intel-llvm-sycl/dpcpp compiler linking error: soft-float modules with double-float modules #17726
Comments
|
I've tried the advice outlined in this thread https://community.intel.com/t5/Intel-Fortran-Compiler/Double-precison-on-ARC-GPU/td-p/1450217with the understanding that those environment flags are for the fortran compiler. The following compilation command:
yields
are there comparable environment variables for the Intel sycl compiler? |
|
Wanted to note this reference link: https://github.com/intel/compute-runtime/blob/master/opencl/doc/FAQ.md#feature-double-precision-emulation-fp64 |
|
Hi @ct-clmsn, thanks for the report. I don't think that this is anyhow related to
Whilst we are trying to reproduce this on our end, could you please repeat the failing command with |
|
Could you please also be a little bit more specific about your environment: what specific OS and CPU you are running on? |
|
I'm currently running on RISC-V with Ubuntu 24.04.1 LTS; I've been working on porting the environment. |
|
compilation command with
output: clang version 19.0.0git |
Describe the bug
I've compiled the intel-llvm-sycl/dpcpp compiler from GitHub here specifically to use the compiler with an Intel GPU: https://github.com/intel/llvm
Right now, I'd like to compile a simple "hello world" program that can print out an the Intel Arc A750 GPU that I've installed on the machine.
I've compiled this program successfully using the Adaptive C++ compiler from here: https://adaptivecpp.github.io/
When using the Intel-LLVM-Sycl/dpcpp compiler with the following flags:
clang++ -DONEDPL_USE_DPCPP_BACKEND=1 -DONEDPL_USE_OPENMP_BACKEND=0 -DONEDPL_USE_TBB_BACKEND=0 -DPSTL_USE_PARALLEL_POLICIES=0 -D_GLIBCXX_USE_TBB_PAR_BACKEND=0 -O3 -DNDEBUG -std=gnu++20 -fsycl -fsycl-unnamed-lambda -fsycl-targets=spir64 -Xsycl-target-backend "-device intel_gpu_dg2_g12" hellosycl.cpp -o hello_dg12clang++ -DONEDPL_USE_DPCPP_BACKEND=1 -DONEDPL_USE_OPENMP_BACKEND=0 -DONEDPL_USE_TBB_BACKEND=0 -DPSTL_USE_PARALLEL_POLICIES=0 -D_GLIBCXX_USE_TBB_PAR_BACKEND=0 -O3 -DNDEBUG -std=gnu++20 -fsycl -fsycl-unnamed-lambda -fsycl-targets=spir64 hellosycl.cpp -o hello_dg12clang++ -DONEDPL_USE_DPCPP_BACKEND=1 -DONEDPL_USE_OPENMP_BACKEND=0 -DONEDPL_USE_TBB_BACKEND=0 -DPSTL_USE_PARALLEL_POLICIES=0 -D_GLIBCXX_USE_TBB_PAR_BACKEND=0 -O3 -DNDEBUG -std=gnu++20 -fsycl -fsycl-unnamed-lambda -fsycl-targets=spir64 hellosycl.cpp -o hello_dg12clang++ -DONEDPL_USE_DPCPP_BACKEND=1 -DONEDPL_USE_OPENMP_BACKEND=0 -DONEDPL_USE_TBB_BACKEND=0 -DPSTL_USE_PARALLEL_POLICIES=0 -D_GLIBCXX_USE_TBB_PAR_BACKEND=0 -O3 -DNDEBUG -std=gnu++20 -fsycl -fsycl-targets=spir64 hellosycl.cpp -o hello_dg12The compilation process terminates with this linking error :
/usr/bin/ld: /tmp/hello_dg12-wrapper-f79096.o: can't link soft-float modules with double-float modulesAny assistance or help would be much appreciated! I get the impression there's a flag that's missing which tells the compiler to emulate 64 bit floating point (double precision) on the GPU architecture.
I also wanted to note that the
-Xsycl-targets=spir64_genflag causes the compiler to fail and terminate with:clang++: error: gen compiler command failed with exit code 1 (use -v to see invocation)To reproduce
#include <sycl/sycl.hpp> int main(int argc, char* argv[]) { sycl::queue q; std::cout << "Running on " << q.get_device().get_info<sycl::info::device::name>() << "\n"; }clang++ -DONEDPL_USE_DPCPP_BACKEND=1 -DONEDPL_USE_OPENMP_BACKEND=0 -DONEDPL_USE_TBB_BACKEND=0 -DPSTL_USE_PARALLEL_POLICIES=0 -D_GLIBCXX_USE_TBB_PAR_BACKEND=0 -O3 -DNDEBUG -std=gnu++20 -fsycl -fsycl-unnamed-lambda -fsycl-targets=spir64_gen hellosycl.cpp -o hello_dg12Environment
Additional context
No response
The text was updated successfully, but these errors were encountered: