Open
Description
Describe the bug
From looking at the source code, there seems to support --expt-extended-lambda but now CUDA has deprecated the above and moved to --extended-lambda. I was wondering if this can be added too to the recognized flags in DPCT.
DPCT Error while doing this command: dpct --extra-arg="-I __CUDA_ARCH__=800" -p ./llama.cpp/build --enable-profiling --use-experimental-features=all --in-root=./llama.cpp --out-root=./dpct_out --migrate-build-script=CMake --process-all
error: unknown argument: '-extended-lambda'
The compilation command from the database on the CUDA side:
"command": "nvcc -c -DGGML_BACKEND_BUILD -DGGML_BACKEND_SHARED -DGGML_CUDA_PEER_MAX_BATCH_SIZE=128 -DGGML_CUDA_USE_GRAPHS -DGGML_SCHED_MAX_COPIES=4 -DGGML_SHARED -D_GNU_SOURCE -D_XOPEN_SOURCE=600 -Dggml_cuda_EXPORTS -I/lus/eagle/projects/NWChemEx_aesp/abagusetty/llama-cuda-to-sycl/llama.cpp/ggml/src/ggml-cuda/.. -I/lus/eagle/projects/NWChemEx_aesp/abagusetty/llama-cuda-to-sycl/llama.cpp/ggml/src/../include -isystem /opt/nvidia/hpc_sdk/Linux_x86_64/23.9/cuda/12.2/include -O3 -DNDEBUG -std=c++17 -extended-lambda -Wmissing-declarations -Wmissing-noreturn -Wall -Wextra -Wpedantic -Wcast-qual -Wno-unused-function -Wno-array-bounds -Wno-pedantic -o CMakeFiles/ggml-cuda.dir/template-instances/fattn-vec-f16-instance-hs256-f16-f16.cu.o -D__CUDACC__=1 /lus/eagle/projects/NWChemEx_aesp/abagusetty/llama-cuda-to-sycl/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs256-f16-f16.cu",
To reproduce
- Include code snippet as short as possible
- Specify the command which should be used to compile the program
- Specify the command which should be used to launch the program
- Indicate what is wrong and what was expected
Environment
- OS: [e.g Windows/Linux]
- Target device and vendor: [e.g. Nvidia GPU]
- DPC++ version: [e.g. commit hash or output of
clang++ --version] - Dependencies version: [e.g. the output of
nvidia-smiandsycl-ls --verbose]
Additional context
No response