row = 0u; row < OUTPUTS_PER_WG; row++) {
@ -1234,7 +1241,7 @@ fn accumulate_vec_dot(thread_id: u32, row_base: u32, src0_batch_offset: u32, src
var x_block: array<array<f32, 16>, NUM_COLS>; var x_block: array<array<f32, 16>, NUM_COLS>;
for (var col = 0u; col < NUM_COLS;col += 1) { for (var col = 0u; col < NUM_COLS;col += 1) {
for (var i = 0u; i < 16u; i++) { for (var i = 0u; i < 16u; i++) {
x_block[col][i] = f32(src1[x_base + col * params.stride_11 + i]); x_block[col][i] = f32(SRC1[x_base + col * params.stride_11 + i]);
} }
} }
for (var row = 0u; row < OUTPUTS_PER_WG; row++) { for (var row = 0u; row < OUTPUTS_PER_WG; row++) {
@ -1302,7 +1309,7 @@ fn accumulate_vec_dot(thread_id: u32, row_base: u32, src0_batch_offset: u32, src
var x_block: array<array<f32, 16>, NUM_COLS>; var x_block: array<array<f32, 16>, NUM_COLS>;
for (var col = 0u; col < NUM_COLS;col += 1) { for (var col = 0u; col < NUM_COLS;col += 1) {
for (var i = 0u; i < 16u; i++) { for (var i = 0u; i < 16u; i++) {
x_block[col][i] = f32(src1[x_base + col * params.stride_11 + i]); x_block[col][i] = f32(SRC1[x_base + col * params.stride_11 + i]);
} }
} }
for (var row = 0u; row < OUTPUTS_PER_WG; row++) { for (var row = 0u; row < OUTPUTS_PER_WG; row++) {
@ -1367,8 +1374,8 @@ fn accumulate_vec_dot(thread_id: u32, row_base: u32, src0_batch_offset: u32, src
var x_block: array<array<f32, ELEMS_PER_THREAD>, NUM_COLS>; var x_block: array<array<f32, ELEMS_PER_THREAD>, NUM_COLS>;
for (var col = 0u; col < NUM_COLS;col += 1) { for (var col = 0u; col < NUM_COLS;col += 1) {
for (var i = 0u; i < ELEMS_PER_THREAD / 2u; i++) { for (var i = 0u; i < ELEMS_PER_THREAD / 2u; i++) {
x_block[col][i] = f32(src1[x_base + col * params.stride_11 + i]); x_block[col][i] = f32(SRC1[x_base + col * params.stride_11 + i]);
x_block[col][i + 4u] = f32(src1[x_base + col * params.stride_11 + i + 16u]); x_block[col][i + 4u] = f32(SRC1[x_base + col * params.stride_11 + i + 16u]);
} }
} }
for (var row = 0u; row < OUTPUTS_PER_WG; row++) { for (var row = 0u; row < OUTPUTS_PER_WG; row++) {
@ -1418,7 +1425,7 @@ fn accumulate_vec_dot(thread_id: u32, row_base: u32, src0_batch_offset: u32, src
var x_block: array<array<f32, 16>, NUM_COLS>; var x_block: array<array<f32, 16>, NUM_COLS>;
for (var col = 0u; col < NUM_COLS;col += 1) { for (var col = 0u; col < NUM_COLS;col += 1) {
for (var i = 0u; i < 16u; i++) { for (var i = 0u; i < 16u; i++) {
x_block[col][i] = f32(src1[x_base + col * params.stride_11 + i]); x_block[col][i] = f32(SRC1[x_base + col * params.stride_11 + i]);
} }
} }
for (var row = 0u; row < OUTPUTS_PER_WG; row++) { for (var row = 0u; row < OUTPUTS_PER_WG; row++) {
@ -1476,8 +1483,8 @@ fn accumulate_vec_dot(thread_id: u32, row_base: u32, src0_batch_offset: u32, src
var x_block: array<array<f32, ELEMS_PER_THREAD>, NUM_COLS>; var x_block: array<array<f32, ELEMS_PER_THREAD>, NUM_COLS>;
for (var col = 0u; col < NUM_COLS;col += 1) { for (var col = 0u; col < NUM_COLS;col += 1) {
for (var i = 0u; i < ELEMS_PER_THREAD / 2; i++) { for (var i = 0u; i < ELEMS_PER_THREAD / 2; i++) {
x_block[col][i] = f32(src1[x_base + col * params.stride_11 + i]); x_block[col][i] = f32(SRC1[x_base + col * params.stride_11 + i]);
x_block[col][i + 4] = f32(src1[x_base + col * params.stride_11 + i + 16]); x_block[col][i + 4] = f32(SRC1[x_base + col * params.stride_11 + i + 16]);
} }
} }
for (var row = 0u; row < OUTPUTS_PER_WG; row++) { for (var row = 0u; row < OUTPUTS_PER_WG; row++) {
@ -1521,8 +1528,8 @@ fn accumulate_vec_dot(thread_id: u32, row_base: u32, src0_batch_offset: u32, src
var x_block: array<array<f32, ELEMS_PER_THREAD>, NUM_COLS>; var x_block: array<array<f32, ELEMS_PER_THREAD>, NUM_COLS>;
for (var col = 0u; col < NUM_COLS;col += 1) { for (var col = 0u; col < NUM_COLS;col += 1) {
for (var i = 0u; i < ELEMS_PER_THREAD / 2; i++) { for (var i = 0u; i < ELEMS_PER_THREAD / 2; i++) {
x_block[col][i] = f32(src1[x_base + col * params.stride_11 + i]); x_block[col][i] = f32(SRC1[x_base + col * params.stride_11 + i]);
x_block[col][i + 8] = f32(src1[x_base + col * params.stride_11 + i + 8]); x_block[col][i + 8] = f32(SRC1[x_base + col * params.stride_11 + i + 8]);
} }
} }
for (var row = 0u; row < OUTPUTS_PER_WG; row++) { for (var row = 0u; row < OUTPUTS_PER_WG; row++) {