Home
last modified time | relevance | path

Searched refs:commutative (Results 1 – 25 of 65) sorted by relevance

123

/llvm-project-15.0.7/llvm/test/MC/ARM/
H A Dthumb2-narrow-dp.ll9 // In addition, some operations are commutative, allowing the transformation
42 // ADD register (not SP) A8.8.6 (commutative)
92 // ADD (SP plus register) A8.8.10 (commutative)
105 // AND (commutative)
108 ANDS r2, r1, r2 // Should choose narrow - commutative
140 ANDEQ r3, r1, r3 // Should choose narrow - commutative
192 // EOR (commutative)
195 EORS r5, r1, r5 // Should choose narrow - commutative
227 EOREQ r2, r1, r2 // Should choose narrow - commutative
486 // ADC (commutative)
[all …]
/llvm-project-15.0.7/llvm/test/CodeGen/AArch64/
H A Daddsub-shifted.ll30 ; Subtraction is not commutative!
61 ; Subtraction is not commutative!
96 ; Subtraction is not commutative!
122 ; Subtraction is not commutative!
155 ; Subtraction is not commutative!
181 ; Subtraction is not commutative!
H A Dsink-addsub-of-const.ll7 ; Outer 'add' is commutative - 2 variants.
31 ; Outer 'add' is commutative - 2 variants.
55 ; Outer 'add' is commutative - 2 variants.
156 ; Outer 'add' is commutative - 2 variants.
184 ; Outer 'add' is commutative - 2 variants.
212 ; Outer 'add' is commutative - 2 variants.
/llvm-project-15.0.7/llvm/test/Transforms/IROutliner/
H A Doutlining-commutative-fp.ll4 ; This test checks that floating point commutative instructions are not treated
5 ; as commutative. Even though an ffadd is technically commutative, the order
H A Doutlining-commutative.ll4 ; This test checks that commutative instructions where the operands are
7 ; It also checks that non-commutative instructions outlined as different
12 ; are commutative, we should still outline from all four as the same
117 ; instructions are not commutative, we should outline the first two functions
H A Doutlining-commutative-operands-opposite-order.ll4 ; This is a test to ensure that when the first instruction is a commutative
/llvm-project-15.0.7/llvm/test/CodeGen/X86/
H A Dsink-addsub-of-const.ll8 ; Outer 'add' is commutative - 2 variants.
50 ; Outer 'add' is commutative - 2 variants.
92 ; Outer 'add' is commutative - 2 variants.
258 ; Outer 'add' is commutative - 2 variants.
294 ; Outer 'add' is commutative - 2 variants.
330 ; Outer 'add' is commutative - 2 variants.
H A Dvector-bo-select.ll161 ; negative test - fsub is not commutative; there is no identity constant for operand 0
226 ; negative test - fsub is not commutative; there is no identity constant for operand 0
906 ; negative test - sub is not commutative; there is no identity constant for operand 0
971 ; negative test - sub is not commutative; there is no identity constant for operand 0
1242 ; negative test - shl is not commutative; there is no identity constant for operand 0
1307 ; negative test - shl is not commutative; there is no identity constant for operand 0
1429 ; negative test - lshr is not commutative; there is no identity constant for operand 0
1494 ; negative test - lshr is not commutative; there is no identity constant for operand 0
1616 ; negative test - ashr is not commutative; there is no identity constant for operand 0
1681 ; negative test - ashr is not commutative; there is no identity constant for operand 0
/llvm-project-15.0.7/llvm/test/CodeGen/AArch64/GlobalISel/
H A Dselect-rotate.mir16 ; operand instead of RHS since they're both rotates and XOR is commutative.
/llvm-project-15.0.7/llvm/test/TableGen/
H A DGlobalISelEmitterCustomPredicate.td132 // Test commutative, standalone pattern.
176 // Test a non-commutative pattern.
208 // Test a commutative pattern using multiple patterns using PatFrags.
/llvm-project-15.0.7/mlir/lib/Dialect/Arithmetic/IR/
H A DArithmeticCanonicalization.td25 // addi is commutative and will be canonicalized to have its constants appear
99 // xori is commutative and will be canonicalized to have its constants appear
/llvm-project-15.0.7/llvm/test/Transforms/InstCombine/
H A Dfold-sub-of-not-to-inc-of-add.ll74 ; The `sub` (and the fold) is not commutative.
H A Dhoist-negation-out-of-bias-calculation.ll40 %unbiasedx = and i8 %x, %negy ; commutative, swapped
H A Dcanonicalize-constant-low-bit-mask-and-icmp-uge-to-icmp-ule.ll99 ; The pattern is not commutative. instsimplify will already take care of it.
H A Dcanonicalize-constant-low-bit-mask-and-icmp-ult-to-icmp-ugt.ll110 ; The pattern is not commutative. instsimplify will already take care of it.
H A Dcanonicalize-constant-low-bit-mask-and-icmp-sle-to-icmp-sle.ll13 ; NOTE: this pattern is not commutative!
H A Dshift-amount-reassociation-in-bittest-with-truncation-shl.ll84 ; While 'and' is commutative, the 'trunc' *always* seems to be getting
340 ; While 'and' is commutative, the 'trunc' *always* seems to be getting
/llvm-project-15.0.7/llvm/test/CodeGen/Mips/compactbranches/
H A Dbeqc-bnec-register-constraint.ll7 ; Cases where $rs > $rt can have the operands swapped as ==,!= are commutative.
/llvm-project-15.0.7/llvm/test/CodeGen/NVPTX/
H A Ddivrem-combine.ll24 ; (This is unfortunately order-sensitive, even though mul is commutative.)
/llvm-project-15.0.7/llvm/test/Transforms/LoadStoreVectorizer/X86/
H A Dcompare-scev-by-complexity.ll6 ; operands of commutative and associative SCEV operators.
/llvm-project-15.0.7/llvm/lib/Target/VE/
H A DVVPInstrInfo.td126 // Binary operator commutative pattern.
/llvm-project-15.0.7/llvm/test/Transforms/VectorCombine/X86/
H A Dshuffle.ll184 ; For commutative instructions, common operand may be swapped.
199 ; For commutative instructions, common operand may be swapped.
214 ; non-commutative binop, but common op0
/llvm-project-15.0.7/llvm/lib/Target/AMDGPU/
H A DAMDGPUInstrInfo.td137 // This is not commutative because this gives the second operand:
153 // FIXME: TableGen doesn't like commutative instructions with more
/llvm-project-15.0.7/llvm/lib/Transforms/Scalar/
H A DGVN.cpp126 bool commutative = false; member
331 e.commutative = true; in createExpr()
342 e.commutative = true; in createExpr()
368 e.commutative = true; in createCmpExpr()
2186 if (Exp.commutative) { in phiTranslateImpl()
/llvm-project-15.0.7/mlir/docs/Rationale/
H A DRationaleGenericDAGRewriter.md147 * The IR being matched often has identities (e.g. when matching commutative
196 * It automates matching across identities, like commutative patterns.

123