| /llvm-project-15.0.7/mlir/lib/Rewrite/ |
| H A D | PatternApplicator.cpp | 26 if (const PDLByteCode *bytecode = frozenPatternList.getPDLByteCode()) { in PatternApplicator() local 28 bytecode->initializeMutableState(*mutableByteCodeState); in PatternApplicator() 55 if (const PDLByteCode *bytecode = frozenPatternList.getPDLByteCode()) { in applyCostModel() local 56 for (const auto &it : llvm::enumerate(bytecode->getPatterns())) in applyCostModel() 120 if (const PDLByteCode *bytecode = frozenPatternList.getPDLByteCode()) { in walkAllPatterns() local 121 for (const Pattern &it : bytecode->getPatterns()) in walkAllPatterns() 135 const PDLByteCode *bytecode = frozenPatternList.getPDLByteCode(); in matchAndRewrite() local 136 if (bytecode) in matchAndRewrite() 137 bytecode->match(op, rewriter, pdlMatches, *mutableByteCodeState); in matchAndRewrite() 194 bytecode->rewrite(rewriter, *pdlMatch, *mutableByteCodeState); in matchAndRewrite()
|
| H A D | ByteCode.cpp | 336 : bytecode(bytecode), generator(generator) {} in ByteCodeWriter() 339 void append(ByteCodeField field) { bytecode.push_back(field); } in append() 340 void append(OpCode opCode) { bytecode.push_back(opCode); } in append() 349 bytecode.append({fieldParts[0], fieldParts[1]}); in append() 357 unresolvedSuccessorRefs[successor].push_back(bytecode.size()); in append() 370 bytecode.push_back(values.size()); in appendPDLValueList() 399 bytecode.push_back(static_cast<ByteCodeField>(kind)); in appendPDLValueKind() 408 bytecode.push_back(generator.getMemIndex(value)); in append() 415 bytecode.push_back(llvm::size(range)); in append() 435 bytecode.append(fieldParts, fieldParts + numParts); in appendInline() [all …]
|
| /llvm-project-15.0.7/llvm/docs/HistoricalNotes/ |
| H A D | 2000-12-06-MeetingSummary.txt | 4 We met to discuss the LLVM instruction format and bytecode representation: 20 bytecode would have to parse and handle it. This would slow down the 38 1. Including dominator information in the LLVM bytecode 42 in the LLVM bytecode program. Basic blocks could be numbered according 43 to the order of occurrence in the bytecode representation. 66 runtime to skip information that it didn't understand in a bytecode
|
| H A D | 2001-02-13-Reference-MemoryResponse.txt | 34 > no way to represent this in either the bytecode or assembly. 35 > B. When parsing assembly/bytecode, we effectively have to do a full 42 how big the bytecode files are before expanding them further. I am pretty 43 keen to explore the implications of LLVM for mobile devices. Both bytecode
|
| H A D | 2001-02-13-Reference-Memory.txt | 30 no way to represent this in either the bytecode or assembly. 31 B. When parsing assembly/bytecode, we effectively have to do a full 36 bytecode to get around this... what do you think?
|
| H A D | 2001-02-09-AdveCommentsResponse.txt | 59 > - it could make the bytecode significantly larger because there could 86 No, it isn't. Remember that the bytecode encoding saves value slots into 87 the bytecode instructions themselves, not constant values. This is
|
| H A D | 2003-06-26-Reoptimizer2.txt | 57 2. Run opt -lowerswitch -paths -emitfuncs on the bytecode.
|
| H A D | 2001-02-09-AdveComments.txt | 41 - it could make the bytecode significantly larger because there could
|
| H A D | 2003-06-25-Reoptimizer1.txt | 6 We use opt to do Bytecode-to-bytecode instrumentation. Look at
|
| H A D | 2000-11-18-EarlyDesignIdeasResp.txt | 26 * Java bytecode is used as our "safe" representation (to avoid
|
| /llvm-project-15.0.7/llvm/test/Assembler/ |
| H A D | 2002-08-16-ConstExprInlined.ll | 1 ; In this testcase, the bytecode reader or writer is not correctly handling the 7 ; disassembling this would be good, but in addition to that, the bytecode
|
| H A D | 2004-10-22-BCWriterUndefBug.ll | 1 ;; The bytecode writer was trying to treat undef values as ConstantArray's when
|
| H A D | 2002-08-19-BytecodeReader.ll | 1 ; Testcase that seems to break the bytecode reader. This comes from the
|
| /llvm-project-15.0.7/llvm/test/Verifier/ |
| H A D | non-integer-gep-index.ll | 8 ; impossible to load from .ll file. Hence in this test we use bytecode input.
|
| /llvm-project-15.0.7/llvm/test/Integer/ |
| H A D | basictest_bt.ll | 6 ; Test "stripped" format where nothing is symbolic... this is how the bytecode
|
| H A D | constpointer_bt.ll | 7 ; declared and involves an icky bytecode encoding. There is no meaningful
|
| /llvm-project-15.0.7/llvm/test/Feature/ |
| H A D | basictest.ll | 5 ; Test "stripped" format where nothing is symbolic... this is how the bytecode
|
| H A D | constpointer.ll | 7 ; declared and involves an icky bytecode encoding. There is no meaningful
|
| /llvm-project-15.0.7/clang/docs/ |
| H A D | ConstantInterpreter.rst | 23 backends: one to generate bytecode for functions (``ByteCodeEmitter``) and 25 generating bytecode (``EvalEmitter``). All functions are compiled to 26 bytecode, while toplevel expressions used in constant contexts are directly 27 evaluated since the bytecode would never be reused. This mechanism aims to 211 Descriptors are generated at bytecode compilation time and contain information 371 * Continue-after-failure: a form of exception handling at the bytecode
|
| /llvm-project-15.0.7/llvm/test/CodeGen/WebAssembly/ |
| H A D | switch-in-loop.ll | 4 ;; bytecode interpreter) lowers reasonably without any irreducible
|
| /llvm-project-15.0.7/llvm/test/Object/ |
| H A D | archive-extract.test | 1 ; This test just makes sure that llvm-ar can extract bytecode members
|
| /llvm-project-15.0.7/llvm/docs/ |
| H A D | CFIVerify.rst | 62 are not implemented due to a lack of information provided by the bytecode.
|
| /llvm-project-15.0.7/llvm/test/Transforms/SampleProfile/ |
| H A D | gcc-simple.ll | 18 ; This test was compiled down to bytecode at -O0 to avoid inlining foo() into
|
| /llvm-project-15.0.7/mlir/test/Integration/Dialect/PDL/CPU/ |
| H A D | multiroot.mlir | 1 // RUN: mlir-opt %s -allow-unregistered-dialect -test-pdl-bytecode-pass -split-input-file | FileCh…
|
| /llvm-project-15.0.7/mlir/include/mlir/Dialect/PDLInterp/IR/ |
| H A D | PDLInterpOps.td | 599 In the bytecode interpreter, this operation is implemented by looping over 600 the values and, for each selection, running the bytecode until we reach
|