hip_check.hpp File Reference

PULSATRIX_HIP_CHECK – converts a HIP runtime failure into a thrown C++ exception at the DeviceBackend boundary, per gpu_backend_programming/context_gpu_cuda_kernel_mechanics.md's Error Checking pattern. Only compiled when PULSATRIX_ENABLE_HIP is set. More...

#include <hip/hip_runtime.h>
#include <stdexcept>
#include <string>

Include dependency graph for hip_check.hpp:

Go to the source code of this file.

Macros
#define	PULSATRIX_HIP_CHECK(call)
	Checks a HIP runtime call's hipError_t result; throws std::runtime_error with the HIP error string and source location if it did not succeed.

Detailed Description

PULSATRIX_HIP_CHECK – converts a HIP runtime failure into a thrown C++ exception at the DeviceBackend boundary, per gpu_backend_programming/context_gpu_cuda_kernel_mechanics.md's Error Checking pattern. Only compiled when PULSATRIX_ENABLE_HIP is set.

Macro Definition Documentation

PULSATRIX_HIP_CHECK

#define PULSATRIX_HIP_CHECK

(

call

)

Value:

    do {                                                                                          \
        hipError_t pulsatrix_hip_check_result = (call);                                                 \
        if (pulsatrix_hip_check_result != hipSuccess) {                                                  \
            throw std::runtime_error(std::string("HIP error: ") +                                    \
                                      hipGetErrorString(pulsatrix_hip_check_result) + " at " __FILE__ +    \
                                      ":" + std::to_string(__LINE__));                                 \
        }                                                                                                \
    } while (0)

Checks a HIP runtime call's hipError_t result; throws std::runtime_error with the HIP error string and source location if it did not succeed.

Note

Kernel launches are asynchronous – this macro catches launch-configuration errors immediately via hipGetLastError(), but in-kernel runtime errors only surface at the next synchronizing call. Route every allocate/copy/kernel-launch/sync call in HIPBackend through this macro, including a hipGetLastError() check right after every kernel launch.

Deliberately identical in shape to PULSATRIX_CUDA_CHECK (cuda_check.hpp) – HIPIFY maps cudaError_t/cudaSuccess/cudaGetErrorString one-to-one onto their hip* counterparts, so there is nothing to redesign here and a divergent shape would only make the two backends harder to compare.