Searched refs:backedges (Results 1 – 25 of 26) sorted by relevance
12
/external/llvm/test/Transforms/LoopSimplify/ |
D | indirectbr-backedge.ll | 3 ; LoopSimplify shouldn't split loop backedges that use indirectbr.
|
/external/llvm-project/llvm/test/Transforms/LoopSimplify/ |
D | indirectbr-backedge.ll | 3 ; LoopSimplify shouldn't split loop backedges that use indirectbr.
|
/external/llvm/test/Analysis/ScalarEvolution/ |
D | unsimplified-loop.ll | 3 ; This loop has no preheader, multiple backedges, etc., but ScalarEvolution
|
D | guards.ll | 4 ; conditions gaurding loop entries and backedges. This isn't intended
|
/external/llvm-project/llvm/test/Analysis/ScalarEvolution/ |
D | unsimplified-loop.ll | 4 ; This loop has no preheader, multiple backedges, etc., but ScalarEvolution
|
D | guards.ll | 4 ; conditions gaurding loop entries and backedges. This isn't intended
|
/external/llvm-project/mlir/lib/Transforms/ |
D | BufferDeallocation.cpp | 509 Backedges backedges(getFunction()); in runOnFunction() local 510 if (backedges.size()) { in runOnFunction()
|
/external/llvm-project/llvm/test/DebugInfo/MIR/X86/ |
D | livedebugvalues_loop_two_backedge_clobbered.mir | 5 ; their locations and have two backedges.
|
D | livedebugvalues_loop_two_backedge.mir | 5 ; backedges and beyond.
|
/external/llvm-project/llvm/test/Transforms/StructurizeCFG/AMDGPU/ |
D | loop-subregion-misordered.ll | 5 ; list to get the order. The only problem with it is that sometimes backedges 6 ; for outer loops will be visited before backedges for inner loops. To solve this problem,
|
/external/llvm-project/mlir/test/Transforms/ |
D | sccp-structured.mlir | 91 /// Check that arguments go to overdefined when loop backedges produce a
|
D | sccp.mlir | 161 /// Check that arguments go to overdefined when loop backedges produce a
|
D | promote-buffers-to-stack.mlir | 503 // that are passed via the backedges. The alloc is converted to an AllocaOp.
|
D | buffer-deallocation.mlir | 912 // that are passed via the backedges.
|
/external/llvm-project/llvm/test/Analysis/BlockFrequencyInfo/ |
D | basic.ll | 12 ; Loop backedges are weighted and thus their bodies have a greater frequency.
|
D | irreducible.ll | 50 ; that are targets of a backedge within it (excluding backedges within true 54 ; intercepts all the edges to the headers. All backedges and entries point to
|
/external/llvm/test/Analysis/BlockFrequencyInfo/ |
D | basic.ll | 12 ; Loop backedges are weighted and thus their bodies have a greater frequency.
|
D | irreducible.ll | 50 ; that are targets of a backedge within it (excluding backedges within true 54 ; intercepts all the edges to the headers. All backedges and entries point to
|
/external/llvm-project/llvm/docs/ |
D | BlockFrequencyTerminology.rst | 67 bottom-up, ignoring backedges; i.e., as a DAG. After each loop is processed,
|
D | LoopTerminology.rst | 239 times any of the backedges is taken before the loop. It is one less than
|
/external/llvm/docs/ |
D | BlockFrequencyTerminology.rst | 67 bottom-up, ignoring backedges; i.e., as a DAG. After each loop is processed,
|
D | Statepoints.rst | 752 loop backedges locations. Extending this to work with return polls would be
|
/external/tensorflow/tensorflow/compiler/jit/ |
D | deadness_analysis.cc | 986 std::vector<string> backedges; in CreateMultipleNextIterationInputsError() local 989 backedges.push_back(absl::StrCat(" ", SummarizeNode(*backedge->src()))); in CreateMultipleNextIterationInputsError() 994 FormatNodeForError(*merge), ": \n", absl::StrJoin(backedges, "\n"), in CreateMultipleNextIterationInputsError()
|
/external/llvm-project/llvm/test/CodeGen/WebAssembly/ |
D | cfg-stackify.ll | 13 ; Test that loops are made contiguous, even in the presence of split backedges. 505 ; Test a case where there are multiple backedges and multiple loop exits
|
/external/llvm/test/CodeGen/WebAssembly/ |
D | cfg-stackify.ll | 14 ; Test that loops are made contiguous, even in the presence of split backedges. 691 ; Test a case where there are multiple backedges and multiple loop exits
|
12