| Issue |
91709
|
| Summary |
icpx: error: ptxas command failed
|
| Labels |
new issue
|
| Assignees |
|
| Reporter |
J-shel
|
Updating this issue:
https://github.com/llvm/llvm-project/issues/90539
Hello, I'm sorry I may misunderstood relationship between icpx and clang++, but when I use icpx with DPC++ plugin to compile cloudsc-sycl, I did met some error messages seems related to LLVM/clang:
/opt/intel/oneapi/compiler/2024.1/bin/compiler/clang-offload-bundler -type=o -input=CMakeFiles/dwarf-cloudsc-scc-sycl-lib.dir/cloudsc/mycpu.cpp.o -list
/opt/intel/oneapi/compiler/2024.1/bin/compiler/clang-offload-bundler -type=ao -targets=sycl-fpga_aocx-intel-unknown -input=/usr/local/cuda/lib64/libcudadevrt.a -check-section -base-temp-dir=/tmp/icpx-e2ed2974b6
/opt/intel/oneapi/compiler/2024.1/bin/compiler/clang-offload-bundler -type=ao -targets=sycl-fpga_aocr-intel-unknown -input=/usr/local/cuda/lib64/libcudadevrt.a -check-section -base-temp-dir=/tmp/icpx-e2ed2974b6
/opt/intel/oneapi/compiler/2024.1/bin/compiler/clang-offload-bundler -type=ao -targets=sycl-fpga_aocr_emu-intel-unknown -input=/usr/local/cuda/lib64/libcudadevrt.a -check-section -base-temp-dir=/tmp/icpx-e2ed2974b6
/opt/intel/oneapi/compiler/2024.1/bin/compiler/clang-offload-bundler -type=ao -input=/usr/local/cuda/lib64/libcudadevrt.a -list
"/opt/intel/oneapi/compiler/2024.1/bin/compiler/llvm-foreach" --out-ext=o --in-file-list=/tmp/icpx-c2c23b019a/load_state-sm_80-c9c042.s --in-replace=/tmp/icpx-c2c23b019a/load_state-sm_80-c9c042.s --out-file-list=/tmp/icpx-c2c23b019a/load_state-sm_80-146846.cubin --out-replace=/tmp/icpx-c2c23b019a/load_state-sm_80-146846.cubin -- /usr/local/cuda-12.0/bin/ptxas -m64 -O3 -v --gpu-name sm_80 --output-file /tmp/icpx-c2c23b019a/load_state-sm_80-146846.cubin /tmp/icpx-c2c23b019a/load_state-sm_80-c9c042.s
ptxas info : 0 bytes gmem
ptxas info : 0 bytes gmem
ptxas fatal : Unresolved extern function 'fmin'
llvm-foreach:
icpx: error: ptxas command failed with exit code 255 (use -v to see invocation)
Intel(R) oneAPI DPC++/C++ Compiler 2024.1.0 (2024.1.0.20240308)
Target: x86_64-unknown-linux-gnu
Thread model: posix
I guess it may be related to CUDA. When using CUDA nvcc, adding -rdc=true as an compiler option may solve the "Unresolved extern function 'xxxxx' " issue. As for icpx, I found the compiler option "-fgpu-rdc", but it didn't solve my problem.
![]()
_______________________________________________
llvm-bugs mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
