error when trying to lower .cu file, should I report this here or on LLVM repo issues?

I read this: https://github.com/llvm/Polygeist/issues/320 and tried to construct a command to lower my .cu file into LLVM IR.

The result:

nyck33@lenovo-gtx1650:/mnt/d/LLVM/Lean/LLVM$ cgeist cudaComputePortfolioRisk.cu --function=* -S --resource-dir=$LLVM_BUILD_DIR/lib/clang/18 --cuda-gpu-arch=sm_75 | polygeist-opt --convert-polygeist-to-llvm
warning: CUDA version 12.1 is only partially supported
warning: CUDA version 12.1 is only partially supported
polygeist-opt: /mnt/d/LLVM/NewPolygeistDir/llvm-project/llvm/include/llvm/Support/Casting.h:566: decltype(auto) llvm::cast(const From&) [with To = mlir::detail::TypedValue<mlir::MemRefType>; From = mlir::Value]: Assertion `isa<To>(Val) && "cast<Ty>() argument of incompatible type!"' failed.
PLEASE submit a bug report to https://github.com/llvm/llvm-project/issues/ and include the crash backtrace.
Stack dump:
0.      Program arguments: polygeist-opt --convert-polygeist-to-llvm
 #0 0x0000557978016ad8 llvm::sys::PrintStackTrace(llvm::raw_ostream&, int) /mnt/d/LLVM/NewPolygeistDir/llvm-project/llvm/lib/Support/Unix/Signals.inc:723:22
 #1 0x0000557978016ef4 PrintStackTraceSignalHandler(void*) /mnt/d/LLVM/NewPolygeistDir/llvm-project/llvm/lib/Support/Unix/Signals.inc:798:1
 #2 0x0000557978014341 llvm::sys::RunSignalHandlers() /mnt/d/LLVM/NewPolygeistDir/llvm-project/llvm/lib/Support/Signals.cpp:105:20
 #3 0x0000557978016370 SignalHandler(int) /mnt/d/LLVM/NewPolygeistDir/llvm-project/llvm/lib/Support/Unix/Signals.inc:413:1
 #4 0x00007f396dd06520 (/usr/lib/x86_64-linux-gnu/libc.so.6+0x42520)
 #5 0x00007f396dd5a9fc pthread_kill (/usr/lib/x86_64-linux-gnu/libc.so.6+0x969fc)
 #6 0x00007f396dd06476 gsignal (/usr/lib/x86_64-linux-gnu/libc.so.6+0x42476)
 #7 0x00007f396dcec7f3 abort (/usr/lib/x86_64-linux-gnu/libc.so.6+0x287f3)
 #8 0x00007f396dcec71b (/usr/lib/x86_64-linux-gnu/libc.so.6+0x2871b)
 #9 0x00007f396dcfde96 (/usr/lib/x86_64-linux-gnu/libc.so.6+0x39e96)
#10 0x0000557972f0f4fc decltype(auto) llvm::cast<mlir::detail::TypedValue<mlir::MemRefType>, mlir::Value>(mlir::Value const&) /mnt/d/LLVM/NewPolygeistDir/llvm-project/llvm/include/llvm/Support/Casting.h:567:42
#11 0x0000557972ed23fa mlir::affine::AffineLoadOp::getMemref() /mnt/d/LLVM/NewPolygeistDir/llvm-project/build/tools/mlir/include/mlir/Dialect/Affine/IR/AffineOps.cpp.inc:1087:1
#12 0x0000557972ec3622 mlir::affine::AffineLoadOp::fold(mlir::affine::AffineLoadOpGenericAdaptor<llvm::ArrayRef<mlir::Attribute>>) /mnt/d/LLVM/NewPolygeistDir/llvm-project/mlir/lib/Dialect/Affine/IR/AffineOps.cpp:3178:31
#13 0x0000557972f7db4c mlir::LogicalResult mlir::Op<mlir::affine::AffineLoadOp, mlir::OpTrait::ZeroRegions, mlir::OpTrait::OneResult, mlir::OpTrait::OneTypedResult<mlir::Type>::Impl, mlir::OpTrait::ZeroSuccessors, mlir::OpTrait::AtLeastNOperands<1u>::Impl, mlir::OpTrait::OpInvariants, mlir::BytecodeOpInterface::Trait, mlir::affine::AffineReadOpInterface::Trait, mlir::affine::AffineMapAccessInterface::Trait, mlir::OpTrait::MemRefsNormalizable, mlir::MemoryEffectOpInterface::Trait>::foldSingleResultHook<mlir::affine::AffineLoadOp>(mlir::Operation*, llvm::ArrayRef<mlir::Attribute>, llvm::SmallVectorImpl<mlir::OpFoldResult>&) /mnt/d/LLVM/NewPolygeistDir/llvm-project/mlir/include/mlir/IR/OpDefinition.h:1898:42
#14 0x0000557972f79e3a mlir::Op<mlir::affine::AffineLoadOp, mlir::OpTrait::ZeroRegions, mlir::OpTrait::OneResult, mlir::OpTrait::OneTypedResult<mlir::Type>::Impl, mlir::OpTrait::ZeroSuccessors, mlir::OpTrait::AtLeastNOperands<1u>::Impl, mlir::OpTrait::OpInvariants, mlir::BytecodeOpInterface::Trait, mlir::affine::AffineReadOpInterface::Trait, mlir::affine::AffineMapAccessInterface::Trait, mlir::OpTrait::MemRefsNormalizable, mlir::MemoryEffectOpInterface::Trait>::getFoldHookFn()::'lambda'(mlir::Operation*, llvm::ArrayRef<mlir::Attribute>, llvm::SmallVectorImpl<mlir::OpFoldResult>&)::operator()(mlir::Operation*, llvm::ArrayRef<mlir::Attribute>, llvm::SmallVectorImpl<mlir::OpFoldResult>&) const /mnt/d/LLVM/NewPolygeistDir/llvm-project/mlir/include/mlir/IR/OpDefinition.h:1872:14
#15 0x0000557972f88b3a mlir::LogicalResult llvm::detail::UniqueFunctionBase<mlir::LogicalResult, mlir::Operation*, llvm::ArrayRef<mlir::Attribute>, llvm::SmallVectorImpl<mlir::OpFoldResult>&>::CallImpl<mlir::Op<mlir::affine::AffineLoadOp, mlir::OpTrait::ZeroRegions, mlir::OpTrait::OneResult, mlir::OpTrait::OneTypedResult<mlir::Type>::Impl, mlir::OpTrait::ZeroSuccessors, mlir::OpTrait::AtLeastNOperands<1u>::Impl, mlir::OpTrait::OpInvariants, mlir::BytecodeOpInterface::Trait, mlir::affine::AffineReadOpInterface::Trait, mlir::affine::AffineMapAccessInterface::Trait, mlir::OpTrait::MemRefsNormalizable, mlir::MemoryEffectOpInterface::Trait>::getFoldHookFn()::'lambda'(mlir::Operation*, llvm::ArrayRef<mlir::Attribute>, llvm::SmallVectorImpl<mlir::OpFoldResult>&) const>(void*, mlir::Operation*, llvm::ArrayRef<mlir::Attribute>, llvm::SmallVectorImpl<mlir::OpFoldResult>&) /mnt/d/LLVM/NewPolygeistDir/llvm-project/llvm/include/llvm/ADT/FunctionExtras.h:221:3
#16 0x0000557972f78c0e llvm::unique_function<mlir::LogicalResult (mlir::Operation*, llvm::ArrayRef<mlir::Attribute>, llvm::SmallVectorImpl<mlir::OpFoldResult>&) const>::operator()(mlir::Operation*, llvm::ArrayRef<mlir::Attribute>, llvm::SmallVectorImpl<mlir::OpFoldResult>&) const /mnt/d/LLVM/NewPolygeistDir/llvm-project/llvm/include/llvm/ADT/FunctionExtras.h:409:3
#17 0x0000557972f73292 mlir::RegisteredOperationName::Model<mlir::affine::AffineLoadOp>::foldHook(mlir::Operation*, llvm::ArrayRef<mlir::Attribute>, llvm::SmallVectorImpl<mlir::OpFoldResult>&) /mnt/d/LLVM/NewPolygeistDir/llvm-project/mlir/include/mlir/IR/OperationSupport.h:539:41
#18 0x0000557977e2c40b mlir::OperationName::foldHook(mlir::Operation*, llvm::ArrayRef<mlir::Attribute>, llvm::SmallVectorImpl<mlir::OpFoldResult>&) const /mnt/d/LLVM/NewPolygeistDir/llvm-project/mlir/include/mlir/IR/OperationSupport.h:267:3
#19 0x0000557977e238c9 mlir::Operation::fold(llvm::ArrayRef<mlir::Attribute>, llvm::SmallVectorImpl<mlir::OpFoldResult>&) /mnt/d/LLVM/NewPolygeistDir/llvm-project/mlir/lib/IR/Operation.cpp:614:16
#20 0x0000557977e23a65 mlir::Operation::fold(llvm::SmallVectorImpl<mlir::OpFoldResult>&) /mnt/d/LLVM/NewPolygeistDir/llvm-project/mlir/lib/IR/Operation.cpp:635:14
#21 0x0000557977c99e4d mlir::OpBuilder::tryFold(mlir::Operation*, llvm::SmallVectorImpl<mlir::Value>&) /mnt/d/LLVM/NewPolygeistDir/llvm-project/mlir/lib/IR/Builders.cpp:480:13
#22 0x00005579777c6c75 (anonymous namespace)::OperationLegalizer::legalizeWithFold(mlir::Operation*, mlir::ConversionPatternRewriter&) /mnt/d/LLVM/NewPolygeistDir/llvm-project/mlir/lib/Transforms/Utils/DialectConversion.cpp:1896:13   
#23 0x00005579777c693f (anonymous namespace)::OperationLegalizer::legalize(mlir::Operation*, mlir::ConversionPatternRewriter&) /mnt/d/LLVM/NewPolygeistDir/llvm-project/mlir/lib/Transforms/Utils/DialectConversion.cpp:1858:16
#24 0x00005579777c7f2f (anonymous namespace)::OperationLegalizer::legalizePatternCreatedOperations(mlir::ConversionPatternRewriter&, mlir::detail::ConversionPatternRewriterImpl&, (anonymous namespace)::RewriterState&, (anonymous namespace)::RewriterState&) /mnt/d/LLVM/NewPolygeistDir/llvm-project/mlir/lib/Transforms/Utils/DialectConversion.cpp:2083:15
#25 0x00005579777c79bb (anonymous namespace)::OperationLegalizer::legalizePatternResult(mlir::Operation*, mlir::Pattern const&, mlir::ConversionPatternRewriter&, (anonymous namespace)::RewriterState&) /mnt/d/LLVM/NewPolygeistDir/llvm-project/mlir/lib/Transforms/Utils/DialectConversion.cpp:2016:13
#26 0x00005579777c71fb (anonymous namespace)::OperationLegalizer::legalizeWithPattern(mlir::Operation*, mlir::ConversionPatternRewriter&)::'lambda1'(mlir::Pattern const&)::operator()(mlir::Pattern const&) const /mnt/d/LLVM/NewPolygeistDir/llvm-project/mlir/lib/Transforms/Utils/DialectConversion.cpp:1951:40
#27 0x00005579777d3eba mlir::LogicalResult llvm::function_ref<mlir::LogicalResult (mlir::Pattern const&)>::callback_fn<(anonymous namespace)::OperationLegalizer::legalizeWithPattern(mlir::Operation*, mlir::ConversionPatternRewriter&)::'lambda1'(mlir::Pattern const&)>(long, mlir::Pattern const&) /mnt/d/LLVM/NewPolygeistDir/llvm-project/llvm/include/llvm/ADT/STLFunctionalExtras.h:47:3
#28 0x00005579779fcb07 llvm::function_ref<mlir::LogicalResult (mlir::Pattern const&)>::operator()(mlir::Pattern const&) const /mnt/d/LLVM/NewPolygeistDir/llvm-project/llvm/include/llvm/ADT/STLFunctionalExtras.h:69:3
#29 0x00005579779f8ff6 mlir::PatternApplicator::matchAndRewrite(mlir::Operation*, mlir::PatternRewriter&, llvm::function_ref<bool (mlir::Pattern const&)>, llvm::function_ref<void (mlir::Pattern const&)>, llvm::function_ref<mlir::LogicalResult (mlir::Pattern const&)>)::'lambda'()::operator()() const /mnt/d/LLVM/NewPolygeistDir/llvm-project/mlir/lib/Rewrite/PatternApplicator.cpp:214:55
#30 0x00005579779f98ad void llvm::function_ref<void ()>::callback_fn<mlir::PatternApplicator::matchAndRewrite(mlir::Operation*, mlir::PatternRewriter&, llvm::function_ref<bool (mlir::Pattern const&)>, llvm::function_ref<void (mlir::Pattern const&)>, llvm::function_ref<mlir::LogicalResult (mlir::Pattern const&)>)::'lambda'()>(long) /mnt/d/LLVM/NewPolygeistDir/llvm-project/llvm/include/llvm/ADT/STLFunctionalExtras.h:46:40
#31 0x0000557976d73e54 llvm::function_ref<void ()>::operator()() const /mnt/d/LLVM/NewPolygeistDir/llvm-project/llvm/include/llvm/ADT/STLFunctionalExtras.h:68:62
#32 0x00005579779fcc55 void mlir::MLIRContext::executeAction<mlir::ApplyPatternAction, mlir::Pattern const&>(llvm::function_ref<void ()>, llvm::ArrayRef<mlir::IRUnit>, mlir::Pattern const&) /mnt/d/LLVM/NewPolygeistDir/llvm-project/mlir/include/mlir/IR/MLIRContext.h:276:3
#33 0x00005579779f965a mlir::PatternApplicator::matchAndRewrite(mlir::Operation*, mlir::PatternRewriter&, llvm::function_ref<bool (mlir::Pattern const&)>, llvm::function_ref<void (mlir::Pattern const&)>, llvm::function_ref<mlir::LogicalResult (mlir::Pattern const&)>) /mnt/d/LLVM/NewPolygeistDir/llvm-project/mlir/lib/Rewrite/PatternApplicator.cpp:227:5
#34 0x00005579777c7412 (anonymous namespace)::OperationLegalizer::legalizeWithPattern(mlir::Operation*, mlir::ConversionPatternRewriter&) /mnt/d/LLVM/NewPolygeistDir/llvm-project/mlir/lib/Transforms/Utils/DialectConversion.cpp:1959:36
#35 0x00005579777c69fd (anonymous namespace)::OperationLegalizer::legalize(mlir::Operation*, mlir::ConversionPatternRewriter&) /mnt/d/LLVM/NewPolygeistDir/llvm-project/mlir/lib/Transforms/Utils/DialectConversion.cpp:1867:16
#36 0x00005579777c913a (anonymous namespace)::OperationConverter::convert(mlir::ConversionPatternRewriter&, mlir::Operation*) /mnt/d/LLVM/NewPolygeistDir/llvm-project/mlir/lib/Transforms/Utils/DialectConversion.cpp:2378:13
#37 0x00005579777c9633 (anonymous namespace)::OperationConverter::convertOperations(llvm::ArrayRef<mlir::Operation*>, llvm::function_ref<void (mlir::Diagnostic&)>) /mnt/d/LLVM/NewPolygeistDir/llvm-project/mlir/lib/Transforms/Utils/DialectConversion.cpp:2432:15
#38 0x00005579777cf06a mlir::applyPartialConversion(llvm::ArrayRef<mlir::Operation*>, mlir::ConversionTarget const&, mlir::FrozenRewritePatternSet const&, llvm::DenseSet<mlir::Operation*, llvm::DenseMapInfo<mlir::Operation*, void>>*) /mnt/d/LLVM/NewPolygeistDir/llvm-project/mlir/lib/Transforms/Utils/DialectConversion.cpp:3400:39
#39 0x00005579777cf0f8 mlir::applyPartialConversion(mlir::Operation*, mlir::ConversionTarget const&, mlir::FrozenRewritePatternSet const&, llvm::DenseSet<mlir::Operation*, llvm::DenseMapInfo<mlir::Operation*, void>>*) /mnt/d/LLVM/NewPolygeistDir/llvm-project/mlir/lib/Transforms/Utils/DialectConversion.cpp:3408:1
#40 0x00005579747bfa94 (anonymous namespace)::ConvertPolygeistToLLVMPass::convertModule(mlir::ModuleOp, bool) /mnt/d/LLVM/NewPolygeistDir/lib/polygeist/Passes/ConvertPolygeistToLLVM.cpp:2972:17
#41 0x00005579747c03a3 (anonymous namespace)::ConvertPolygeistToLLVMPass::runOnOperation() /mnt/d/LLVM/NewPolygeistDir/lib/polygeist/Passes/ConvertPolygeistToLLVM.cpp:3025:3
#42 0x0000557977adffcd mlir::detail::OpToOpPassAdaptor::run(mlir::Pass*, mlir::Operation*, mlir::AnalysisManager, bool, unsigned int)::'lambda0'()::operator()() const /mnt/d/LLVM/NewPolygeistDir/llvm-project/mlir/lib/Pass/Pass.cpp:500:57
#43 0x0000557977ae3caa void llvm::function_ref<void ()>::callback_fn<mlir::detail::OpToOpPassAdaptor::run(mlir::Pass*, mlir::Operation*, mlir::AnalysisManager, bool, unsigned int)::'lambda0'()>(long) /mnt/d/LLVM/NewPolygeistDir/llvm-project/llvm/include/llvm/ADT/STLFunctionalExtras.h:46:40
#44 0x0000557976d73e54 llvm::function_ref<void ()>::operator()() const /mnt/d/LLVM/NewPolygeistDir/llvm-project/llvm/include/llvm/ADT/STLFunctionalExtras.h:68:62
#45 0x0000557977ae9a2b void mlir::MLIRContext::executeAction<mlir::PassExecutionAction, mlir::Pass&>(llvm::function_ref<void ()>, llvm::ArrayRef<mlir::IRUnit>, mlir::Pass&) /mnt/d/LLVM/NewPolygeistDir/llvm-project/mlir/include/mlir/IR/MLIRContext.h:276:3
#46 0x0000557977ae03ef mlir::detail::OpToOpPassAdaptor::run(mlir::Pass*, mlir::Operation*, mlir::AnalysisManager, bool, unsigned int) /mnt/d/LLVM/NewPolygeistDir/llvm-project/mlir/lib/Pass/Pass.cpp:509:23
#47 0x0000557977ae06ca mlir::detail::OpToOpPassAdaptor::runPipeline(mlir::OpPassManager&, mlir::Operation*, mlir::AnalysisManager, bool, unsigned int, mlir::PassInstrumentor*, mlir::PassInstrumentation::PipelineParentInfo const*) /mnt/d/LLVM/NewPolygeistDir/llvm-project/mlir/lib/Pass/Pass.cpp:569:15
#48 0x0000557977ae25f0 mlir::PassManager::runPasses(mlir::Operation*, mlir::AnalysisManager) /mnt/d/LLVM/NewPolygeistDir/llvm-project/mlir/lib/Pass/Pass.cpp:880:40
#49 0x0000557977ae2448 mlir::PassManager::run(mlir::Operation*) /mnt/d/LLVM/NewPolygeistDir/llvm-project/mlir/lib/Pass/Pass.cpp:860:69
#50 0x0000557974c65003 performActions(llvm::raw_ostream&, std::shared_ptr<llvm::SourceMgr> const&, mlir::MLIRContext*, mlir::MlirOptMainConfig const&) /mnt/d/LLVM/NewPolygeistDir/llvm-project/mlir/lib/Tools/mlir-opt/MlirOptMain.cpp:376:13
#51 0x0000557974c655d7 processBuffer(llvm::raw_ostream&, std::unique_ptr<llvm::MemoryBuffer, std::default_delete<llvm::MemoryBuffer>>, mlir::MlirOptMainConfig const&, mlir::DialectRegistry&, llvm::ThreadPool*) /mnt/d/LLVM/NewPolygeistDir/llvm-project/mlir/lib/Tools/mlir-opt/MlirOptMain.cpp:431:26
#52 0x0000557974c65711 mlir::MlirOptMain(llvm::raw_ostream&, std::unique_ptr<llvm::MemoryBuffer, std::default_delete<llvm::MemoryBuffer>>, mlir::DialectRegistry&, mlir::MlirOptMainConfig const&)::'lambda'(std::unique_ptr<llvm::MemoryBuffer, std::default_delete<llvm::MemoryBuffer>>, llvm::raw_ostream&)::operator()(std::unique_ptr<llvm::MemoryBuffer, std::default_delete<llvm::MemoryBuffer>>, llvm::raw_ostream&) const /mnt/d/LLVM/NewPolygeistDir/llvm-project/mlir/lib/Tools/mlir-opt/MlirOptMain.cpp:472:25
#53 0x0000557974c66850 mlir::LogicalResult llvm::function_ref<mlir::LogicalResult (std::unique_ptr<llvm::MemoryBuffer, std::default_delete<llvm::MemoryBuffer>>, llvm::raw_ostream&)>::callback_fn<mlir::MlirOptMain(llvm::raw_ostream&, std::unique_ptr<llvm::MemoryBuffer, std::default_delete<llvm::MemoryBuffer>>, mlir::DialectRegistry&, mlir::MlirOptMainConfig const&)::'lambda'(std::unique_ptr<llvm::MemoryBuffer, std::default_delete<llvm::MemoryBuffer>>, llvm::raw_ostream&)>(long, std::unique_ptr<llvm::MemoryBuffer, std::default_delete<llvm::MemoryBuffer>>, llvm::raw_ostream&) /mnt/d/LLVM/NewPolygeistDir/llvm-project/llvm/include/llvm/ADT/STLFunctionalExtras.h:45:52
#54 0x0000557977e8dd93 llvm::function_ref<mlir::LogicalResult (std::unique_ptr<llvm::MemoryBuffer, std::default_delete<llvm::MemoryBuffer>>, llvm::raw_ostream&)>::operator()(std::unique_ptr<llvm::MemoryBuffer, std::default_delete<llvm::MemoryBuffer>>, llvm::raw_ostream&) const /mnt/d/LLVM/NewPolygeistDir/llvm-project/llvm/include/llvm/ADT/STLFunctionalExtras.h:68:12
#55 0x0000557977e8d65d mlir::splitAndProcessBuffer(std::unique_ptr<llvm::MemoryBuffer, std::default_delete<llvm::MemoryBuffer>>, llvm::function_ref<mlir::LogicalResult (std::unique_ptr<llvm::MemoryBuffer, std::default_delete<llvm::MemoryBuffer>>, llvm::raw_ostream&)>, llvm::raw_ostream&, bool, bool) /mnt/d/LLVM/NewPolygeistDir/llvm-project/mlir/lib/Support/ToolUtilities.cpp:28:30
#56 0x0000557974c658c6 mlir::MlirOptMain(llvm::raw_ostream&, std::unique_ptr<llvm::MemoryBuffer, std::default_delete<llvm::MemoryBuffer>>, mlir::DialectRegistry&, mlir::MlirOptMainConfig const&) /mnt/d/LLVM/NewPolygeistDir/llvm-project/mlir/lib/Tools/mlir-opt/MlirOptMain.cpp:475:31
#57 0x0000557974c65e83 mlir::MlirOptMain(int, char**, llvm::StringRef, mlir::DialectRegistry&) /mnt/d/LLVM/NewPolygeistDir/llvm-project/mlir/lib/Tools/mlir-opt/MlirOptMain.cpp:531:13
#58 0x0000557972e8e6e5 main /mnt/d/LLVM/NewPolygeistDir/tools/polygeist-opt/polygeist-opt.cpp:110:22
#59 0x00007f396dcedd90 (/usr/lib/x86_64-linux-gnu/libc.so.6+0x29d90)
#60 0x00007f396dcede40 __libc_start_main (/usr/lib/x86_64-linux-gnu/libc.so.6+0x29e40)
#61 0x0000557972e8e0e5 _start (/mnt/d/LLVM/NewPolygeistDir/build/bin/polygeist-opt+0x41610e5)
Aborted

My .cu file:

#include <cuda.h>

#include <cuda_runtime.h>

// CUDA kernel to perform the matrix multiplication A * B
__global__ void matrixMultiply(double* A, double* B, double* C, int ARows, int ACols, int BCols) {
    int row = blockIdx.y * blockDim.y + threadIdx.y;
    int col = blockIdx.x * blockDim.x + threadIdx.x;

    if (row < ARows && col < BCols) {
        double sum = 0.0;
        for (int k = 0; k < ACols; ++k) {
            sum += A[row * ACols + k] * B[k * BCols + col];
        }
        C[row * BCols + col] = sum;
    }
}

// Host function to initialize memory, call the kernels, and clean up
extern "C" void computeCovarianceMatrix(double* S, double* R, double* Sigma, int sRows, int sCols) {
    double *d_S, *d_R, *d_T, *d_Sigma;

    // Allocate Unified Memory – accessible from CPU or GPU
    cudaMallocManaged(&d_S, sRows * sCols * sizeof(double));
    cudaMallocManaged(&d_R, sCols * sCols * sizeof(double));
    cudaMallocManaged(&d_T, sRows * sCols * sizeof(double)); // Intermediate result
    cudaMallocManaged(&d_Sigma, sRows * sRows * sizeof(double));

    // Copy data into managed memory
    cudaMemcpy(d_S, S, sRows * sCols * sizeof(double), cudaMemcpyHostToDevice);
    cudaMemcpy(d_R, R, sCols * sCols * sizeof(double), cudaMemcpyHostToDevice);

    dim3 threadsPerBlock(16, 16);
    dim3 blocksPerGrid1((sCols + 15) / 16, (sRows + 15) / 16);
    dim3 blocksPerGrid2((sRows + 15) / 16, (sRows + 15) / 16);

    // Perform S * R = T
    matrixMultiply<<<blocksPerGrid1, threadsPerBlock>>>(d_S, d_R, d_T, sRows, sCols, sCols);

    // Perform T * S^T = Sigma (assuming S is square and sCols == sRows)
    matrixMultiply<<<blocksPerGrid2, threadsPerBlock>>>(d_T, d_S, d_Sigma, sRows, sCols, sRows);

    // Wait for GPU to finish before accessing on host
    cudaDeviceSynchronize();

    // Copy the result matrix back to the host memory
    cudaMemcpy(Sigma, d_Sigma, sRows * sRows * sizeof(double), cudaMemcpyDeviceToHost);

    // Free the device memory
    cudaFree(d_S);
    cudaFree(d_R);
    cudaFree(d_T);
    cudaFree(d_Sigma);
}

Start with the reported cgeist and polygeist-opt command, then inspect the stack frames in mlir/lib/Dialect/Affine/IR/AffineOps.cpp and lib/polygeist/Passes/ConvertPolygeistToLLVM.cpp:2972. Reproduce the crash with the CUDA input and reduce it to a minimal .cu example; done means identifying a reproducible failure suitable for the appropriate issue tracker.