diff options
| -rw-r--r-- | llvm/test/Transforms/InstCombine/AArch64/sve-intrinsic-opts-reinterpret.ll | 208 |
1 files changed, 141 insertions, 67 deletions
diff --git a/llvm/test/Transforms/InstCombine/AArch64/sve-intrinsic-opts-reinterpret.ll b/llvm/test/Transforms/InstCombine/AArch64/sve-intrinsic-opts-reinterpret.ll index 0748792..868a5d7 100644 --- a/llvm/test/Transforms/InstCombine/AArch64/sve-intrinsic-opts-reinterpret.ll +++ b/llvm/test/Transforms/InstCombine/AArch64/sve-intrinsic-opts-reinterpret.ll @@ -1,11 +1,13 @@ +; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 2 ; RUN: opt -S -passes=instcombine < %s | FileCheck %s target triple = "aarch64" define <vscale x 8 x i1> @reinterpret_test_h(<vscale x 8 x i1> %a) { -; CHECK-LABEL: @reinterpret_test_h( -; CHECK-NOT: convert -; CHECK: ret <vscale x 8 x i1> %a +; CHECK-LABEL: define <vscale x 8 x i1> @reinterpret_test_h +; CHECK-SAME: (<vscale x 8 x i1> [[A:%.*]]) { +; CHECK-NEXT: ret <vscale x 8 x i1> [[A]] +; %1 = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %a) %2 = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %1) ret <vscale x 8 x i1> %2 @@ -14,19 +16,22 @@ define <vscale x 8 x i1> @reinterpret_test_h(<vscale x 8 x i1> %a) { ; Reinterprets are not redundant because the second reinterpret zeros the ; lanes that don't exist within its input. define <vscale x 16 x i1> @reinterpret_test_h_rev(<vscale x 16 x i1> %a) { -; CHECK-LABEL: @reinterpret_test_h_rev( -; CHECK: %1 = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %a) -; CHECK-NEXT: %2 = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %1) -; CHECK-NEXT: ret <vscale x 16 x i1> %2 +; CHECK-LABEL: define <vscale x 16 x i1> @reinterpret_test_h_rev +; CHECK-SAME: (<vscale x 16 x i1> [[A:%.*]]) { +; CHECK-NEXT: [[TMP1:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[A]]) +; CHECK-NEXT: [[TMP2:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[TMP1]]) +; CHECK-NEXT: ret <vscale x 16 x i1> [[TMP2]] +; %1 = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %a) %2 = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %1) ret <vscale x 16 x i1> %2 } define <vscale x 4 x i1> @reinterpret_test_w(<vscale x 4 x i1> %a) { -; CHECK-LABEL: @reinterpret_test_w( -; CHECK-NOT: convert -; CHECK: ret <vscale x 4 x i1> %a +; CHECK-LABEL: define <vscale x 4 x i1> @reinterpret_test_w +; CHECK-SAME: (<vscale x 4 x i1> [[A:%.*]]) { +; CHECK-NEXT: ret <vscale x 4 x i1> [[A]] +; %1 = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %a) %2 = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %1) ret <vscale x 4 x i1> %2 @@ -35,19 +40,22 @@ define <vscale x 4 x i1> @reinterpret_test_w(<vscale x 4 x i1> %a) { ; Reinterprets are not redundant because the second reinterpret zeros the ; lanes that don't exist within its input. define <vscale x 16 x i1> @reinterpret_test_w_rev(<vscale x 16 x i1> %a) { -; CHECK-LABEL: @reinterpret_test_w_rev( -; CHECK: %1 = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %a) -; CHECK-NEXT: %2 = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %1) -; CHECK-NEXT: ret <vscale x 16 x i1> %2 +; CHECK-LABEL: define <vscale x 16 x i1> @reinterpret_test_w_rev +; CHECK-SAME: (<vscale x 16 x i1> [[A:%.*]]) { +; CHECK-NEXT: [[TMP1:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[A]]) +; CHECK-NEXT: [[TMP2:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP1]]) +; CHECK-NEXT: ret <vscale x 16 x i1> [[TMP2]] +; %1 = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %a) %2 = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %1) ret <vscale x 16 x i1> %2 } define <vscale x 2 x i1> @reinterpret_test_d(<vscale x 2 x i1> %a) { -; CHECK-LABEL: @reinterpret_test_d( -; CHECK-NOT: convert -; CHECK: ret <vscale x 2 x i1> %a +; CHECK-LABEL: define <vscale x 2 x i1> @reinterpret_test_d +; CHECK-SAME: (<vscale x 2 x i1> [[A:%.*]]) { +; CHECK-NEXT: ret <vscale x 2 x i1> [[A]] +; %1 = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %a) %2 = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %1) ret <vscale x 2 x i1> %2 @@ -56,18 +64,22 @@ define <vscale x 2 x i1> @reinterpret_test_d(<vscale x 2 x i1> %a) { ; Reinterprets are not redundant because the second reinterpret zeros the ; lanes that don't exist within its input. define <vscale x 16 x i1> @reinterpret_test_d_rev(<vscale x 16 x i1> %a) { -; CHECK-LABEL: @reinterpret_test_d_rev( -; CHECK: %1 = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %a) -; CHECK-NEXT: %2 = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %1) -; CHECK-NEXT: ret <vscale x 16 x i1> %2 +; CHECK-LABEL: define <vscale x 16 x i1> @reinterpret_test_d_rev +; CHECK-SAME: (<vscale x 16 x i1> [[A:%.*]]) { +; CHECK-NEXT: [[TMP1:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[A]]) +; CHECK-NEXT: [[TMP2:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[TMP1]]) +; CHECK-NEXT: ret <vscale x 16 x i1> [[TMP2]] +; %1 = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %a) %2 = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %1) ret <vscale x 16 x i1> %2 } define <vscale x 2 x i1> @reinterpret_test_full_chain(<vscale x 2 x i1> %a) { -; CHECK-LABEL: @reinterpret_test_full_chain( -; CHECK: ret <vscale x 2 x i1> %a +; CHECK-LABEL: define <vscale x 2 x i1> @reinterpret_test_full_chain +; CHECK-SAME: (<vscale x 2 x i1> [[A:%.*]]) { +; CHECK-NEXT: ret <vscale x 2 x i1> [[A]] +; %1 = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %a) %2 = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %1) %3 = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %2) @@ -80,10 +92,12 @@ define <vscale x 2 x i1> @reinterpret_test_full_chain(<vscale x 2 x i1> %a) { ; The last two reinterprets are not necessary, since they are doing the same ; work as the first two. define <vscale x 4 x i1> @reinterpret_test_partial_chain(<vscale x 2 x i1> %a) { -; CHECK-LABEL: @reinterpret_test_partial_chain( -; CHECK: %1 = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %a) -; CHECK-NEXT: %2 = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %1) -; CHECK-NEXT: ret <vscale x 4 x i1> %2 +; CHECK-LABEL: define <vscale x 4 x i1> @reinterpret_test_partial_chain +; CHECK-SAME: (<vscale x 2 x i1> [[A:%.*]]) { +; CHECK-NEXT: [[TMP1:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[A]]) +; CHECK-NEXT: [[TMP2:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[TMP1]]) +; CHECK-NEXT: ret <vscale x 4 x i1> [[TMP2]] +; %1 = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %a) %2 = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %1) %3 = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %2) @@ -94,12 +108,14 @@ define <vscale x 4 x i1> @reinterpret_test_partial_chain(<vscale x 2 x i1> %a) { ; The chain cannot be reduced because of the second reinterpret, which causes ; zeroing. define <vscale x 8 x i1> @reinterpret_test_irreducible_chain(<vscale x 8 x i1> %a) { -; CHECK-LABEL: @reinterpret_test_irreducible_chain( -; CHECK: %1 = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %a) -; CHECK-NEXT: %2 = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %1) -; CHECK-NEXT: %3 = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %2) -; CHECK-NEXT: %4 = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %3) -; CHECK-NEXT: ret <vscale x 8 x i1> %4 +; CHECK-LABEL: define <vscale x 8 x i1> @reinterpret_test_irreducible_chain +; CHECK-SAME: (<vscale x 8 x i1> [[A:%.*]]) { +; CHECK-NEXT: [[TMP1:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[A]]) +; CHECK-NEXT: [[TMP2:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[TMP1]]) +; CHECK-NEXT: [[TMP3:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[TMP2]]) +; CHECK-NEXT: [[TMP4:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[TMP3]]) +; CHECK-NEXT: ret <vscale x 8 x i1> [[TMP4]] +; %1 = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %a) %2 = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %1) %3 = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %2) @@ -110,10 +126,12 @@ define <vscale x 8 x i1> @reinterpret_test_irreducible_chain(<vscale x 8 x i1> % ; Here, the candidate list is larger than the number of instructions that we ; end up removing. define <vscale x 4 x i1> @reinterpret_test_keep_some_candidates(<vscale x 8 x i1> %a) { -; CHECK-LABEL: @reinterpret_test_keep_some_candidates( -; CHECK: %1 = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %a) -; CHECK-NEXT: %2 = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %1) -; CHECK-NEXT: ret <vscale x 4 x i1> %2 +; CHECK-LABEL: define <vscale x 4 x i1> @reinterpret_test_keep_some_candidates +; CHECK-SAME: (<vscale x 8 x i1> [[A:%.*]]) { +; CHECK-NEXT: [[TMP1:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> [[A]]) +; CHECK-NEXT: [[TMP2:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[TMP1]]) +; CHECK-NEXT: ret <vscale x 4 x i1> [[TMP2]] +; %1 = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %a) %2 = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %1) %3 = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %2) @@ -122,17 +140,28 @@ define <vscale x 4 x i1> @reinterpret_test_keep_some_candidates(<vscale x 8 x i1 } define <vscale x 2 x i1> @reinterpret_reductions(i32 %cond, <vscale x 2 x i1> %a, <vscale x 2 x i1> %b, <vscale x 2 x i1> %c) { -; CHECK-LABEL: reinterpret_reductions -; CHECK-NOT: convert -; CHECK-NOT: phi <vscale x 16 x i1> -; CHECK: phi <vscale x 2 x i1> [ %a, %br_phi_a ], [ %b, %br_phi_b ], [ %c, %br_phi_c ] -; CHECK-NOT: convert -; CHECK: ret +; CHECK-LABEL: define <vscale x 2 x i1> @reinterpret_reductions +; CHECK-SAME: (i32 [[COND:%.*]], <vscale x 2 x i1> [[A:%.*]], <vscale x 2 x i1> [[B:%.*]], <vscale x 2 x i1> [[C:%.*]]) { +; CHECK-NEXT: entry: +; CHECK-NEXT: switch i32 [[COND]], label [[BR_PHI_C:%.*]] [ +; CHECK-NEXT: i32 43, label [[BR_PHI_A:%.*]] +; CHECK-NEXT: i32 45, label [[BR_PHI_B:%.*]] +; CHECK-NEXT: ] +; CHECK: br_phi_a: +; CHECK-NEXT: br label [[JOIN:%.*]] +; CHECK: br_phi_b: +; CHECK-NEXT: br label [[JOIN]] +; CHECK: br_phi_c: +; CHECK-NEXT: br label [[JOIN]] +; CHECK: join: +; CHECK-NEXT: [[PG1:%.*]] = phi <vscale x 2 x i1> [ [[A]], [[BR_PHI_A]] ], [ [[B]], [[BR_PHI_B]] ], [ [[C]], [[BR_PHI_C]] ] +; CHECK-NEXT: ret <vscale x 2 x i1> [[PG1]] +; entry: switch i32 %cond, label %br_phi_c [ - i32 43, label %br_phi_a - i32 45, label %br_phi_b + i32 43, label %br_phi_a + i32 45, label %br_phi_b ] br_phi_a: @@ -156,17 +185,32 @@ join: ; No transform as the reinterprets are converting from different types (nxv2i1 & nxv4i1) ; As the incoming values to the phi must all be the same type, we cannot remove the reinterprets. define <vscale x 2 x i1> @reinterpret_reductions_1(i32 %cond, <vscale x 2 x i1> %a, <vscale x 4 x i1> %b, <vscale x 2 x i1> %c) { -; CHECK-LABEL: reinterpret_reductions_1 -; CHECK: convert -; CHECK: phi <vscale x 16 x i1> [ %a1, %br_phi_a ], [ %b1, %br_phi_b ], [ %c1, %br_phi_c ] -; CHECK-NOT: phi <vscale x 2 x i1> -; CHECK: tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg) -; CHECK: ret +; CHECK-LABEL: define <vscale x 2 x i1> @reinterpret_reductions_1 +; CHECK-SAME: (i32 [[COND:%.*]], <vscale x 2 x i1> [[A:%.*]], <vscale x 4 x i1> [[B:%.*]], <vscale x 2 x i1> [[C:%.*]]) { +; CHECK-NEXT: entry: +; CHECK-NEXT: switch i32 [[COND]], label [[BR_PHI_C:%.*]] [ +; CHECK-NEXT: i32 43, label [[BR_PHI_A:%.*]] +; CHECK-NEXT: i32 45, label [[BR_PHI_B:%.*]] +; CHECK-NEXT: ] +; CHECK: br_phi_a: +; CHECK-NEXT: [[A1:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[A]]) +; CHECK-NEXT: br label [[JOIN:%.*]] +; CHECK: br_phi_b: +; CHECK-NEXT: [[B1:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[B]]) +; CHECK-NEXT: br label [[JOIN]] +; CHECK: br_phi_c: +; CHECK-NEXT: [[C1:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[C]]) +; CHECK-NEXT: br label [[JOIN]] +; CHECK: join: +; CHECK-NEXT: [[PG:%.*]] = phi <vscale x 16 x i1> [ [[A1]], [[BR_PHI_A]] ], [ [[B1]], [[BR_PHI_B]] ], [ [[C1]], [[BR_PHI_C]] ] +; CHECK-NEXT: [[PG1:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG]]) +; CHECK-NEXT: ret <vscale x 2 x i1> [[PG1]] +; entry: switch i32 %cond, label %br_phi_c [ - i32 43, label %br_phi_a - i32 45, label %br_phi_b + i32 43, label %br_phi_a + i32 45, label %br_phi_b ] br_phi_a: @@ -190,17 +234,31 @@ join: ; No transform. Similar to the the test above, but here only two of the arguments need to ; be converted to svbool. define <vscale x 2 x i1> @reinterpret_reductions_2(i32 %cond, <vscale x 2 x i1> %a, <vscale x 16 x i1> %b, <vscale x 2 x i1> %c) { -; CHECK-LABEL: reinterpret_reductions_2 -; CHECK: convert -; CHECK: phi <vscale x 16 x i1> [ %a1, %br_phi_a ], [ %b, %br_phi_b ], [ %c1, %br_phi_c ] -; CHECK-NOT: phi <vscale x 2 x i1> -; CHECK: tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg) -; CHECK: ret +; CHECK-LABEL: define <vscale x 2 x i1> @reinterpret_reductions_2 +; CHECK-SAME: (i32 [[COND:%.*]], <vscale x 2 x i1> [[A:%.*]], <vscale x 16 x i1> [[B:%.*]], <vscale x 2 x i1> [[C:%.*]]) { +; CHECK-NEXT: entry: +; CHECK-NEXT: switch i32 [[COND]], label [[BR_PHI_C:%.*]] [ +; CHECK-NEXT: i32 43, label [[BR_PHI_A:%.*]] +; CHECK-NEXT: i32 45, label [[BR_PHI_B:%.*]] +; CHECK-NEXT: ] +; CHECK: br_phi_a: +; CHECK-NEXT: [[A1:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[A]]) +; CHECK-NEXT: br label [[JOIN:%.*]] +; CHECK: br_phi_b: +; CHECK-NEXT: br label [[JOIN]] +; CHECK: br_phi_c: +; CHECK-NEXT: [[C1:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[C]]) +; CHECK-NEXT: br label [[JOIN]] +; CHECK: join: +; CHECK-NEXT: [[PG:%.*]] = phi <vscale x 16 x i1> [ [[A1]], [[BR_PHI_A]] ], [ [[B]], [[BR_PHI_B]] ], [ [[C1]], [[BR_PHI_C]] ] +; CHECK-NEXT: [[PG1:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG]]) +; CHECK-NEXT: ret <vscale x 2 x i1> [[PG1]] +; entry: switch i32 %cond, label %br_phi_c [ - i32 43, label %br_phi_a - i32 45, label %br_phi_b + i32 43, label %br_phi_a + i32 45, label %br_phi_b ] br_phi_a: @@ -223,15 +281,31 @@ join: ; Similar to reinterpret_reductions but the reinterprets remain because the ; original phi cannot be removed (i.e. prefer reinterprets over multiple phis). define <vscale x 16 x i1> @reinterpret_reductions3(i32 %cond, <vscale x 2 x i1> %a, <vscale x 2 x i1> %b, <vscale x 2 x i1> %c) { -; CHECK-LABEL: reinterpret_reductions3 -; CHECK: phi <vscale x 16 x i1> [ %a1, %br_phi_a ], [ %b1, %br_phi_b ], [ %c1, %br_phi_c ] -; CHECK-NOT: phi <vscale x 2 x i1> -; CHECK: ret <vscale x 16 x i1> %pg +; CHECK-LABEL: define <vscale x 16 x i1> @reinterpret_reductions3 +; CHECK-SAME: (i32 [[COND:%.*]], <vscale x 2 x i1> [[A:%.*]], <vscale x 2 x i1> [[B:%.*]], <vscale x 2 x i1> [[C:%.*]]) { +; CHECK-NEXT: entry: +; CHECK-NEXT: switch i32 [[COND]], label [[BR_PHI_C:%.*]] [ +; CHECK-NEXT: i32 43, label [[BR_PHI_A:%.*]] +; CHECK-NEXT: i32 45, label [[BR_PHI_B:%.*]] +; CHECK-NEXT: ] +; CHECK: br_phi_a: +; CHECK-NEXT: [[A1:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[A]]) +; CHECK-NEXT: br label [[JOIN:%.*]] +; CHECK: br_phi_b: +; CHECK-NEXT: [[B1:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[B]]) +; CHECK-NEXT: br label [[JOIN]] +; CHECK: br_phi_c: +; CHECK-NEXT: [[C1:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[C]]) +; CHECK-NEXT: br label [[JOIN]] +; CHECK: join: +; CHECK-NEXT: [[PG:%.*]] = phi <vscale x 16 x i1> [ [[A1]], [[BR_PHI_A]] ], [ [[B1]], [[BR_PHI_B]] ], [ [[C1]], [[BR_PHI_C]] ] +; CHECK-NEXT: ret <vscale x 16 x i1> [[PG]] +; entry: switch i32 %cond, label %br_phi_c [ - i32 43, label %br_phi_a - i32 45, label %br_phi_b + i32 43, label %br_phi_a + i32 45, label %br_phi_b ] br_phi_a: |