Home
last modified time | relevance | path

Searched full:batched (Results 1 – 20 of 20) sorted by relevance

/llvm-project/mlir/test/Dialect/Vector/
H A Dvector-unroll-options.mlir3 // RUN: mlir-opt %s -test-vector-unrolling-patterns="unroll-based-on-type unroll-order=0,3,1,2" | FileCheck %s --check-prefix=BATCHED
291 // BATCHED-LABEL: vector_contract_batched
292 // BATCHED-COUNT-16: vector.contract
293 // BATCHED-NOT: vector.contract
294 // BATCHED: return
/llvm-project/llvm/docs/HistoricalNotes/
H A D2002-06-25-MegaPatchInfo.txt47 include every .h file that it used. Now things are batched a little bit more
/llvm-project/mlir/include/mlir/Dialect/SparseTensor/IR/
H A DSparseTensorType.h253 /// Returns the batched level-rank.
256 /// Returns the batched level-shape.
H A DSparseTensorAttrDefs.td242 // Batched sorted coordinate scheme, with high encoding.
/llvm-project/mlir/python/mlir/dialects/linalg/opdsl/ops/
H A Dcore_named_ops.py472 """Performs a batched matrix-matrix-transpose multiplication of two
473 batched-4D (5D) inputs.
493 """Performs a batched matrix multiplication of two 3D inputs.
511 """Performs a batched matrix multiplication of two 3D inputs where lhs operand
530 """Performs a batched matrix multiplication of two 3D inputs where rhs operand
551 """Performs a batched matrix multiplication of two 3D inputs.
617 """Performs a batched matrix-vector multiplication.
635 """Performs a batched matrix-vector multiplication.
/llvm-project/llvm/utils/
H A Dupdate_mca_test_checks.py134 # point so flush any batched-up lines one prefix at a time to the output
161 # Flush any remaining batched-up lines one prefix at a time to the output.
/llvm-project/llvm/lib/IR/
H A DPass.cpp10 // responsible with ensuring that passes are executed and batched together
/llvm-project/mlir/include/mlir/Dialect/Linalg/IR/
H A DLinalgInterfaces.td103 batched vector-matrix multiplication.
125 batched matrix-vector multiplication.
H A DLinalgNamedStructuredOps.yaml1401 Performs a batched matrix-matrix-transpose multiplication of two
1402 batched-4D (5D) inputs.
1479 Performs a batched matrix multiplication of two 3D inputs.
1548 Performs a batched matrix multiplication of two 3D inputs where lhs operand
1618 Performs a batched matrix multiplication of two 3D inputs where rhs operand
1688 Performs a batched matrix multiplication of two 3D inputs.
1997 Performs a batched matrix-vector multiplication.
2065 Performs a batched matrix-vector multiplication.
/llvm-project/mlir/docs/Dialects/
H A DTransform.md64 object ("batched execution"). Deviations from this convention are described in
/llvm-project/mlir/lib/Dialect/SparseTensor/Transforms/
H A DSparseTensorCodegen.cpp1397 SmallVector<Value> batched(stt.getBatchLvlRank(), in matchAndRewrite()
1399 batched.push_back(posBack); in matchAndRewrite()
1400 memSize = genIndexLoad(rewriter, loc, desc.getPosMemRef(lvl), batched); in matchAndRewrite()
1388 SmallVector<Value> batched(stt.getBatchLvlRank(), matchAndRewrite() local
/llvm-project/polly/include/polly/
H A DScopDetection.h143 BatchAAResults BAA; // The batched alias analysis results.
/llvm-project/llvm/lib/Transforms/IPO/
H A DInliner.cpp308 // We expect the calls to typically be batched with sequences of calls that in run()
/llvm-project/mlir/lib/Dialect/Linalg/Transforms/
H A DConvertConv2DToImg2Col.cpp599 // Because we didn't transpose the filters we don't actually have a batched in rewriteInIm2Col()
H A DDropUnitDims.cpp966 /// Patterns for unbatching batched contraction ops
/llvm-project/mlir/include/mlir/Dialect/Linalg/Transforms/
H A DTransforms.h1287 /// the N input. For the case where N > 1 its a batched matrix-matrix
1306 /// outer most the computation is a batched matrix-vector product.
/llvm-project/mlir/docs/
H A DBufferization.md220 // Batched TOSA matrix multiplication. %A and %B are the
/llvm-project/mlir/include/mlir/Dialect/Tosa/IR/
H A DTosaOps.td191 Performs a batched complex 2D Fast Fourier Transform over the input. The
309 Performs a batched 2D real-valued Fast Fourier Transform over the input where
/llvm-project/mlir/include/mlir/Dialect/Linalg/TransformOps/
H A DLinalgTransformOps.td2449 the N input. For the case where N > 1 its a batched matrxi-matrix
2730 Winograd Conv2D algorithm will convert linalg Conv2D operation into batched
2732 input into a format suitable for batched matrix multiply. After the matrix
/llvm-project/llvm/lib/Analysis/
H A DLoopAccessAnalysis.cpp760 /// Batched alias analysis results. in adjustAliasScopeList()