sleepy/llama.cpp
1
0
Fork 0
Code Issues 16 Pull Requests Actions 34 Packages Projects Releases Wiki Activity

CANN: support gated linear attn (#18653)

Browse Source 
* CANN: support gated linear attn

This change adds support for the GGML_OP_GATED_LINEAR_ATTN operator.
The feature was implemented by YushengZhao. Because the previous
submission was based on an outdated codebase, this PR was rebased to
merge.

Co-authored-by: YushengZhao <yusheng.chao@outlook.com>
Co-authored-by: hipudding <huafengchun@gmail.com>

* CANN: optimize OP gla

Optimize gla for high preformance

* Remove unused comments

---------

Co-authored-by: 赵禹昇 <2501112001@cninfer02.localdomain>
Co-authored-by: YushengZhao <yusheng.chao@outlook.com>
This commit is contained in:
hipudding
2026-01-16 16:18:49 +08:00
committed by GitHub
parent 785a710085
commit baa4ba0aec
3 changed files with 186 additions and 161 deletions
Download Patch File Download Diff File Expand all files Collapse all files
ggml/src/ggml-cann/aclnn_ops.cpp
+143 -77
View File
@@ -58,6 +58,7 @@
#include <aclnnop/aclnn_mean.h> #include <aclnnop/aclnn_mean.h>
#include <aclnnop/aclnn_mm.h> #include <aclnnop/aclnn_mm.h>
#include <aclnnop/aclnn_mul.h> #include <aclnnop/aclnn_mul.h>
#include <aclnnop/aclnn_mv.h>
#include <aclnnop/aclnn_permute.h> #include <aclnnop/aclnn_permute.h>
#include <aclnnop/aclnn_pow.h> #include <aclnnop/aclnn_pow.h>
#include <aclnnop/aclnn_pow_tensor_tensor.h> #include <aclnnop/aclnn_pow_tensor_tensor.h>
@@ -2338,20 +2339,21 @@ static void aclnn_rope_cache_init(ggml_backend_cann_context & ctx,
// Step1.2: prepare rope_yarn_ramp, if this part updated, should update theta_scale_tensor. // Step1.2: prepare rope_yarn_ramp, if this part updated, should update theta_scale_tensor.
// TODO: acl_yarn_ramp_tensor use rope cache. // TODO: acl_yarn_ramp_tensor use rope cache.
bool yarn_ramp_tensor_updated = false; bool yarn_ramp_tensor_updated = false;
acl_tensor_ptr acl_yarn_ramp_tensor; acl_tensor_ptr acl_yarn_ramp_tensor;
if (ext_factor != 0 && (theta_scale_updated || ctx.rope_cache.theta_scale_length != theta_scale_length || if (ext_factor != 0 && (theta_scale_updated || ctx.rope_cache.theta_scale_length != theta_scale_length ||
ctx.rope_cache.freq_scale != freq_scale)) { ctx.rope_cache.freq_scale != freq_scale)) {
yarn_ramp_tensor_updated = true; yarn_ramp_tensor_updated = true;
if (ctx.rope_cache.yarn_ramp_cache != nullptr) { if (ctx.rope_cache.yarn_ramp_cache != nullptr) {
ACL_CHECK(aclrtFree(ctx.rope_cache.yarn_ramp_cache)); ACL_CHECK(aclrtFree(ctx.rope_cache.yarn_ramp_cache));
} }
ACL_CHECK(aclrtMalloc(&ctx.rope_cache.yarn_ramp_cache, theta_scale_length * sizeof(float), ACL_MEM_MALLOC_HUGE_FIRST)); ACL_CHECK(aclrtMalloc(&ctx.rope_cache.yarn_ramp_cache, theta_scale_length * sizeof(float),
ACL_MEM_MALLOC_HUGE_FIRST));
// -rope_yarn_ramp // -rope_yarn_ramp
// const float y = (i0 / 2 - low) / MAX(0.001f, high - low); // const float y = (i0 / 2 - low) / MAX(0.001f, high - low);
// return MIN(1, MAX(0, y)) - 1; // return MIN(1, MAX(0, y)) - 1;
acl_yarn_ramp_tensor = acl_yarn_ramp_tensor = ggml_cann_create_tensor(ctx.rope_cache.yarn_ramp_cache, ACL_FLOAT, sizeof(float),
ggml_cann_create_tensor(ctx.rope_cache.yarn_ramp_cache, ACL_FLOAT, sizeof(float), theta_scale_ne, theta_scale_nb, 1); theta_scale_ne, theta_scale_nb, 1);
float zero_value = 0, one_value = 1; float zero_value = 0, one_value = 1;
float denom_safe_value = MAX(0.001f, corr_dims[1] - corr_dims[0]); float denom_safe_value = MAX(0.001f, corr_dims[1] - corr_dims[0]);
acl_scalar_ptr low = ggml_cann_create_scalar(&corr_dims[0], aclDataType::ACL_FLOAT); acl_scalar_ptr low = ggml_cann_create_scalar(&corr_dims[0], aclDataType::ACL_FLOAT);
@@ -2382,8 +2384,8 @@ static void aclnn_rope_cache_init(ggml_backend_cann_context & ctx,
GGML_CANN_CALL_ACLNN_OP(ctx, InplaceMuls, acl_yarn_ramp_tensor.get(), freq_scale_1_sc.get()); GGML_CANN_CALL_ACLNN_OP(ctx, InplaceMuls, acl_yarn_ramp_tensor.get(), freq_scale_1_sc.get());
GGML_CANN_CALL_ACLNN_OP(ctx, InplaceAdds, acl_yarn_ramp_tensor.get(), freq_scale_sc.get(), one.get()); GGML_CANN_CALL_ACLNN_OP(ctx, InplaceAdds, acl_yarn_ramp_tensor.get(), freq_scale_sc.get(), one.get());
} else { } else {
acl_yarn_ramp_tensor = acl_yarn_ramp_tensor = ggml_cann_create_tensor(ctx.rope_cache.yarn_ramp_cache, ACL_FLOAT, sizeof(float),
ggml_cann_create_tensor(ctx.rope_cache.yarn_ramp_cache, ACL_FLOAT, sizeof(float), theta_scale_ne, theta_scale_nb, 1); theta_scale_ne, theta_scale_nb, 1);
} }
// Step 1.3: update theta_scale_tensor according to ext_factor or freq_scale. // Step 1.3: update theta_scale_tensor according to ext_factor or freq_scale.
if (ext_factor != 0) { if (ext_factor != 0) {
@@ -2991,20 +2993,20 @@ void ggml_cann_argmax(ggml_backend_cann_context & ctx, ggml_tensor * dst) {
GGML_CANN_CALL_ACLNN_OP(ctx, ArgMax, acl_src.get(), 3, false, acl_dst.get()); GGML_CANN_CALL_ACLNN_OP(ctx, ArgMax, acl_src.get(), 3, false, acl_dst.get());
} }
void ggml_cann_conv_transpose_1d(ggml_backend_cann_context& ctx, ggml_tensor* dst){ void ggml_cann_conv_transpose_1d(ggml_backend_cann_context & ctx, ggml_tensor * dst) {
ggml_tensor * src0 = dst->src[0]; ggml_tensor * src0 = dst->src[0];
ggml_tensor * src1 = dst->src[1]; ggml_tensor * src1 = dst->src[1];
// stride // stride
int64_t s0 = ((const int32_t*)(dst->op_params))[0]; int64_t s0 = ((const int32_t *) (dst->op_params))[0];
acl_tensor_ptr acl_input = ggml_cann_create_tensor(src1, src1->ne, src1->nb, 3, ACL_FORMAT_NCL); acl_tensor_ptr acl_input = ggml_cann_create_tensor(src1, src1->ne, src1->nb, 3, ACL_FORMAT_NCL);
acl_tensor_ptr acl_weight = ggml_cann_create_tensor(src0, src0->ne, src0->nb, 3, ACL_FORMAT_NCL); acl_tensor_ptr acl_weight = ggml_cann_create_tensor(src0, src0->ne, src0->nb, 3, ACL_FORMAT_NCL);
acl_tensor_ptr acl_dst = ggml_cann_create_tensor(dst, dst->ne, dst->nb, 3, ACL_FORMAT_NCL); acl_tensor_ptr acl_dst = ggml_cann_create_tensor(dst, dst->ne, dst->nb, 3, ACL_FORMAT_NCL);
// get base information of input and kernel // get base information of input and kernel
int64_t input_len = *(src1->ne); int64_t input_len = *(src1->ne);
int64_t dst_len = *(dst->ne); int64_t dst_len = *(dst->ne);
int64_t kernel_size = *(src0->ne); int64_t kernel_size = *(src0->ne);
// set the max kernel size for each conv // set the max kernel size for each conv
@@ -3012,56 +3014,55 @@ void ggml_cann_conv_transpose_1d(ggml_backend_cann_context& ctx, ggml_tensor* ds
// compute the partition of kernel // compute the partition of kernel
int64_t part_num = 1; int64_t part_num = 1;
part_num = (kernel_size + max_kernel_size - 1) / max_kernel_size; part_num = (kernel_size + max_kernel_size - 1) / max_kernel_size;
int64_t strideVal[1]; int64_t strideVal[1];
strideVal[0] = s0; strideVal[0] = s0;
acl_int_array_ptr stride = ggml_cann_create_int_array(strideVal, 1); acl_int_array_ptr stride = ggml_cann_create_int_array(strideVal, 1);
int64_t paddingVal[] = {0}; int64_t paddingVal[] = { 0 };
acl_int_array_ptr padding = ggml_cann_create_int_array(paddingVal, 1); acl_int_array_ptr padding = ggml_cann_create_int_array(paddingVal, 1);
int64_t dilationVal[] = {1}; int64_t dilationVal[] = { 1 };
acl_int_array_ptr dilation = ggml_cann_create_int_array(dilationVal, 1); acl_int_array_ptr dilation = ggml_cann_create_int_array(dilationVal, 1);
bool transposed = true; bool transposed = true;
int64_t groups = 1; int64_t groups = 1;
int8_t cubeMathType = 0; int8_t cubeMathType = 0;
#ifdef ASCEND_310P #ifdef ASCEND_310P
cubeMathType = 1; cubeMathType = 1;
#endif #endif
auto weight_type = ggml_cann_type_mapping(src0->type); auto weight_type = ggml_cann_type_mapping(src0->type);
auto dst_type = ggml_cann_type_mapping(dst->type); auto dst_type = ggml_cann_type_mapping(dst->type);
// slice the kernel to make each conv available // slice the kernel to make each conv available
int64_t slice_dim = -1; int64_t slice_dim = -1;
int64_t slice_start = 0; int64_t slice_start = 0;
int64_t slice_end = max_kernel_size; int64_t slice_end = max_kernel_size;
int64_t slice_step = 1; int64_t slice_step = 1;
int64_t interval = max_kernel_size; int64_t interval = max_kernel_size;
int64_t left_pad_len = dilationVal[0] * (max_kernel_size - 1) + 1 - 2 * paddingVal[0]; int64_t left_pad_len = dilationVal[0] * (max_kernel_size - 1) + 1 - 2 * paddingVal[0];
int64_t right_pad_len = 0; int64_t right_pad_len = 0;
acl_scalar_ptr alpha = nullptr; acl_scalar_ptr alpha = nullptr;
float alphaValue = 1.0; float alphaValue = 1.0;
alpha = ggml_cann_create_scalar(&alphaValue, aclDataType::ACL_FLOAT); alpha = ggml_cann_create_scalar(&alphaValue, aclDataType::ACL_FLOAT);
// set zero to destination // set zero to destination
GGML_CANN_CALL_ACLNN_OP(ctx, InplaceZero, acl_dst.get()); GGML_CANN_CALL_ACLNN_OP(ctx, InplaceZero, acl_dst.get());
for(int k = 0; k < part_num; k++){ for (int k = 0; k < part_num; k++) {
// create part kernel tensor and slice from big kernel // create part kernel tensor and slice from big kernel
slice_start = max_kernel_size * k; slice_start = max_kernel_size * k;
if(k == part_num - 1){ if (k == part_num - 1) {
slice_end = kernel_size; slice_end = kernel_size;
interval = kernel_size - max_kernel_size * k; interval = kernel_size - max_kernel_size * k;
}else{ } else {
slice_end = max_kernel_size * (k+1); slice_end = max_kernel_size * (k + 1);
} }
int64_t part_ne[4]; int64_t part_ne[4];
for(int i = 0; i < 4; i++) { for (int i = 0; i < 4; i++) {
part_ne[i] = *(src0->ne + i); part_ne[i] = *(src0->ne + i);
} }
part_ne[0] = interval; part_ne[0] = interval;
@@ -3074,16 +3075,17 @@ void ggml_cann_conv_transpose_1d(ggml_backend_cann_context& ctx, ggml_tensor* ds
ggml_cann_pool_alloc part_kernel_allocator; ggml_cann_pool_alloc part_kernel_allocator;
part_kernel_allocator.alloc(ctx.pool(), part_nb[3]); part_kernel_allocator.alloc(ctx.pool(), part_nb[3]);
void* part_kernel_buf = part_kernel_allocator.get(); void * part_kernel_buf = part_kernel_allocator.get();
acl_tensor_ptr part_kernel = ggml_cann_create_tensor(part_kernel_buf, weight_type, acl_tensor_ptr part_kernel = ggml_cann_create_tensor(part_kernel_buf, weight_type, ggml_element_size(src0),
ggml_element_size(src0), part_ne, part_nb, 3, ACL_FORMAT_NCL); part_ne, part_nb, 3, ACL_FORMAT_NCL);
GGML_CANN_CALL_ACLNN_OP(ctx, Slice, acl_weight.get(), slice_dim, slice_start, slice_end, slice_step, part_kernel.get()); GGML_CANN_CALL_ACLNN_OP(ctx, Slice, acl_weight.get(), slice_dim, slice_start, slice_end, slice_step,
part_kernel.get());
// create the part conv result tensor // create the part conv result tensor
int64_t part_dst_ne[4]; int64_t part_dst_ne[4];
for(int i = 0; i < 4; i++){ for (int i = 0; i < 4; i++) {
part_dst_ne[i] = *(dst->ne + i); part_dst_ne[i] = *(dst->ne + i);
} }
part_dst_ne[0] = (input_len - 1) * strideVal[0] - 2 * paddingVal[0] + dilationVal[0] * (part_ne[0] - 1) + 1; part_dst_ne[0] = (input_len - 1) * strideVal[0] - 2 * paddingVal[0] + dilationVal[0] * (part_ne[0] - 1) + 1;
@@ -3095,32 +3097,33 @@ void ggml_cann_conv_transpose_1d(ggml_backend_cann_context& ctx, ggml_tensor* ds
} }
ggml_cann_pool_alloc part_dst_allocator; ggml_cann_pool_alloc part_dst_allocator;
part_dst_allocator.alloc(ctx.pool(), part_dst_nb[3]); part_dst_allocator.alloc(ctx.pool(), part_dst_nb[3]);
void* part_dst_buf = part_dst_allocator.get(); void * part_dst_buf = part_dst_allocator.get();
acl_tensor_ptr acl_part_dst = ggml_cann_create_tensor(part_dst_buf, dst_type, ggml_element_size(dst), acl_tensor_ptr acl_part_dst = ggml_cann_create_tensor(part_dst_buf, dst_type, ggml_element_size(dst),
part_dst_ne, part_dst_nb, 3, ACL_FORMAT_NCL); part_dst_ne, part_dst_nb, 3, ACL_FORMAT_NCL);
GGML_CANN_CALL_ACLNN_OP(ctx, InplaceZero, acl_part_dst.get()); GGML_CANN_CALL_ACLNN_OP(ctx, InplaceZero, acl_part_dst.get());
// compute part conv transpose 1d // compute part conv transpose 1d
GGML_CANN_CALL_ACLNN_OP(ctx, Convolution, acl_input.get(), part_kernel.get(), nullptr, stride.get(), GGML_CANN_CALL_ACLNN_OP(ctx, Convolution, acl_input.get(), part_kernel.get(), nullptr, stride.get(),
padding.get(), dilation.get(), transposed, padding.get(), groups, acl_part_dst.get(), cubeMathType); padding.get(), dilation.get(), transposed, padding.get(), groups, acl_part_dst.get(),
cubeMathType);
// compute the position of part result in final result // compute the position of part result in final result
int64_t global_start = slice_start; int64_t global_start = slice_start;
int64_t global_end = std::min((input_len - 1) * strideVal[0] + slice_end, dst_len); int64_t global_end = std::min((input_len - 1) * strideVal[0] + slice_end, dst_len);
left_pad_len = global_start; left_pad_len = global_start;
right_pad_len = dst_len - global_end; right_pad_len = dst_len - global_end;
std::vector<int64_t> padDataVal = {left_pad_len,right_pad_len}; std::vector<int64_t> padDataVal = { left_pad_len, right_pad_len };
acl_int_array_ptr padData = ggml_cann_create_int_array(padDataVal.data(), 2); acl_int_array_ptr padData = ggml_cann_create_int_array(padDataVal.data(), 2);
acl_scalar_ptr pad_value = nullptr; acl_scalar_ptr pad_value = nullptr;
float pad_valueVal = 0.0; float pad_valueVal = 0.0;
pad_value = ggml_cann_create_scalar(&pad_valueVal, aclDataType::ACL_FLOAT); pad_value = ggml_cann_create_scalar(&pad_valueVal, aclDataType::ACL_FLOAT);
int64_t conv_result_ne[4]; int64_t conv_result_ne[4];
for(int i = 0; i < 4; i++){ for (int i = 0; i < 4; i++) {
conv_result_ne[i] = *(dst->ne + i); conv_result_ne[i] = *(dst->ne + i);
} }
@@ -3132,13 +3135,14 @@ void ggml_cann_conv_transpose_1d(ggml_backend_cann_context& ctx, ggml_tensor* ds
ggml_cann_pool_alloc conv_result_allocator; ggml_cann_pool_alloc conv_result_allocator;
conv_result_allocator.alloc(ctx.pool(), conv_result_nb[3]); conv_result_allocator.alloc(ctx.pool(), conv_result_nb[3]);
void* conv_result_buf = conv_result_allocator.get(); void * conv_result_buf = conv_result_allocator.get();
acl_tensor_ptr conv_result = ggml_cann_create_tensor(conv_result_buf, dst_type, ggml_element_size(dst), acl_tensor_ptr conv_result = ggml_cann_create_tensor(conv_result_buf, dst_type, ggml_element_size(dst),
conv_result_ne, conv_result_nb, 3, ACL_FORMAT_NCL); conv_result_ne, conv_result_nb, 3, ACL_FORMAT_NCL);
GGML_CANN_CALL_ACLNN_OP(ctx, InplaceZero, conv_result.get()); GGML_CANN_CALL_ACLNN_OP(ctx, InplaceZero, conv_result.get());
GGML_CANN_CALL_ACLNN_OP(ctx, ConstantPadNd, acl_part_dst.get(), padData.get(), pad_value.get(), conv_result.get()); GGML_CANN_CALL_ACLNN_OP(ctx, ConstantPadNd, acl_part_dst.get(), padData.get(), pad_value.get(),
conv_result.get());
GGML_CANN_CALL_ACLNN_OP(ctx, InplaceAdd, acl_dst.get(), conv_result.get(), alpha.get()); GGML_CANN_CALL_ACLNN_OP(ctx, InplaceAdd, acl_dst.get(), conv_result.get(), alpha.get());
} }
} }
@@ -3742,15 +3746,15 @@ void ggml_cann_ssm_conv(ggml_backend_cann_context & ctx, ggml_tensor * dst) {
// we want a view: ne_w = { nc, 1, nr } // [K, 1, C] // we want a view: ne_w = { nc, 1, nr } // [K, 1, C]
// so that reversed dims -> [C, 1, K] which matches // so that reversed dims -> [C, 1, K] which matches
// [out_channels, in_channels/groups, kernel_size] // [out_channels, in_channels/groups, kernel_size]
int64_t w_ne[GGML_MAX_DIMS] = { nc, 1, nr, 1 }; // [K, 1 input ch. per group, C groups] int64_t w_ne[GGML_MAX_DIMS] = { nc, 1, nr, 1 }; // [K, 1 input ch. per group, C groups]
// Layout: src1 data is [K, C] with // Layout: src1 data is [K, C] with
// offset(k, c) = k*nb0 + c*nb1 // offset(k, c) = k*nb0 + c*nb1
// We want offset_w(k, 0, c) = k*nb0 + c*nb1, // We want offset_w(k, 0, c) = k*nb0 + c*nb1,
// so we can reuse nb0 and nb1, and set nb2 = nb1. // so we can reuse nb0 and nb1, and set nb2 = nb1.
size_t w_nb[GGML_MAX_DIMS] = { src1->nb[0], src1->nb[1], src1->nb[1], src1->nb[3] }; // same as src1 size_t w_nb[GGML_MAX_DIMS] = { src1->nb[0], src1->nb[1], src1->nb[1], src1->nb[3] }; // same as src1
acl_tensor_ptr acl_w = ggml_cann_create_tensor( acl_tensor_ptr acl_w = ggml_cann_create_tensor(src1->data, ggml_cann_type_mapping(src1->type),
src1->data, ggml_cann_type_mapping(src1->type), ggml_type_size(src1->type), w_ne, w_nb, 3, ACL_FORMAT_NCL); ggml_type_size(src1->type), w_ne, w_nb, 3, ACL_FORMAT_NCL);
// 3) Output: dst is { d_inner, n_t, n_s } (CLN) // 3) Output: dst is { d_inner, n_t, n_s } (CLN)
// //
@@ -3768,11 +3772,12 @@ void ggml_cann_ssm_conv(ggml_backend_cann_context & ctx, ggml_tensor * dst) {
// nb_y[0] = nr * sizeof(float); // step in L // nb_y[0] = nr * sizeof(float); // step in L
// nb_y[1] = sizeof(float); // step in C // nb_y[1] = sizeof(float); // step in C
// nb_y[2] = nr * n_t * sizeof(float); // step in N // nb_y[2] = nr * n_t * sizeof(float); // step in N
int64_t y_ne[GGML_MAX_DIMS] = { n_t, nr, n_s, 1 }; // [L_out, C, N] int64_t y_ne[GGML_MAX_DIMS] = { n_t, nr, n_s, 1 }; // [L_out, C, N]
size_t y_nb[GGML_MAX_DIMS] = { dst->ne[0] * sizeof(float), sizeof(float), dst->ne[0] * dst->ne[1] * sizeof(float), dst->nb[3] }; // [nr, 1, nr * n_t] size_t y_nb[GGML_MAX_DIMS] = { dst->ne[0] * sizeof(float), sizeof(float), dst->ne[0] * dst->ne[1] * sizeof(float),
dst->nb[3] }; // [nr, 1, nr * n_t]
acl_tensor_ptr acl_y = ggml_cann_create_tensor( acl_tensor_ptr acl_y = ggml_cann_create_tensor(dst->data, ggml_cann_type_mapping(dst->type),
dst->data, ggml_cann_type_mapping(dst->type), ggml_type_size(dst->type), y_ne, y_nb, 3, ACL_FORMAT_NCL); ggml_type_size(dst->type), y_ne, y_nb, 3, ACL_FORMAT_NCL);
// --- Conv1d parameters: depthwise, stride 1, no padding ("valid") --- // --- Conv1d parameters: depthwise, stride 1, no padding ("valid") ---
int64_t strideVal[1] = { 1 }; int64_t strideVal[1] = { 1 };
@@ -3791,22 +3796,15 @@ void ggml_cann_ssm_conv(ggml_backend_cann_context & ctx, ggml_tensor * dst) {
cubeMathType = 1; cubeMathType = 1;
#endif #endif
GGML_CANN_CALL_ACLNN_OP(ctx, GGML_CANN_CALL_ACLNN_OP(ctx, Convolution,
Convolution,
acl_x.get(), // input: N, C, L_in = ncs acl_x.get(), // input: N, C, L_in = ncs
acl_w.get(), // weight: [C, 1, K] with groups=nr acl_w.get(), // weight: [C, 1, K] with groups=nr
nullptr, // bias nullptr, // bias
stride.get(), stride.get(), padding.get(), dilation.get(), transposed,
padding.get(), padding.get(), // output padding (unused for non-transposed)
dilation.get(), groups, acl_y.get(), cubeMathType);
transposed,
padding.get(), // output padding (unused for non-transposed)
groups,
acl_y.get(),
cubeMathType);
} }
void ggml_cann_op_add_rms_norm_fused(ggml_backend_cann_context & ctx, void ggml_cann_op_add_rms_norm_fused(ggml_backend_cann_context & ctx,
ggml_tensor * add_node, ggml_tensor * add_node,
ggml_tensor * rms_norm_node) { ggml_tensor * rms_norm_node) {
@@ -3860,3 +3858,71 @@ void ggml_cann_op_add_rms_norm_fused(ggml_backend_cann_context & ctx,
eps, // double type eps, // double type
acl_yout.get(), acl_rstd.get(), acl_xout.get()); acl_yout.get(), acl_rstd.get(), acl_xout.get());
} }
void ggml_cann_gated_linear_attn(ggml_backend_cann_context & ctx, ggml_tensor * dst) {
ggml_tensor * k = dst->src[0];
ggml_tensor * v = dst->src[1];
ggml_tensor * q = dst->src[2];
ggml_tensor * g = dst->src[3];
ggml_tensor * s = dst->src[4];
int64_t B = dst->src[4]->ne[1];
int64_t T = dst->src[0]->ne[2];
int64_t H = dst->src[0]->ne[1];
int64_t C = dst->ne[0];
int64_t D = C / H;
int64_t L = T / B;
int64_t ne_qkg[2] = { 1, D };
int64_t ne_s[2] = { D, D };
int64_t ne_st[2] = { ne_s[1], ne_s[0] };
int64_t ne_vo[2] = { D, 1 };
int64_t ne_q[1] = { D };
size_t nb_base = ggml_type_size(k->type);
size_t nb_qkg[2] = { nb_base, nb_base };
size_t nb_s[2] = { nb_base, D * nb_base };
size_t nb_st[2] = { nb_s[1], nb_s[0] };
size_t nb_vo[2] = { nb_base, D * nb_base };
size_t nb_q[1] = { nb_base };
const float scale = ggml_get_op_params_f32(dst, 0);
acl_tensor_ptr acl_s = ggml_cann_create_tensor(s, s->ne, s->nb, 2, ACL_FORMAT_ND);
acl_tensor_ptr new_state = ggml_cann_create_tensor(dst, s->ne, s->nb, 2, ACL_FORMAT_ND, (B * L * H * D) * nb_base);
cann_copy(ctx, acl_s.get(), new_state.get());
for (int64_t b = 0; b < B; b++) {
for (int64_t h = 0; h < H; h++) {
size_t s_offset = (b * (H * D * D) + h * (D * D)) * nb_base;
// D * D
acl_tensor_ptr acl_s_new =
ggml_cann_create_tensor(dst, ne_s, nb_s, 2, ACL_FORMAT_ND, (B * L * H * D) * nb_base + s_offset);
acl_tensor_ptr acl_s_new_t =
ggml_cann_create_tensor(dst, ne_st, nb_st, 2, ACL_FORMAT_ND, (B * L * H * D) * nb_base + s_offset);
for (int64_t l = 0; l < L; l++) {
size_t qkvgo_offset = (b * (L * H * D) + l * (H * D) + h * (D)) * nb_base;
// D * 1
acl_tensor_ptr acl_k = ggml_cann_create_tensor(k, ne_qkg, nb_qkg, 2, ACL_FORMAT_ND, qkvgo_offset);
acl_tensor_ptr acl_g = ggml_cann_create_tensor(g, ne_qkg, nb_qkg, 2, ACL_FORMAT_ND, qkvgo_offset);
// D
acl_tensor_ptr acl_q = ggml_cann_create_tensor(q, ne_q, nb_q, 1, ACL_FORMAT_ND, qkvgo_offset);
// 1 * D
acl_tensor_ptr acl_v = ggml_cann_create_tensor(v, ne_vo, nb_vo, 2, ACL_FORMAT_ND, qkvgo_offset);
// D
acl_tensor_ptr acl_o = ggml_cann_create_tensor(dst, ne_q, nb_q, 1, ACL_FORMAT_ND, qkvgo_offset);
// k ⊗ v
size_t buf_size = D * D * nb_base;
ggml_cann_pool_alloc buffer_allocator(ctx.pool(), buf_size);
acl_tensor_ptr tmp_tensor = ggml_cann_create_tensor(
buffer_allocator.get(), ggml_cann_type_mapping(k->type), nb_base, ne_s, nb_s, 2);
aclnn_mul(ctx, acl_k.get(), acl_v.get(), tmp_tensor.get());
//s_new = g ⊗ s_old + k ⊗ v
aclnn_mul(ctx, acl_s_new.get(), acl_g.get(), nullptr);
aclnn_add(ctx, acl_s_new.get(), tmp_tensor.get(), nullptr);
// compute output
GGML_CANN_CALL_ACLNN_OP(ctx, Mv, acl_s_new_t.get(), acl_q.get(), acl_o.get(), 1);
aclnn_muls(ctx, acl_o.get(), scale, nullptr, true);
}
}
}
}
ggml/src/ggml-cann/aclnn_ops.h
+39 -84
View File
@@ -814,67 +814,20 @@ void ggml_cann_step(ggml_backend_cann_context & ctx, ggml_tensor * dst);
*/ */
void ggml_cann_flash_attn_ext(ggml_backend_cann_context & ctx, ggml_tensor * dst); void ggml_cann_flash_attn_ext(ggml_backend_cann_context & ctx, ggml_tensor * dst);
/*
* @brief A generic wrapper for ACL resources with custom deleter support.
*/
using any_acl_resource = std::unique_ptr<void, std::function<void(void *)>>;
/** /**
* @brief Trait structure used to define how to destroy a given ACL resource type. * @brief Forward Gated Linear Attention on the CANN backend.
* *
* @tparam T ACL resource type. * Expects dst->src[0..4] = {k, v, q, g, s} with shape conventions:
*/ * k, v, q, g: [D] with outer dims T x H batched as ne[2]=T, ne[1]=H
template <typename T> struct acl_resource_traits; * s: initial state [B, H, D, D], where B is batch and D=C/H
* dst holds both outputs (o) and updated state; a scale factor is read from op params.
/**
* @brief Specialization for aclTensor, defines how to destroy an aclTensor resource.
*/
template <> struct acl_resource_traits<aclTensor> {
static void destroy(void * p) { ACL_CHECK(aclDestroyTensor(static_cast<aclTensor *>(p))); }
};
/**
* @brief Specialization for aclIntArray, defines how to destroy an aclIntArray resource.
*/
template <> struct acl_resource_traits<aclIntArray> {
static void destroy(void * p) { ACL_CHECK(aclDestroyIntArray(static_cast<aclIntArray *>(p))); }
};
/**
* @brief Specialization for aclScalar, defines how to destroy an aclScalar resource.
*/
template <> struct acl_resource_traits<aclScalar> {
static void destroy(void * p) { ACL_CHECK(aclDestroyScalar(static_cast<aclScalar *>(p))); }
};
/**
* @brief Specialization for aclTensorList, defines how to destroy an aclTensorList resource.
*/
template <> struct acl_resource_traits<aclTensorList> {
static void destroy(void * p) { ACL_CHECK(aclDestroyTensorList(static_cast<aclTensorList *>(p))); }
};
/**
* @brief Creates a generic ACL resource wrapper with proper destruction logic.
* *
* @tparam T ACL resource type. * The kernel updates per time step l: S_new = g ⊗ S_old + k ⊗ v, then computes o = (S_new^T q) * scale.
* @param ptr Raw pointer to ACL resource.
* @return any_acl_resource Smart pointer that handles destruction.
*/
template <typename T> any_acl_resource make_acl_resource(T * ptr) {
return any_acl_resource(static_cast<void *>(ptr), [](void * p) { acl_resource_traits<T>::destroy(p); });
}
/**
* @brief Registers multiple ACL resources into a vector for lifetime management.
* *
* @tparam Args Variadic list of ACL resource types. * @param ctx Backend context providing stream/allocator utilities.
* @param vec Target vector to hold ACL resources. * @param dst Output tensor; src deps are k, v, q, g, s as above.
* @param args Raw pointers to ACL resources.
*/ */
template <typename... Args> void register_acl_resources(std::vector<any_acl_resource> & vec, Args *... args) { void ggml_cann_gated_linear_attn(ggml_backend_cann_context & ctx, ggml_tensor * dst);
(vec.emplace_back(make_acl_resource(args)), ...);
}
/** /**
* @brief Launches an asynchronous task using the memory allocator. * @brief Launches an asynchronous task using the memory allocator.
@@ -894,19 +847,19 @@ template <typename... Args> void register_acl_resources(std::vector<any_acl_reso
* same stream are executed in queue order. * same stream are executed in queue order.
*/ */
#define GGML_CANN_CALL_ACLNN_OP(CTX, OP_NAME, ...) \ # define GGML_CANN_CALL_ACLNN_OP(CTX, OP_NAME, ...) \
do { \ do { \
uint64_t workspaceSize = 0; \ uint64_t workspaceSize = 0; \
aclOpExecutor * executor; \ aclOpExecutor * executor; \
void * workspaceAddr = nullptr; \ void * workspaceAddr = nullptr; \
ACL_CHECK(aclnn##OP_NAME##GetWorkspaceSize(__VA_ARGS__, &workspaceSize, &executor)); \ ACL_CHECK(aclnn##OP_NAME##GetWorkspaceSize(__VA_ARGS__, &workspaceSize, &executor)); \
/* workspace should alloced in main thread to keep malloc order when using vmm. */ \ /* workspace should alloced in main thread to keep malloc order when using vmm. */ \
if (workspaceSize > 0) { \ if (workspaceSize > 0) { \
ggml_cann_pool_alloc workspace_allocator(CTX.pool(), workspaceSize); \ ggml_cann_pool_alloc workspace_allocator(CTX.pool(), workspaceSize); \
workspaceAddr = workspace_allocator.get(); \ workspaceAddr = workspace_allocator.get(); \
} \ } \
ACL_CHECK(aclnn##OP_NAME(workspaceAddr, workspaceSize, executor, CTX.stream())); \ ACL_CHECK(aclnn##OP_NAME(workspaceAddr, workspaceSize, executor, CTX.stream())); \
} while (0) } while (0)
/** /**
* @brief Performs sparse expert-based matrix multiplication using the CANN backend. * @brief Performs sparse expert-based matrix multiplication using the CANN backend.
@@ -947,7 +900,9 @@ void ggml_cann_mul_mat_id(ggml_backend_cann_context & ctx, ggml_tensor * dst);
* @param rms_norm_tensor The RMS_NORM operation node, contains the gamma weights * @param rms_norm_tensor The RMS_NORM operation node, contains the gamma weights
* and epsilon parameter. * and epsilon parameter.
*/ */
void ggml_cann_op_add_rms_norm_fused(ggml_backend_cann_context & ctx, ggml_tensor * add_node, ggml_tensor * rms_norm_node); void ggml_cann_op_add_rms_norm_fused(ggml_backend_cann_context & ctx,
ggml_tensor * add_node,
ggml_tensor * rms_norm_node);
/** /**
* @brief Check whether a tensor is a weight tensor for matrix multiplication. * @brief Check whether a tensor is a weight tensor for matrix multiplication.
@@ -1104,13 +1059,13 @@ void ggml_cann_op_unary_gated(std::function<void(ggml_backend_cann_context &, ac
* @see ggml_cann_op_unary * @see ggml_cann_op_unary
* @see GGML_CANN_CALL_ACLNN_OP * @see GGML_CANN_CALL_ACLNN_OP
*/ */
#define GGML_CANN_CALL_OP_UNARY(OP_NAME) \ # define GGML_CANN_CALL_OP_UNARY(OP_NAME) \
do { \ do { \
auto lambda = [](ggml_backend_cann_context & ctx, aclTensor * acl_src, aclTensor * acl_dst) { \ auto lambda = [](ggml_backend_cann_context & ctx, aclTensor * acl_src, aclTensor * acl_dst) { \
GGML_CANN_CALL_ACLNN_OP(ctx, OP_NAME, acl_src, acl_dst); \ GGML_CANN_CALL_ACLNN_OP(ctx, OP_NAME, acl_src, acl_dst); \
}; \ }; \
ggml_cann_op_unary(lambda, ctx, dst); \ ggml_cann_op_unary(lambda, ctx, dst); \
} while (0) } while (0)
/** /**
* @brief Helper macro to call a gated unary ACL operator via ggml_cann_op_unary_gated. * @brief Helper macro to call a gated unary ACL operator via ggml_cann_op_unary_gated.
@@ -1133,13 +1088,13 @@ void ggml_cann_op_unary_gated(std::function<void(ggml_backend_cann_context &, ac
* @see ggml_cann_op_unary_gated * @see ggml_cann_op_unary_gated
* @see GGML_CANN_CALL_ACLNN_OP * @see GGML_CANN_CALL_ACLNN_OP
*/ */
#define GGML_CANN_CALL_OP_UNARY_GATED(OP_NAME) \ # define GGML_CANN_CALL_OP_UNARY_GATED(OP_NAME) \
do { \ do { \
auto lambda = [](ggml_backend_cann_context & ctx, aclTensor * acl_src, aclTensor * acl_dst) { \ auto lambda = [](ggml_backend_cann_context & ctx, aclTensor * acl_src, aclTensor * acl_dst) { \
GGML_CANN_CALL_ACLNN_OP(ctx, OP_NAME, acl_src, acl_dst); \ GGML_CANN_CALL_ACLNN_OP(ctx, OP_NAME, acl_src, acl_dst); \
}; \ }; \
ggml_cann_op_unary_gated(lambda, ctx, dst); \ ggml_cann_op_unary_gated(lambda, ctx, dst); \
} while (0) } while (0)
#endif // CANN_ACLNN_OPS #endif // CANN_ACLNN_OPS
ggml/src/ggml-cann/ggml-cann.cpp
+4
View File
@@ -1889,6 +1889,9 @@ static bool ggml_cann_compute_forward(ggml_backend_cann_context & ctx, struct gg
case GGML_OP_OUT_PROD: case GGML_OP_OUT_PROD:
ggml_cann_out_prod(ctx, dst); ggml_cann_out_prod(ctx, dst);
break; break;
case GGML_OP_GATED_LINEAR_ATTN:
ggml_cann_gated_linear_attn(ctx, dst);
break;
case GGML_OP_SSM_CONV: case GGML_OP_SSM_CONV:
ggml_cann_ssm_conv(ctx, dst); ggml_cann_ssm_conv(ctx, dst);
break; break;
@@ -2454,6 +2457,7 @@ static bool ggml_backend_cann_supports_op(ggml_backend_dev_t dev, const ggml_ten
case GGML_OP_MEAN: case GGML_OP_MEAN:
case GGML_OP_PAD_REFLECT_1D: case GGML_OP_PAD_REFLECT_1D:
case GGML_OP_COUNT_EQUAL: case GGML_OP_COUNT_EQUAL:
case GGML_OP_GATED_LINEAR_ATTN:
return true; return true;
case GGML_OP_OUT_PROD: case GGML_OP_OUT_PROD:
{ {