/llvm-project/mlir/python/mlir/dialects/linalg/opdsl/ops/ |
H A D | core_named_ops.py | 17 Numeric casting is performed on the input operand, promoting it to the same 33 Numeric casting is performed on the input operand, promoting it to the same 205 Numeric casting is performed on the input operand, promoting it to the same 396 Numeric casting is performed on the operands to the inner multiply, promoting 417 Numeric casting is performed on the operands to the inner multiply, promoting 435 Numeric casting is performed on the operands to the inner multiply, promoting 495 Numeric casting is performed on the operands to the inner multiply, promoting 514 Numeric casting is performed on the operands to the inner multiply, promoting 533 Numeric casting is performed on the operands to the inner multiply, promoting 553 Numeric casting is performed on the operands to the inner multiply, promoting [all...] |
/llvm-project/mlir/include/mlir/Dialect/Linalg/IR/ |
H A D | LinalgNamedStructuredOps.yaml | 10 Numeric casting is performed on the input operand, promoting it to the same 53 Numeric casting is performed on the input operand, promoting it to the same 558 Numeric casting is performed on the input operand, promoting it to the same 1074 Numeric casting is performed on the operands to the inner multiply, promoting 1179 Numeric casting is performed on the operands to the inner multiply, promoting 1252 Numeric casting is performed on the operands to the inner multiply, promoting 1481 Numeric casting is performed on the operands to the inner multiply, promoting 1551 Numeric casting is performed on the operands to the inner multiply, promoting 1621 Numeric casting is performed on the operands to the inner multiply, promoting 1690 Numeric casting is performed on the operands to the inner multiply, promoting [all...] |
/llvm-project/llvm/docs/ |
H A D | AssignmentTracking.md | 168 **Promoting** allocas and store/loads: `#dbg_assign` records implicitly 170 necessarily the case after promoting (or partially promoting) the
|
/llvm-project/llvm/lib/Transforms/IPO/ |
H A D | ArgumentPromotion.cpp | 124 // that we are *not* promoting. For the ones that we do promote, the parameter in doPromotion() 161 << "promoting argument " << ore::NV("ArgName", I->getName()) in doPromotion() 188 LLVM_DEBUG(dbgs() << "ARG PROMOTION: Promoting to:" << *NF << "\n" in doPromotion() 522 // Promoting the argument causes it to be loaded in the caller in findArgParts() 567 // then promoting it might lead to recursive promotion. in findArgParts() 577 // We limit promotion to only promoting up to a fixed number of elements of in findArgParts() 691 // We limit promotion to only promoting up to a fixed number of elements in areTypesABICompatible() 847 // FIXME: Support promoting whole chain of musttail functions in run() 906 // Iterate until we stop promoting from this SCC.
|
/llvm-project/llvm/test/Transforms/GlobalOpt/ |
H A D | 2012-05-11-blockaddress.ll | 3 ; from promoting @f to fastcc.
|
/llvm-project/llvm/test/Transforms/Mem2Reg/ |
H A D | 2003-06-26-IterativePromote.ll | 1 ; Promoting some values allows promotion of other values.
|
/llvm-project/clang-tools-extra/docs/clang-tidy/checks/readability/ |
H A D | math-missing-parentheses.rst | 14 ambiguity and errors, promoting clearer and more maintainable code.
|
/llvm-project/mlir/include/mlir/Transforms/ |
H A D | Mem2Reg.h | 26 /// retries the promotion of all slots as promoting one slot might enable
|
/llvm-project/llvm/test/CodeGen/AMDGPU/ |
H A D | vector-alloca-addrspacecast.ll | 3 ; Should give up promoting alloca to vector with an addrspacecast.
|
H A D | promote-alloca-non-constant-index.ll | 4 ; Check that promoting an alloca to a vector form works correctly when a variable
|
/llvm-project/clang/test/Sema/ |
H A D | bitint-bitfield-promote.c | 4 // type was promoting to int rather than the type of the bit-field.
|
/llvm-project/llvm/lib/Target/AMDGPU/ |
H A D | AMDGPUPromoteAlloca.cpp | 18 // much later in the pipeline. This runs after SROA because promoting to 339 // promoting to LDS instead. in run() 444 /// be promoted now. This happens when promoting requires \p 654 llvm_unreachable("Unsupported call when promoting alloca to vector"); in promoteAllocaUserToVector() 661 llvm_unreachable("Did not return after promoting instruction!"); in promoteAllocaUserToVector() 1113 // Figure out which operand is the one we might not be promoting. in isCallPromotable() 1255 LLVM_DEBUG(dbgs() << "Function has local memory argument. Promoting to " in hasSufficientLocalMem() 1320 "local memory. Promoting to local memory " in hasSufficientLocalMem() 1451 LLVM_DEBUG(dbgs() << "Promoting alloca to local memory\n"); in tryPromoteAllocaToLDS()
|
/llvm-project/llvm/test/Transforms/ArgumentPromotion/ |
H A D | 2008-07-02-array-indexing.ll | 5 ; This test tries to convince CHECK about promoting the load from %A + 2,
|
H A D | reserve-tbaa.ll | 8 ; original loads to the newly-created loads when promoting pointer arguments.
|
/llvm-project/mlir/lib/Transforms/ |
H A D | Mem2Reg.cpp | 163 /// The MemorySlotPromoter handles the state of promoting a memory slot. It 174 /// Actually promotes the slot by mutating IR. Promoting a slot DOES 681 // A break is required, since promoting a slot may invalidate the in tryToPromoteMemorySlots() 731 // Attempt promoting as many of the slots as possible.
|
/llvm-project/llvm/test/Transforms/InstCombine/ |
H A D | 2007-02-01-LoadSinkAlloca.ll | 7 ; cond_next. Doing so prevents mem2reg from promoting the B alloca.
|
/llvm-project/llvm/test/CodeGen/SystemZ/ |
H A D | fp-conv-10.ll | 8 ; Promoting to i64 doesn't generate an inexact condition for values that are
|
H A D | fp-strict-conv-10.ll | 8 ; Promoting to i64 doesn't generate an inexact condition for values that are
|
/llvm-project/llvm/test/Transforms/Attributor/ArgumentPromotion/ |
H A D | 2008-07-02-array-indexing.ll | 6 ; This test tries to convince CHECK about promoting the load from %A + 2,
|
/llvm-project/mlir/lib/Dialect/LLVMIR/Transforms/ |
H A D | OptimizeForNVVM.cpp | 30 // This as accurate as promoting the division to fp32 in the NVPTX backend, but
|
/llvm-project/mlir/test/lib/Dialect/Affine/ |
H A D | TestAffineDataCopy.cpp | 124 // Promoting single iteration loops could lead to simplification of in runOnOperation()
|
/llvm-project/llvm/test/DebugInfo/Generic/ |
H A D | mem2reg-promote-alloca-3.ll | 6 ;; promoting the alloca %local after 'esc' has been inlined. Without this we
|
H A D | mem2reg-promote-alloca-1.ll | 5 ;; when promoting the alloca %param.addr.
|
/llvm-project/llvm/lib/CodeGen/ |
H A D | TypePromotion.cpp | 81 // We achieve this by type promoting the IR to i32 like so for this example: 459 LLVM_DEBUG(dbgs() << "IR Promotion: Promoting sources:\n"); in ExtendSources() 657 LLVM_DEBUG(dbgs() << "IR Promotion: Promoting use-def chains to " in Mutate() 809 // Ignore GEPs because they don't need promoting and the constant indices in TryToPromote()
|
/llvm-project/llvm/include/llvm/Transforms/Utils/ |
H A D | CallPromotionUtils.h | 9 // This file declares utilities useful for promoting indirect call sites to
|