chore: import upstream snapshot with attribution

2026-07-13 13:36:55 +08:00
commit c8a779b1bb
1887 changed files with 3245738 additions and 0 deletions
@@ -0,0 +1,96 @@
+/*
+ * SPDX-FileCopyrightText: Copyright (c) 1993-2024 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+#include "common/kernels/kernel.h"
+namespace nvinfer1
+{
+namespace plugin
+{
+#define CUBLAS_CHECK(condition)                                                                                        \
+    do                                                                                                                 \
+    {                                                                                                                  \
+        cublasStatus_t status = condition;                                                                             \
+        if (status != CUBLAS_STATUS_SUCCESS)                                                                           \
+        {                                                                                                              \
+            printf("%s %d CUBLAS FAIL %s\n", __FILE__, __LINE__, cublasGetErrorString(status));                        \
+        }                                                                                                              \
+    } while (0)
+
+// this scatter kernel works on a 2d table writing rows
+// index is 1-D array
+// updates is 2-D array
+// output is 2-D array
+// output[index[i]] = updates[i]
+__global__ void scatterKernel(
+    char* output,
+    const char* updates,
+    const int* indices,
+    int pitch,
+    int rowSize)
+{
+    int idx = indices[blockIdx.x];
+    char* pDst = (char*)output + idx * pitch;
+    const char* pSrc = updates + blockIdx.x * rowSize;
+    memcpy(pDst, pSrc, rowSize);
+}
+
+// Transform nd index to 1 - d index
+__global__ void transformIdxKernel(
+    int* output,
+    const int* transformCoeff, // these are actually the output pitches of the respective dimensions
+    const int* indices,
+    int sliceRank)
+{
+    const int* idx = indices + sliceRank * blockIdx.x;
+    int transformedIdx = 0;
+    for (int i = 0; i < sliceRank; i++)
+    {
+        transformedIdx += idx[i] * transformCoeff[i];
+    }
+    output[blockIdx.x] = transformedIdx;
+}
+
+
+pluginStatus_t scatterNDInference(
+    cudaStream_t stream,
+    int* transformCoeff,
+    int nOutputDims,
+    int sliceRank,
+    int nRows,
+    int rowSize,
+    int copySize,
+    int sizeOfElementInBytes,
+    const void* index,
+    const void* updates,
+    const void* data,
+    void* output,
+    void* workspace)
+{
+    const int* _index = (const int*)(index);
+    const char* _updates = (const char*)(updates);
+    char* _output = (char*)(output);
+    int* wo = (int*)(workspace);
+    int* transformedIdx = wo + sizeof(int)*nOutputDims;
+    int* deviceTransformCoeff = wo;
+    CSC(cudaMemcpy(workspace, transformCoeff, sizeof(int) * nOutputDims, cudaMemcpyHostToDevice), STATUS_FAILURE);
+    transformIdxKernel<<<nRows, 1, 0, stream>>>(transformedIdx, deviceTransformCoeff, _index, sliceRank);
+    CSC(cudaMemcpy(output, data, copySize, cudaMemcpyDeviceToDevice), STATUS_FAILURE);
+    // assuming output pitch = rowSize i.e no padding
+    scatterKernel<<<nRows, 1, 0, stream>>>(_output, _updates, transformedIdx, rowSize * 4, rowSize * 4);
+    return STATUS_SUCCESS;
+}
+} // namespace plugin
+} // namespace nvinfer1