[SYCL][CUDA][L0] fatal error: translation unit is too large for Clang to process #20371
Description
Describe the bug
When building a project with just JIT mode results in the following linking error. Stack trace and reproducer below. This was confirmed with L0 & CUDA backend. Expecting similar behavior for HIP backend as well.
We see a similar issue with 2025.2.0 version with agama 1146.12 on PVC (L0) as well (ticket if folks have access to CMPLRLLVM-70962)
Reproducer:
git clone https://github.com/abagusetty/ExchCXX.git ExchCXX_test
cd ExchCXX_test
f=src/sycl/exchcxx_sycl.cmake
# insert link option (device-code-split) after the compile option within the block
sed -i -E '
/^if[[:space:]]*\( *EXCHCXX_SYCL_DEVICE_CODE_SPLIT_PER_SOURCE *\)/,/^endif[[:space:]]*\(\)/ {
/target_compile_options[[:space:]]*\([[:space:]]*exchcxx[[:space:]]+PRIVATE/,/^[[:space:]]*\)[[:space:]]*$/ {
/^[[:space:]]*\)[[:space:]]*$/ a\
target_link_options( exchcxx PRIVATE -fsycl-device-code-split=per_source )
}
}
' "$f"
find . -type f -path "*/src/sycl/*.cmake" -exec sed -i -e 's/per_source/per_kernel/g' {} +
find . -type f -path "*/src/sycl/*.cmake" -exec sed -i -e 's/PER_SOURCE/PER_KERNEL/g' {} +
cmake -H. -Bbuild_sycl_jit_kernel -DEXCHCXX_ENABLE_SYCL=ON -DBUILD_SHARED_LIBS=ON -DCMAKE_BUILD_TYPE=RelWithDebInfo -DEXCHCXX_ENABLE_TESTS=OFF -DEXCHCXX_ENABLE_LIBXC=OFF -DCMAKE_INSTALL_PREFIX=$PWD/install_jit -DCMAKE_CXX_FLAGS="-fno-system-debug" -DCMAKE_C_COMPILER=`which clang` -DCMAKE_CXX_COMPILER=`which clang++`
cd build_sycl_jit_kernel
make VERBOSE=1 -j48Error:
fatal error: translation unit is too large for Clang to process: ran out of source locations
note: 2B (2B) in local locations, 0B (0B) in locations loaded from AST files, for a total of
2B (2B) (0% of available space)
note: file entered 1 time using 1B (1B) of space
clang: /vast/users/abagusetty/compilers/llvm/clang/lib/Frontend/CompilerInstance.cpp:1003: static bool clang::CompilerInstance::InitializeSourceManager(const clang::FrontendInputFile&, clang::DiagnosticsEngine&, clang::FileManager&, clang::SourceManager&): Assertion `SourceMgr.getMainFileID().isValid() && "Couldn't establish MainFileID!"' failed.
PLEASE submit a bug report to https://github.com/intel/llvm/issues and include the crash backtrace, preprocessed source, and associated run script.
Stack dump:
0. Program arguments: /vast/users/abagusetty/gpu4pyscf/install_llvm_cuda12.9.1_09-16-2025/bin/clang --target=x86_64-unknown-linux-gnu -c -o /tmp/libexchcxx-f896a5.o /tmp/wrapper-dc8b3b.bc -fPIC
#0 0x00000000024fce8b llvm::sys::PrintStackTrace(llvm::raw_ostream&, int) (/vast/users/abagusetty/gpu4pyscf/install_llvm_cuda12.9.1_09-16-2025/bin/clang+0x24fce8b)
#1 0x00000000024f9dfb llvm::sys::RunSignalHandlers() (/vast/users/abagusetty/gpu4pyscf/install_llvm_cuda12.9.1_09-16-2025/bin/clang+0x24f9dfb)
#2 0x000000000243d7d8 CrashRecoverySignalHandler(int) CrashRecoveryContext.cpp:0:0
#3 0x00007f35f7057900 __restore_rt (/lib64/libc.so.6+0x57900)
#4 0x00007f35f70a8dfc __pthread_kill_implementation (/lib64/libc.so.6+0xa8dfc)
#5 0x00007f35f7057842 gsignal (/lib64/libc.so.6+0x57842)
#6 0x00007f35f703f5cf abort (/lib64/libc.so.6+0x3f5cf)
#7 0x00007f35f703f4e7 _nl_load_domain.cold (/lib64/libc.so.6+0x3f4e7)
#8 0x00007f35f704fb32 (/lib64/libc.so.6+0x4fb32)
#9 0x00000000031af258 (/vast/users/abagusetty/gpu4pyscf/install_llvm_cuda12.9.1_09-16-2025/bin/clang+0x31af258)
#10 0x0000000003249b57 clang::FrontendAction::BeginSourceFile(clang::CompilerInstance&, clang::FrontendInputFile const&) (/vast/users/abagusetty/gpu4pyscf/install_llvm_cuda12.9.1_09-16-2025/bin/clang+0x3249b57)
#11 0x00000000031ba14a clang::CompilerInstance::ExecuteAction(clang::FrontendAction&) (/vast/users/abagusetty/gpu4pyscf/install_llvm_cuda12.9.1_09-16-2025/bin/clang+0x31ba14a)
#12 0x000000000332d0fb clang::ExecuteCompilerInvocation(clang::CompilerInstance*) (/vast/users/abagusetty/gpu4pyscf/install_llvm_cuda12.9.1_09-16-2025/bin/clang+0x332d0fb)
#13 0x0000000000d2f022 cc1_main(llvm::ArrayRef<char const*>, char const*, void*) (/vast/users/abagusetty/gpu4pyscf/install_llvm_cuda12.9.1_09-16-2025/bin/clang+0xd2f022)
#14 0x0000000000d24ab7 ExecuteCC1Tool(llvm::SmallVectorImpl<char const*>&, llvm::ToolContext const&) driver.cpp:0:0
#15 0x0000000002f4a6a9 void llvm::function_ref<void ()>::callback_fn<clang::driver::CC1Command::Execute(llvm::ArrayRef<std::optional<llvm::StringRef>>, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char>>*, bool*) const::'lambda'()>(long) Job.cpp:0:0
#16 0x000000000243dc32 llvm::CrashRecoveryContext::RunSafely(llvm::function_ref<void ()>) (/vast/users/abagusetty/gpu4pyscf/install_llvm_cuda12.9.1_09-16-2025/bin/clang+0x243dc32)
#17 0x0000000002f4c9f8 clang::driver::CC1Command::Execute(llvm::ArrayRef<std::optional<llvm::StringRef>>, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char>>*, bool*) const (.part.0) Job.cpp:0:0
#18 0x0000000002eec3da clang::driver::Compilation::ExecuteCommand(clang::driver::Command const&, clang::driver::Command const*&, bool) const (/vast/users/abagusetty/gpu4pyscf/install_llvm_cuda12.9.1_09-16-2025/bin/clang+0x2eec3da)
#19 0x0000000002eed5a6 clang::driver::Compilation::ExecuteJobs(clang::driver::JobList const&, llvm::SmallVectorImpl<std::pair<int, clang::driver::Command const*>>&, bool) const (/vast/users/abagusetty/gpu4pyscf/install_llvm_cuda12.9.1_09-16-2025/bin/clang+0x2eed5a6)
#20 0x0000000002efd414 clang::driver::Driver::ExecuteCompilation(clang::driver::Compilation&, llvm::SmallVectorImpl<std::pair<int, clang::driver::Command const*>>&) (/vast/users/abagusetty/gpu4pyscf/install_llvm_cuda12.9.1_09-16-2025/bin/clang+0x2efd414)
#21 0x0000000000d2a27b clang_main(int, char**, llvm::ToolContext const&) (/vast/users/abagusetty/gpu4pyscf/install_llvm_cuda12.9.1_09-16-2025/bin/clang+0xd2a27b)
#22 0x0000000000c14863 main (/vast/users/abagusetty/gpu4pyscf/install_llvm_cuda12.9.1_09-16-2025/bin/clang+0xc14863)
#23 0x00007f35f7040e6c __libc_start_call_main (/lib64/libc.so.6+0x40e6c)
#24 0x00007f35f7040f35 __libc_start_main@GLIBC_2.2.5 (/lib64/libc.so.6+0x40f35)
#25 0x0000000000d24041 _start /home/abuild/rpmbuild/BUILD/glibc-2.38/csu/../sysdeps/x86_64/start.S:117:0
clang: error: clang frontend command failed with exit code 134 (use -v to see invocation)
clang version 21.0.0git (https://github.com/intel/llvm.git 68f3fdf41ca373e413c74da2949d807d3d7d777f)
Target: x86_64-unknown-linux-gnu
Thread model: posix
InstalledDir: /vast/users/abagusetty/gpu4pyscf/install_llvm_cuda12.9.1_09-16-2025/bin
Build config: +assertions
clang: note: diagnostic msg: Error generating preprocessed source(s) - no preprocessable inputs.
clang++: error: clang-offload-wrapper command failed with exit code 1 (use -v to see invocation)To reproduce
No response
Environment
- OS: Linux
- Target device and vendor: Intel PVC, Nvidia H100
- DPC++ version:
$ clang++ -v
clang version 21.0.0git (https://github.com/intel/llvm.git 68f3fdf41ca373e413c74da2949d807d3d7d777f)
Target: x86_64-unknown-linux-gnu
Thread model: posix
InstalledDir: /vast/users/abagusetty/gpu4pyscf/install_llvm_cuda12.9.1_09-16-2025/bin
Build config: +assertions
Found candidate GCC installation: /usr/lib64/gcc/x86_64-suse-linux/14
Found candidate GCC installation: /usr/lib64/gcc/x86_64-suse-linux/7
Selected GCC installation: /usr/lib64/gcc/x86_64-suse-linux/14
Candidate multilib: .;@m64
Selected multilib: .;@m64
Found CUDA installation: /soft/compilers/cuda/cuda-12.9.1, version
Additional context
No response