Searched refs:divup (Results 1 – 5 of 5) sorted by relevance
/external/eigen/unsupported/Eigen/CXX11/src/Tensor/ |
D | TensorDeviceThreadPool.h | 207 Index block_count = divup(n, block_size); in parallelFor() 212 (divup<int>(block_count, numThreads()) * numThreads()); in parallelFor() 218 Index coarser_block_size = divup(n, prev_block_count - 1); in parallelFor() 228 const Index coarser_block_count = divup(n, coarser_block_size); in parallelFor() 233 (divup<int>(coarser_block_count, numThreads()) * numThreads()); in parallelFor() 257 Index mid = first + divup((last - first) / 2, block_size) * block_size; in parallelFor()
|
D | TensorContractionThreadPool.h | 244 Index nm0 = divup(m, bm); 245 Index nn0 = divup(n, bn); 246 Index nk = divup(k, bk); 264 Index nm = divup(nm0, gm); 265 Index nn = divup(nn0, gn); 358 divup<size_t>(bm_ * bk_ * sizeof(LhsScalar), align) * align; 360 divup<size_t>(bn_ * bk_ * sizeof(RhsScalar), align) * align; 659 Index nm0 = divup(m, bm); 665 while (gm1 <= nm0 && nm1 == divup(nm0, gm1)) gm1++; 671 nm1 = divup(nm0, gm1); [all …]
|
D | TensorReductionCuda.h | 293 const int num_blocks = divup<int>(num_coeffs, block_size * num_per_thread); 320 const int num_blocks = divup<int>(num_coeffs, block_size * num_per_thread); 386 const Index input_col_blocks = divup<Index>(num_coeffs_to_reduce, blockDim.x * NumPerThread); 457 const Index input_col_blocks = divup<Index>(num_coeffs_to_reduce, blockDim.x * NumPerThread * 2); 458 const Index num_input_blocks = divup<Index>(input_col_blocks * num_preserved_coeffs, 2); 561 const int dyn_blocks = divup<int>(num_coeffs, block_size * num_per_thread); 569 const int dyn_blocks = divup<int>(num_preserved_vals, 1024); 607 const int dyn_blocks = divup<int>(num_coeffs, block_size * num_per_thread); 615 const int dyn_blocks = divup<int>(num_preserved_vals, 1024); 680 const Index max_iter = num_preserved_coeffs * divup<Index>(num_coeffs_to_reduce, NumPerThread); [all …]
|
D | TensorMeta.h | 30 T divup(const X x, const Y y) { in divup() function 36 T divup(const T x, const T y) { in divup() function
|
D | TensorExecutor.h | 258 …const int num_blocks = numext::maxi<int>(numext::mini<int>(max_blocks, divup<int>(size, block_size…
|