3 NewOffloadDriver tests failing on CUDA #16413
Description
Describe the bug
These tests are marked as XFAIL on CUDA. Upon resolution, please remove XFAIL from the test source:
SYCL :: NewOffloadDriver/multisource.cpp
SYCL :: NewOffloadDriver/split-per-source-main.cpp
SYCL :: NewOffloadDriver/sycl-external-with-optional-features.cpp
To reproduce
Similar error for all 3:
FAIL: SYCL :: NewOffloadDriver/multisource.cpp (1680 of 2242)
******************** TEST 'SYCL :: NewOffloadDriver/multisource.cpp' FAILED ********************
Exit Code: 1
Command Output (stdout):
--
# RUN: at line 10
/__w/llvm/llvm/toolchain/bin//clang++ -Werror -fsycl -fsycl-targets=nvptx64-nvidia-cuda /__w/llvm/llvm/llvm/sycl/test-e2e/NewOffloadDriver/multisource.cpp --offload-new-driver -c -o /__w/llvm/llvm/build-e2e/NewOffloadDriver/Output/multisource.cpp.tmp.kernel.o -DINIT_KERNEL -DCALC_KERNEL
# executed command: /__w/llvm/llvm/toolchain/bin//clang++ -Werror -fsycl -fsycl-targets=nvptx64-nvidia-cuda /__w/llvm/llvm/llvm/sycl/test-e2e/NewOffloadDriver/multisource.cpp --offload-new-driver -c -o /__w/llvm/llvm/build-e2e/NewOffloadDriver/Output/multisource.cpp.tmp.kernel.o -DINIT_KERNEL -DCALC_KERNEL
# note: command had no output on stdout or stderr
# RUN: at line 11
/__w/llvm/llvm/toolchain/bin//clang++ -Werror -fsycl -fsycl-targets=nvptx64-nvidia-cuda /__w/llvm/llvm/llvm/sycl/test-e2e/NewOffloadDriver/multisource.cpp --offload-new-driver -c -o /__w/llvm/llvm/build-e2e/NewOffloadDriver/Output/multisource.cpp.tmp.main.o -DMAIN_APP
# executed command: /__w/llvm/llvm/toolchain/bin//clang++ -Werror -fsycl -fsycl-targets=nvptx64-nvidia-cuda /__w/llvm/llvm/llvm/sycl/test-e2e/NewOffloadDriver/multisource.cpp --offload-new-driver -c -o /__w/llvm/llvm/build-e2e/NewOffloadDriver/Output/multisource.cpp.tmp.main.o -DMAIN_APP
# note: command had no output on stdout or stderr
# RUN: at line 12
/__w/llvm/llvm/toolchain/bin//clang++ -Werror -Wno-error=unused-command-line-argument -fsycl -fsycl-targets=nvptx64-nvidia-cuda --offload-new-driver /__w/llvm/llvm/build-e2e/NewOffloadDriver/Output/multisource.cpp.tmp.kernel.o /__w/llvm/llvm/build-e2e/NewOffloadDriver/Output/multisource.cpp.tmp.main.o -o /__w/llvm/llvm/build-e2e/NewOffloadDriver/Output/multisource.cpp.tmp1.fat
# executed command: /__w/llvm/llvm/toolchain/bin//clang++ -Werror -Wno-error=unused-command-line-argument -fsycl -fsycl-targets=nvptx64-nvidia-cuda --offload-new-driver /__w/llvm/llvm/build-e2e/NewOffloadDriver/Output/multisource.cpp.tmp.kernel.o /__w/llvm/llvm/build-e2e/NewOffloadDriver/Output/multisource.cpp.tmp.main.o -o /__w/llvm/llvm/build-e2e/NewOffloadDriver/Output/multisource.cpp.tmp1.fat
# .---command stderr------------
# | /__w/llvm/llvm/toolchain/bin/clang-linker-wrapper: warning: Compatible SYCL device library binary not found
# | ptxas /tmp/lit-tmp-ukydi1bc/multisource.cpp.tmp1.fat.nvptx64.sm_50-65fb11.img, line 1; fatal : Missing .version directive at start of file '/tmp/lit-tmp-ukydi1bc/multisource.cpp.tmp1.fat.nvptx64.sm_50-65fb11.img'
# | ptxas fatal : Ptx assembly aborted due to errors
# | /__w/llvm/llvm/toolchain/bin/clang-linker-wrapper: error: 'ptxas' failed
# | clang++: error: linker command failed with exit code 1 (use -v to see invocation)
# `-----------------------------
# error: command failed with exit status: 1
Environment
- OS: [e.g Windows/Linux]
- Target device and vendor: NVIDIA CUDA
- DPC++ version: da317d04df8c6ffc873dfe1943c1c7c26e8de1e5
- Dependencies version:
- 2024-12-17T21:58:43.7814040Z �[36;1msycl-ls --verbose�[0m
2024-12-17T21:58:43.7814444Z shell: sh -e {0}
2024-12-17T21:58:43.7814623Z env:
2024-12-17T21:58:43.7815136Z PATH: /__w/llvm/llvm/toolchain/bin/:/usr/local/nvidia/bin:/usr/local/cuda/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin
2024-12-17T21:58:43.7815893Z LD_LIBRARY_PATH: /__w/llvm/llvm/toolchain/lib/:/usr/local/nvidia/lib:/usr/local/nvidia/lib64
2024-12-17T21:58:43.7816306Z ##[endgroup]
2024-12-17T21:58:44.7889537Z [cuda:gpu][cuda:0] NVIDIA CUDA BACKEND, NVIDIA GeForce RTX 3090 8.6 [CUDA 12.2]
2024-12-17T21:58:44.7890811Z [native_cpu:cpu][native_cpu:0] SYCL_NATIVE_CPU, SYCL Native CPU 0.1 [0.0.0]
2024-12-17T21:58:44.7891600Z
2024-12-17T21:58:44.7891859Z Platforms: 2
2024-12-17T21:58:44.7892384Z Platform [#1]:
2024-12-17T21:58:44.7893015Z Version : CUDA 12.2
2024-12-17T21:58:44.7893647Z Name : NVIDIA CUDA BACKEND
2024-12-17T21:58:44.7894417Z Vendor : NVIDIA Corporation
2024-12-17T21:58:44.7894873Z Devices : 1
2024-12-17T21:58:44.7895202Z Device [#0]:
2024-12-17T21:58:44.7895571Z Type : gpu
2024-12-17T21:58:44.7895994Z Version : 8.6
2024-12-17T21:58:44.7896460Z Name : NVIDIA GeForce RTX 3090
2024-12-17T21:58:44.7897034Z Vendor : NVIDIA Corporation
2024-12-17T21:58:44.7897582Z Driver : CUDA 12.2
2024-12-17T21:58:44.7898120Z UUID : 151202218110820252671222472153911713872
2024-12-17T21:58:44.7898682Z DeviceID : 0
2024-12-17T21:58:44.7899113Z Num SubDevices : 0
2024-12-17T21:58:44.7899560Z Num SubSubDevices : 0
2024-12-17T21:58:44.7901299Z Images are not fully supported by the CUDA BE, their support is disabled by default. Their partial support can be activated by setting UR_CUDA_ENABLE_IMAGE_SUPPORT environment variable at runtime.
2024-12-17T21:58:44.7914284Z Aspects : gpu fp16 fp64 online_compiler online_linker queue_profiling usm_device_allocations usm_host_allocations usm_shared_allocations ext_intel_pci_address usm_atomic_shared_allocations atomic64 ext_intel_device_info_uuid ext_oneapi_native_assert ext_oneapi_cuda_async_barrier ext_intel_free_memory ext_intel_device_id ext_intel_memory_clock_rate ext_intel_memory_bus_width ext_oneapi_bindless_images ext_oneapi_bindless_images_shared_usm ext_oneapi_bindless_images_1d_usm ext_oneapi_bindless_images_2d_usm ext_oneapi_external_memory_import ext_oneapi_external_semaphore_import ext_oneapi_mipmap ext_oneapi_mipmap_anisotropy ext_oneapi_mipmap_level_reference ext_oneapi_ballot_group ext_oneapi_fixed_size_group ext_oneapi_opportunistic_group ext_oneapi_graph ext_oneapi_limited_graph ext_oneapi_cubemap ext_oneapi_cubemap_seamless_filtering ext_oneapi_bindless_sampled_image_fetch_1d_usm ext_oneapi_bindless_sampled_image_fetch_2d_usm ext_oneapi_bindless_sampled_image_fetch_2d ext_oneapi_bindless_sampled_image_fetch_3d ext_oneapi_queue_profiling_tag ext_oneapi_virtual_mem ext_oneapi_image_array ext_oneapi_unique_addressing_per_dim ext_oneapi_bindless_images_sample_1d_usm ext_oneapi_bindless_images_sample_2d_usm
2024-12-17T21:58:44.7931906Z info::device::sub_group_sizes: 32
2024-12-17T21:58:44.7932863Z Architecture: nvidia_gpu_sm_86
2024-12-17T21:58:44.7933758Z Platform [#2]:
2024-12-17T21:58:44.7934438Z Version : 0.1
2024-12-17T21:58:44.7935152Z Name : SYCL_NATIVE_CPU
2024-12-17T21:58:44.7935953Z Vendor : tbd
2024-12-17T21:58:44.7936651Z Devices : 1
2024-12-17T21:58:44.7937353Z Device [#0]:
2024-12-17T21:58:44.7938085Z Type : cpu
2024-12-17T21:58:44.7938885Z Version : 0.1
2024-12-17T21:58:44.7939713Z Name : SYCL Native CPU
2024-12-17T21:58:44.7940641Z Vendor : Intel(R) Corporation
2024-12-17T21:58:44.7941541Z Driver : 0.0.0
2024-12-17T21:58:44.7942355Z DeviceID : UNKNOWN
2024-12-17T21:58:44.7943485Z Num SubDevices : 0
2024-12-17T21:58:44.7944275Z Num SubSubDevices : 0
2024-12-17T21:58:44.7947019Z Aspects : cpu fp16 fp64 online_compiler online_linker queue_profiling usm_device_allocations usm_host_allocations usm_shared_allocations usm_system_allocations usm_atomic_host_allocations usm_atomic_shared_allocations atomic64
2024-12-17T21:58:44.7949755Z info::device::sub_group_sizes: 1
2024-12-17T21:58:44.7950612Z Architecture: unknown
2024-12-17T21:58:44.7951795Z default_selector() : gpu, NVIDIA CUDA BACKEND, NVIDIA GeForce RTX 3090 8.6 [CUDA 12.2]
2024-12-17T21:58:44.7953336Z accelerator_selector() : No device of requested type available.
2024-12-17T21:58:44.7954685Z cpu_selector() : cpu, SYCL_NATIVE_CPU, SYCL Native CPU 0.1 [0.0.0]
2024-12-17T21:58:44.7956134Z gpu_selector() : gpu, NVIDIA CUDA BACKEND, NVIDIA GeForce RTX 3090 8.6 [CUDA 12.2]
2024-12-17T21:58:44.7957796Z custom_selector(gpu) : gpu, NVIDIA CUDA BACKEND, NVIDIA GeForce RTX 3090 8.6 [CUDA 12.2]
2024-12-17T21:58:44.7959264Z custom_selector(cpu) : cpu, SYCL_NATIVE_CPU, SYCL Native CPU 0.1 [0.0.0]
2024-12-17T21:58:44.7960574Z custom_selector(acc) : No device of requested type available.
Additional context
No response