Home
last modified time | relevance | path

Searched refs:IVs (Results 1 – 21 of 21) sorted by relevance

/llvm-project-15.0.7/mlir/test/Dialect/Affine/
H A Dload-store.mlir5 // Test with just loop IVs.
19 // Test with loop IVs and constants.
35 // Test with loop IVs and function args without 'symbol' keyword (should
52 // Test with loop IVs and function args with 'symbol' keyword (should
71 // Test with loop IVs, symbols and constants in nested affine expressions.
89 // Test with swizzled loop IVs.
107 // Test with swizzled loop IVs, duplicate args, and function args used as dims.
129 // Test with swizzled loop IVs, duplicate args, and function args used as syms.
184 // Test with loop IVs and constants.
201 // Test with just loop IVs.
[all …]
H A Ddma.mlir5 // Test with loop IVs.
26 // Test with loop IVs and optional stride arguments.
49 // Test with loop IVs and symbols (without symbol keyword).
71 // Test with loop IVs and symbols (with symbol keyword).
94 // Test with loop IVs, symbols and constants in nested affine expressions.
H A Dloop-coalescing.mlir28 // Reconstruct original IVs from the linearized one.
111 // The IVs are rewritten.
H A Ddma-generate.mlir179 // dependent on outer loop IVs.
/llvm-project-15.0.7/llvm/test/Transforms/IndVarSimplify/
H A D2011-09-19-vectoriv.ll2 ; PR10946: Vector IVs are not SCEVable.
H A Dconst_phi.ll9 ; congruent IVs.
H A Dphi-uses-value-multiple-times.ll3 ; Check that IndVarSimplify is not creating unnecessary canonical IVs
H A Dvariable-stride-ivs-0.ll3 ; Test that -indvars can reduce variable stride IVs. If it can reduce variable
H A Delim-extend.ll184 ; IVs in this loop. sext elimination plus LFTR currently results in 2 final
185 ; IVs. Waiting to remove LFTR.
H A Dno-iv-rewrite.ll4 ; Make sure that indvars isn't inserting canonical IVs.
9 ; We should only have 2 IVs.
62 ; We should only have 2 IVs.
H A Dlftr-dead-ivs.ll227 ; For integer IVs, we handle this trigger case by stripping the problematic
H A Dlftr-multi-exit.ll278 ;; one IV when exit tests involve multiple IVs).
H A Dloop-predication.ll635 ; Demonstrate that this approach works with IVs of different steps, and types
/llvm-project-15.0.7/llvm/lib/Transforms/Scalar/
H A DSCCP.cpp109 std::vector<ValueLatticeElement> IVs = Solver.getStructLatticeValueFor(V); in tryToReplaceWithConstant() local
110 if (llvm::any_of(IVs, isOverdefined)) in tryToReplaceWithConstant()
115 ValueLatticeElement V = IVs[i]; in tryToReplaceWithConstant()
/llvm-project-15.0.7/llvm/test/Transforms/LoopVectorize/X86/
H A Dconsecutive-ptr-cg-bug.ll23 ; solution is to detect the problematic IVs in LV (%7 and %8) and perform an
/llvm-project-15.0.7/llvm/test/Transforms/LoopStrengthReduce/ARM/
H A D2012-06-15-lsr-noaddrmode.ll8 ; to coalesce the IV. See <rdar://problem/11680670> [coalescer] IVs
/llvm-project-15.0.7/mlir/include/mlir/Dialect/Affine/IR/
H A DAffineOps.td508 same type as the elements of the memref. An affine expression of loop IVs
870 same type as the elements of the memref. An affine expression of loop IVs
945 An affine expression of loop IVs and symbols must be specified for each
1012 An affine expression of loop IVs and symbols must be specified for each
/llvm-project-15.0.7/mlir/docs/Dialects/
H A DAffine.md320 elements of the memref. An affine expression of loop IVs and symbols must be
350 elements of the memref. An affine expression of loop IVs and symbols must be
/llvm-project-15.0.7/llvm/test/Transforms/IndVarSimplify/ARM/
H A Dcode-size.ll725 ; Demonstrate that this approach works with IVs of different steps, and types
/llvm-project-15.0.7/mlir/test/Transforms/
H A Dloop-fusion.mlir917 // loop IVs, so we should slice at depth 1 and insert the slice at depth 2.
/llvm-project-15.0.7/llvm/test/Transforms/LoopVectorize/
H A Dinduction.ll4318 ; Check that we generate vectorized IVs in the pre-header
5700 ; correctly after all the phis. These new phis correspond to new IVs
5701 ; that are generated by optimizing non-free truncs of IVs to IVs themselves.