Skip to content

GitLab

Unverified Commit 8708851e authored by q.yao's avatar q.yao Committed by GitHub
Browse files

[Fix] Fix bbox overlap fp16 (#1958) 

* add CUDA_ARCH check

* add check in cuh
parent a3b4640b
Hide whitespace changes
Inline Side-by-side
Showing with 4 additions and 0 deletions
mmcv/ops/csrc/common/cuda/bbox_overlaps_cuda_kernel.cuh
View file @ 8708851e
...@@ -88,6 +88,7 @@ __global__ void bbox_overlaps_cuda_kernel(const T* bbox1, const T* bbox2, ...@@ -88,6 +88,7 @@ __global__ void bbox_overlaps_cuda_kernel(const T* bbox1, const T* bbox2,
} }
} }
#if __CUDA_ARCH__ >= 530
__device__ __forceinline__ __half __half_area(const __half x1, const __half y1, __device__ __forceinline__ __half __half_area(const __half x1, const __half y1,
const __half x2, const __half y2, const __half x2, const __half y2,
const __half offset) { const __half offset) {
...@@ -141,5 +142,6 @@ __device__ void bbox_overlaps_cuda_kernel_half( ...@@ -141,5 +142,6 @@ __device__ void bbox_overlaps_cuda_kernel_half(
ious[index] = __hdiv(interS, baseS); ious[index] = __hdiv(interS, baseS);
} }
} }
#endif // __CUDA_ARCH__ >= 530
#endif // BBOX_OVERLAPS_CUDA_KERNEL_CUH #endif // BBOX_OVERLAPS_CUDA_KERNEL_CUH
mmcv/ops/csrc/pytorch/cuda/bbox_overlaps_cuda.cu
View file @ 8708851e
...@@ -4,6 +4,7 @@ ...@@ -4,6 +4,7 @@
// Disable fp16 on ROCm device // Disable fp16 on ROCm device
#ifndef HIP_DIFF #ifndef HIP_DIFF
#if __CUDA_ARCH__ >= 530
template <> template <>
__global__ void bbox_overlaps_cuda_kernel<at::Half>( __global__ void bbox_overlaps_cuda_kernel<at::Half>(
const at::Half* bbox1, const at::Half* bbox2, at::Half* ious, const at::Half* bbox1, const at::Half* bbox2, at::Half* ious,
...@@ -14,6 +15,7 @@ __global__ void bbox_overlaps_cuda_kernel<at::Half>( ...@@ -14,6 +15,7 @@ __global__ void bbox_overlaps_cuda_kernel<at::Half>(
reinterpret_cast<__half*>(ious), num_bbox1, reinterpret_cast<__half*>(ious), num_bbox1,
num_bbox2, mode, aligned, offset); num_bbox2, mode, aligned, offset);
} }
#endif // __CUDA_ARCH__ >= 530
#endif // HIP_DIFF #endif // HIP_DIFF
void BBoxOverlapsCUDAKernelLauncher(const Tensor bboxes1, const Tensor bboxes2, void BBoxOverlapsCUDAKernelLauncher(const Tensor bboxes1, const Tensor bboxes2,
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment