CS348Project/llama.cpp
5
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-11-16 11:27:03 +00:00
Code Issues Packages Projects Releases Wiki Activity

sycl: add usage of enqueue_functions extension (#14244)

Browse Source 
* Add header and namespace to use enqueue_functions extension

* Convert submit and parallel_for to use new extension in convert.cpp

* Convert submit and parallel_for to use extension in ggml-sycl.cpp

* Convert submit and parallel_for to use extension in gla.cpp

* Convert submit and parallel_for in mmq.cpp

* Convert submit and parallel_for in mmvq.cpp

* Convert submit and parallel_for in remaining files

* Convert all simple parallel_for to nd_launch from enqueue_functions
extension

* Wrapping extension in general function

Create a general function that enable the enqueue_functions extension if
it is enable in the compiler, otherwise call the general SYCL function
to launch kernels.

---------

Signed-off-by: nscipione <nicolo.scipione@codeplay.com>
This commit is contained in:
Nicolò Scipione
2025-06-20 15:07:21 +02:00
committed by GitHub
parent 6369be0735
commit 8308f98c7f
19 changed files with 750 additions and 986 deletions
Download Patch File Download Diff File Expand all files Collapse all files

116
ggml/src/ggml-sycl/dmmv.cpp
View File

@@ -208,12 +208,10 @@ static void convert_mul_mat_vec_f16_sycl(const void *vx, const dfloat *y,
dpct::has_capability_or_fail(stream->get_device(),
{sycl::aspect::fp16});
stream->parallel_for(
sycl::nd_range<3>(block_nums * block_dims, block_dims),
[=](sycl::nd_item<3> item_ct1) [[sycl::reqd_sub_group_size(WARP_SIZE)]] {
dequantize_mul_mat_vec<1, 1, convert_f16>(vx, y, dst, ncols,
nrows, item_ct1);
});
sycl_parallel_for(stream, sycl::nd_range<3>(block_nums * block_dims, block_dims),
[=](sycl::nd_item<3> item_ct1) [[sycl::reqd_sub_group_size(WARP_SIZE)]] {
dequantize_mul_mat_vec<1, 1, convert_f16>(vx, y, dst, ncols, nrows, item_ct1);
});
}
}
@@ -877,12 +875,11 @@ static void dequantize_mul_mat_vec_q4_0_sycl_reorder(const void *vx, const dfloa
dpct::has_capability_or_fail(stream->get_device(),
{sycl::aspect::fp16});
stream->parallel_for(
sycl::nd_range<3>(block_nums * block_dims, block_dims),
[=](sycl::nd_item<3> item_ct1) [[sycl::reqd_sub_group_size(WARP_SIZE)]] {
dequantize_mul_mat_vec_reorder<QK4_0, QR4_0, dequantize_q4_0_reorder>(
vx, y, dst, ncols, nrows, item_ct1);
});
sycl_parallel_for(stream, sycl::nd_range<3>(block_nums * block_dims, block_dims),
[=](sycl::nd_item<3> item_ct1) [[sycl::reqd_sub_group_size(WARP_SIZE)]] {
dequantize_mul_mat_vec_reorder<QK4_0, QR4_0, dequantize_q4_0_reorder>(vx, y, dst, ncols,
nrows, item_ct1);
});
}
}
@@ -900,12 +897,10 @@ static void dequantize_mul_mat_vec_q4_0_sycl(const void *vx, const dfloat *y,
dpct::has_capability_or_fail(stream->get_device(),
{sycl::aspect::fp16});
stream->parallel_for(
sycl::nd_range<3>(block_nums * block_dims, block_dims),
[=](sycl::nd_item<3> item_ct1) [[sycl::reqd_sub_group_size(WARP_SIZE)]] {
dequantize_mul_mat_vec<QK4_0, QR4_0, dequantize_q4_0>(
vx, y, dst, ncols, nrows, item_ct1);
});
sycl_parallel_for(stream, sycl::nd_range<3>(block_nums * block_dims, block_dims),
[=](sycl::nd_item<3> item_ct1) [[sycl::reqd_sub_group_size(WARP_SIZE)]] {
dequantize_mul_mat_vec<QK4_0, QR4_0, dequantize_q4_0>(vx, y, dst, ncols, nrows, item_ct1);
});
}
}
@@ -921,12 +916,10 @@ static void dequantize_mul_mat_vec_q4_1_sycl(const void *vx, const dfloat *y,
dpct::has_capability_or_fail(stream->get_device(),
{sycl::aspect::fp16});
stream->parallel_for(
sycl::nd_range<3>(block_nums * block_dims, block_dims),
[=](sycl::nd_item<3> item_ct1) [[sycl::reqd_sub_group_size(WARP_SIZE)]] {
dequantize_mul_mat_vec<QK4_1, QR4_1, dequantize_q4_1>(
vx, y, dst, ncols, nrows, item_ct1);
});
sycl_parallel_for(stream, sycl::nd_range<3>(block_nums * block_dims, block_dims),
[=](sycl::nd_item<3> item_ct1) [[sycl::reqd_sub_group_size(WARP_SIZE)]] {
dequantize_mul_mat_vec<QK4_1, QR4_1, dequantize_q4_1>(vx, y, dst, ncols, nrows, item_ct1);
});
}
}
@@ -942,12 +935,10 @@ static void dequantize_mul_mat_vec_q5_0_sycl(const void *vx, const dfloat *y,
dpct::has_capability_or_fail(stream->get_device(),
{sycl::aspect::fp16});
stream->parallel_for(
sycl::nd_range<3>(block_nums * block_dims, block_dims),
[=](sycl::nd_item<3> item_ct1) [[sycl::reqd_sub_group_size(WARP_SIZE)]] {
dequantize_mul_mat_vec<QK5_0, QR5_0, dequantize_q5_0>(
vx, y, dst, ncols, nrows, item_ct1);
});
sycl_parallel_for(stream, sycl::nd_range<3>(block_nums * block_dims, block_dims),
[=](sycl::nd_item<3> item_ct1) [[sycl::reqd_sub_group_size(WARP_SIZE)]] {
dequantize_mul_mat_vec<QK5_0, QR5_0, dequantize_q5_0>(vx, y, dst, ncols, nrows, item_ct1);
});
}
}
@@ -963,12 +954,10 @@ static void dequantize_mul_mat_vec_q5_1_sycl(const void *vx, const dfloat *y,
dpct::has_capability_or_fail(stream->get_device(),
{sycl::aspect::fp16});
stream->parallel_for(
sycl::nd_range<3>(block_nums * block_dims, block_dims),
[=](sycl::nd_item<3> item_ct1) [[sycl::reqd_sub_group_size(WARP_SIZE)]] {
dequantize_mul_mat_vec<QK5_1, QR5_1, dequantize_q5_1>(
vx, y, dst, ncols, nrows, item_ct1);
});
sycl_parallel_for(stream, sycl::nd_range<3>(block_nums * block_dims, block_dims),
[=](sycl::nd_item<3> item_ct1) [[sycl::reqd_sub_group_size(WARP_SIZE)]] {
dequantize_mul_mat_vec<QK5_1, QR5_1, dequantize_q5_1>(vx, y, dst, ncols, nrows, item_ct1);
});
}
}
@@ -984,12 +973,10 @@ static void dequantize_mul_mat_vec_q8_0_sycl(const void *vx, const dfloat *y,
dpct::has_capability_or_fail(stream->get_device(),
{sycl::aspect::fp16});
stream->parallel_for(
sycl::nd_range<3>(block_nums * block_dims, block_dims),
[=](sycl::nd_item<3> item_ct1) [[sycl::reqd_sub_group_size(WARP_SIZE)]] {
dequantize_mul_mat_vec<QK8_0, QR8_0, dequantize_q8_0>(
vx, y, dst, ncols, nrows, item_ct1);
});
sycl_parallel_for(stream, sycl::nd_range<3>(block_nums * block_dims, block_dims),
[=](sycl::nd_item<3> item_ct1) [[sycl::reqd_sub_group_size(WARP_SIZE)]] {
dequantize_mul_mat_vec<QK8_0, QR8_0, dequantize_q8_0>(vx, y, dst, ncols, nrows, item_ct1);
});
}
}
@@ -1002,11 +989,10 @@ static void dequantize_mul_mat_vec_q2_K_sycl(const void *vx, const float *y,
const int block_num_y = (nrows + ny - 1) / ny;
const sycl::range<3> block_nums(1, 1, block_num_y);
const sycl::range<3> block_dims(1, ny, QK_WARP_SIZE);
stream->parallel_for(
sycl::nd_range<3>(block_nums * block_dims, block_dims),
[=](sycl::nd_item<3> item_ct1) [[sycl::reqd_sub_group_size(QK_WARP_SIZE)]] {
dequantize_mul_mat_vec_q2_k(vx, y, dst, ncols, nrows, item_ct1);
});
sycl_parallel_for(stream, sycl::nd_range<3>(block_nums * block_dims, block_dims),
[=](sycl::nd_item<3> item_ct1) [[sycl::reqd_sub_group_size(QK_WARP_SIZE)]] {
dequantize_mul_mat_vec_q2_k(vx, y, dst, ncols, nrows, item_ct1);
});
}
static void dequantize_mul_mat_vec_q3_K_sycl(const void *vx, const float *y,
@@ -1018,11 +1004,10 @@ static void dequantize_mul_mat_vec_q3_K_sycl(const void *vx, const float *y,
const int block_num_y = (nrows + ny - 1) / ny;
const sycl::range<3> block_nums(1, 1, block_num_y);
const sycl::range<3> block_dims(1, ny, QK_WARP_SIZE);
stream->parallel_for(
sycl::nd_range<3>(block_nums * block_dims, block_dims),
[=](sycl::nd_item<3> item_ct1) [[sycl::reqd_sub_group_size(QK_WARP_SIZE)]] {
dequantize_mul_mat_vec_q3_k(vx, y, dst, ncols, nrows, item_ct1);
});
sycl_parallel_for(stream, sycl::nd_range<3>(block_nums * block_dims, block_dims),
[=](sycl::nd_item<3> item_ct1) [[sycl::reqd_sub_group_size(QK_WARP_SIZE)]] {
dequantize_mul_mat_vec_q3_k(vx, y, dst, ncols, nrows, item_ct1);
});
}
static void dequantize_mul_mat_vec_q4_K_sycl(const void *vx, const float *y,
@@ -1034,11 +1019,10 @@ static void dequantize_mul_mat_vec_q4_K_sycl(const void *vx, const float *y,
const int block_num_y = (nrows + ny - 1) / ny;
const sycl::range<3> block_nums(1, 1, block_num_y);
const sycl::range<3> block_dims(1, ny, QK_WARP_SIZE);
stream->parallel_for(
sycl::nd_range<3>(block_nums * block_dims, block_dims),
[=](sycl::nd_item<3> item_ct1) [[sycl::reqd_sub_group_size(QK_WARP_SIZE)]] {
dequantize_mul_mat_vec_q4_k(vx, y, dst, ncols, nrows, item_ct1);
});
sycl_parallel_for(stream, sycl::nd_range<3>(block_nums * block_dims, block_dims),
[=](sycl::nd_item<3> item_ct1) [[sycl::reqd_sub_group_size(QK_WARP_SIZE)]] {
dequantize_mul_mat_vec_q4_k(vx, y, dst, ncols, nrows, item_ct1);
});
}
static void dequantize_mul_mat_vec_q5_K_sycl(const void *vx, const float *y,
@@ -1047,11 +1031,10 @@ static void dequantize_mul_mat_vec_q5_K_sycl(const void *vx, const float *y,
dpct::queue_ptr stream) {
GGML_ASSERT(ncols % QK_K == 0);
const sycl::range<3> block_dims(1, 1, QK_WARP_SIZE);
stream->parallel_for(
sycl::nd_range<3>(sycl::range<3>(1, 1, nrows) * block_dims, block_dims),
[=](sycl::nd_item<3> item_ct1) [[sycl::reqd_sub_group_size(QK_WARP_SIZE)]] {
dequantize_mul_mat_vec_q5_k(vx, y, dst, ncols, item_ct1);
});
sycl_parallel_for(stream, sycl::nd_range<3>(sycl::range<3>(1, 1, nrows) * block_dims, block_dims),
[=](sycl::nd_item<3> item_ct1) [[sycl::reqd_sub_group_size(QK_WARP_SIZE)]] {
dequantize_mul_mat_vec_q5_k(vx, y, dst, ncols, item_ct1);
});
}
static void dequantize_mul_mat_vec_q6_K_sycl(const void *vx, const float *y,
@@ -1063,11 +1046,10 @@ static void dequantize_mul_mat_vec_q6_K_sycl(const void *vx, const float *y,
const int block_num_y = (nrows + ny - 1) / ny;
const sycl::range<3> block_nums(1, 1, block_num_y);
const sycl::range<3> block_dims(1, ny, QK_WARP_SIZE);
stream->parallel_for(
sycl::nd_range<3>(block_nums * block_dims, block_dims),
[=](sycl::nd_item<3> item_ct1) [[sycl::reqd_sub_group_size(QK_WARP_SIZE)]] {
dequantize_mul_mat_vec_q6_k(vx, y, dst, ncols, nrows, item_ct1);
});
sycl_parallel_for(stream, sycl::nd_range<3>(block_nums * block_dims, block_dims),
[=](sycl::nd_item<3> item_ct1) [[sycl::reqd_sub_group_size(QK_WARP_SIZE)]] {
dequantize_mul_mat_vec_q6_k(vx, y, dst, ncols, nrows, item_ct1);
});
}
void ggml_sycl_op_dequantize_mul_mat_vec(