TensorContractionThreadPool.h | 157 // - task grain sizes (number of kernels executed per task: gm, gn) 254 Index gn = 1; local 257 gm = coarsenM(m, n, bm, bn, bk, gn, num_threads, shard_by_col); 258 gn = coarsenN(m, n, bm, bn, bk, gm, num_threads, shard_by_col); 260 gn = coarsenN(m, n, bm, bn, bk, gm, num_threads, shard_by_col); 261 gm = coarsenM(m, n, bm, bn, bk, gn, num_threads, shard_by_col); 265 Index nn = divup(nn0, gn); 293 k, bm, bn, bk, nm, nn, nk, gm, gn, nm0, nn0, 306 Index gn, Index nm0, Index nn0, bool shard_by_col, 326 gn_(gn), 621 Index gn(Index n) const { return n + 1 < nn_ ? gn_ : nn0_ + gn_ - gn_ * nn_; } function in class:Eigen::TensorEvaluator::Context 681 Index gn = 1; local [all...] |