diff --git a/ggml-sycl.cpp b/ggml-sycl.cpp index 28c9e7606..1931b8094 100644 --- a/ggml-sycl.cpp +++ b/ggml-sycl.cpp @@ -9712,7 +9712,7 @@ static void ggml_sycl_op_flatten(const ggml_tensor *src0, src1_ddf = (float *) src1_extra->data_device[g_main_device_index]; } else { src1_ddf = src1_f.alloc(ggml_nelements(src1)); - // SYCL_CHECK(ggml_sycl_cpy_tensor_2d(src1_ddf, src1, 0, 0, 0, nrows1, main_stream)); + SYCL_CHECK(ggml_sycl_cpy_tensor_2d(src1_ddf, src1, 0, 0, 0, nrows1, main_stream)); } } if (dst_on_device) { diff --git a/run.sh b/run.sh index 38315a465..254a18efc 100755 --- a/run.sh +++ b/run.sh @@ -11,9 +11,9 @@ else export GGML_SYCL_DEVICE=0 fi echo GGML_SYCL_DEVICE=$GGML_SYCL_DEVICE -export GGML_SYCL_DEBUG=1 +#export GGML_SYCL_DEBUG=1 #export GGML_SYCL_LIST_DEVICE=1 #./build/bin/main -m models/llama-2-7b.Q4_0.gguf -p "${INPUT1}" -e -n 400 -ngl 33 -c 2048 -#./build/bin/main -m models/llama-2-7b.Q4_0.gguf -p "${INPUT2}" -n 400 -e -ngl 33 -./build/bin/main -m models/llama-2-7b.Q4_0.gguf -p "${INPUT2}" -n 5 -e -ngl 33 -t 1 -s 0 +./build/bin/main -m models/llama-2-7b.Q4_0.gguf -p "${INPUT2}" -n 400 -e -ngl 33 +#./build/bin/main -m models/llama-2-7b.Q4_0.gguf -p "${INPUT2}" -n 5 -e -ngl 33 -t 1 -s 0