[PATCH] D101836: [LoopVectorize] Enable strict reductions when allowReordering() returns false

[David Sherwood via Phabricator via llvm-commits]

david-arm added inline comments.


================
Comment at: llvm/test/Transforms/LoopVectorize/AArch64/strict-fadd.ll:81
 define float @fadd_strict_unroll_last_val(float* noalias nocapture readonly %a, float* noalias nocapture readonly %b, i64 %n) {
-; CHECK-LABEL: @fadd_strict_unroll_last_val
-; CHECK: vector.body
-; CHECK: %[[VEC_PHI1:.*]] = phi float [ 0.000000e+00, %vector.ph ], [ %[[RDX4:.*]], %vector.body ]
-; CHECK-NOT: phi float [ 0.000000e+00, %vector.ph ], [ %[[RDX4]], %vector.body ]
-; CHECK: %[[LOAD1:.*]] = load <8 x float>, <8 x float>*
-; CHECK: %[[LOAD2:.*]] = load <8 x float>, <8 x float>*
-; CHECK: %[[LOAD3:.*]] = load <8 x float>, <8 x float>*
-; CHECK: %[[LOAD4:.*]] = load <8 x float>, <8 x float>*
-; CHECK: %[[RDX1:.*]] = call float @llvm.vector.reduce.fadd.v8f32(float %[[VEC_PHI1]], <8 x float> %[[LOAD1]])
-; CHECK: %[[RDX2:.*]] = call float @llvm.vector.reduce.fadd.v8f32(float %[[RDX1]], <8 x float> %[[LOAD2]])
-; CHECK: %[[RDX3:.*]] = call float @llvm.vector.reduce.fadd.v8f32(float %[[RDX2]], <8 x float> %[[LOAD3]])
-; CHECK: %[[RDX4]] = call float @llvm.vector.reduce.fadd.v8f32(float %[[RDX3]], <8 x float> %[[LOAD4]])
-; CHECK: for.body
-; CHECK: %[[SUM_PHI:.*]] = phi float [ %[[FADD:.*]], %for.body ], [ {{.*}}, %scalar.ph ]
-; CHECK: %[[LOAD5:.*]] = load float, float*
-; CHECK: %[[FADD]] =  fadd float %[[SUM_PHI]], %[[LOAD5]]
-; CHECK: for.cond.cleanup
-; CHECK: %[[FADD_LCSSA:.*]] = phi float [ %[[FADD]], %for.body ], [ %[[RDX4]], %middle.block ]
-; CHECK: %[[FADD_42:.*]] = fadd float %[[FADD_LCSSA]], 4.200000e+01
-; CHECK: store float %[[FADD_42]], float* %b
-; CHECK: for.end
-; CHECK: %[[SUM_LCSSA:.*]] = phi float [ %[[FADD_LCSSA]], %for.cond.cleanup ], [ 0.000000e+00, %entry ]
-; CHECK: ret float %[[SUM_LCSSA]]
+; CHECK-VF8UF2-LABEL: @fadd_strict_unroll_last_val
+; CHECK-VF8UF2: vector.body
----------------
Hi @kmclaughlin, I think maybe this is meant to be CHECK-VF8UF4?


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D101836/new/

https://reviews.llvm.org/D101836