tide-GPR 0.0.9__py3-none-manylinux_2_28_x86_64.whl

tide/csrc/common_gpu.h

@@ -0,0 +1,56 @@
+#ifndef COMMON_GPU_H
+#define COMMON_GPU_H
+
+#include <cuda_runtime.h>
+
+// Define TIDE_DTYPE_FLOAT based on TIDE_DTYPE (for vectorized load optimization)
+// Note: TIDE_DTYPE_FLOAT is now defined in CMakelists.txt for better compatibility
+#ifndef TIDE_DTYPE_FLOAT
+#ifdef TIDE_DTYPE
+#if TIDE_DTYPE == float
+#define TIDE_DTYPE_FLOAT 1
+#elif TIDE_DTYPE == double
+#define TIDE_DTYPE_FLOAT 0
+#endif
+#endif
+#endif
+
+// Macro to check for CUDA kernel errors
+#define CHECK_KERNEL_ERROR                                              \
+  {                                                                     \
+    cudaError_t err = cudaGetLastError();                              \
+    if (err != cudaSuccess) {                                          \
+      fprintf(stderr, "CUDA kernel error: %s at %s:%d\n",              \
+              cudaGetErrorString(err), __FILE__, __LINE__);            \
+      exit(EXIT_FAILURE);                                              \
+    }                                                                   \
+  }
+
+// GPU error checking helper function
+inline void gpuAssert(cudaError_t code, const char *file, int line, bool abort = true) {
+  if (code != cudaSuccess) {
+    fprintf(stderr, "GPUassert: %s %s %d\n", cudaGetErrorString(code), file, line);
+    if (abort) exit(code);
+  }
+}
+
+// Custom atomicAdd for double precision (needed for older compute capabilities)
+// This is a software emulation using atomicCAS
+#if !defined(__CUDA_ARCH__) || __CUDA_ARCH__ < 600
+__device__ __forceinline__ double atomicAddDouble(double* address, double val) {
+    unsigned long long int* address_as_ull = (unsigned long long int*)address;
+    unsigned long long int old = *address_as_ull, assumed;
+    do {
+        assumed = old;
+        old = atomicCAS(address_as_ull, assumed,
+                        __double_as_longlong(val + __longlong_as_double(assumed)));
+    } while (assumed != old);
+    return __longlong_as_double(old);
+}
+#else
+__device__ __forceinline__ double atomicAddDouble(double* address, double val) {
+    return atomicAdd(address, val);
+}
+#endif
+
+#endif