Skip to content
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
2 changes: 1 addition & 1 deletion ggml/src/ggml-sycl/common.hpp
Original file line number Diff line number Diff line change
Expand Up @@ -35,7 +35,7 @@
void* ggml_sycl_host_malloc(size_t size);
void ggml_sycl_host_free(void* ptr);

static int g_ggml_sycl_debug = 0;
extern int g_ggml_sycl_debug;
#define GGML_SYCL_DEBUG(...) \
do { \
if (g_ggml_sycl_debug) \
Expand Down
3 changes: 2 additions & 1 deletion ggml/src/ggml-sycl/ggml-sycl.cpp
Original file line number Diff line number Diff line change
Expand Up @@ -41,6 +41,7 @@
#include "ggml-sycl/gemm.hpp"

static bool g_sycl_loaded = false;
int g_ggml_sycl_debug = 0;

static ggml_sycl_device_info ggml_sycl_init() {
ggml_sycl_device_info info = {};
Expand Down Expand Up @@ -157,8 +158,8 @@ static void ggml_check_sycl() try {
static bool initialized = false;

if (!initialized) {
GGML_SYCL_DEBUG("[SYCL] call ggml_check_sycl\n");
g_ggml_sycl_debug = get_sycl_env("GGML_SYCL_DEBUG", 0);
GGML_SYCL_DEBUG("[SYCL] call ggml_check_sycl\n");
GGML_LOG_INFO("GGML_SYCL_DEBUG: %d\n", g_ggml_sycl_debug);
#if defined(GGML_SYCL_FORCE_MMQ)
GGML_LOG_INFO("GGML_SYCL_FORCE_MMQ: yes\n");
Expand Down
3 changes: 3 additions & 0 deletions ggml/src/ggml-sycl/softmax.cpp
Original file line number Diff line number Diff line change
Expand Up @@ -249,13 +249,16 @@ void ggml_sycl_op_soft_max(ggml_backend_sycl_context & ctx, ggml_tensor * dst) {

if (dst->src[1] && dst->src[1]->type == GGML_TYPE_F16) {
const sycl::half * src1_dd = static_cast<sycl::half *>(dst->src[1]->data);
GGML_SYCL_DEBUG("%s: F16 mask\n", __func__);
soft_max_f32_sycl<sycl::half>(src0_dd, src1_dd, dst_dd, ne00, nrows_x, nrows_y, scale, max_bias,
main_stream, ctx.device);
} else if (dst->src[1] && dst->src[1]->type == GGML_TYPE_F32) {
const float * src1_dd = static_cast<const float *>(dst->src[1]->data);
GGML_SYCL_DEBUG("%s: F32 mask\n", __func__);
soft_max_f32_sycl<float>(src0_dd, src1_dd, dst_dd, ne00, nrows_x, nrows_y, scale, max_bias, main_stream, ctx.device);
} else {
/* mask unavailable */
GGML_SYCL_DEBUG("%s: No mask\n", __func__);
soft_max_f32_sycl<float>(src0_dd, nullptr, dst_dd, ne00, nrows_x, nrows_y, scale, max_bias, main_stream, ctx.device);
}
}
Loading