[clang] [Clang][SME2] Fix PSEL builtin predicates (PR #77097)
Kerry McLaughlin via cfe-commits
cfe-commits at lists.llvm.org
Fri Jan 5 06:15:03 PST 2024
https://github.com/kmclaughlin-arm created https://github.com/llvm/llvm-project/pull/77097
PSEL intrinsics which return a predicate-as-counter are available
in SVE2p1 & SME2.
>From 0cea7a1c7d72493de5533815903aec868543d544 Mon Sep 17 00:00:00 2001
From: Kerry McLaughlin <kerry.mclaughlin at arm.com>
Date: Fri, 5 Jan 2024 11:44:53 +0000
Subject: [PATCH] [Clang][SME2] Fix PSEL builtin predicates
PSEL intrinsics which return a predicate-as-counter are available
in SVE2p1 & SME2.
---
clang/include/clang/Basic/arm_sve.td | 9 +-
.../acle_sve2p1_psel.c | 92 +--------------
.../acle_sve2p1_psel_svcount.c | 110 ++++++++++++++++++
3 files changed, 118 insertions(+), 93 deletions(-)
create mode 100644 clang/test/CodeGen/aarch64-sve2p1-intrinsics/acle_sve2p1_psel_svcount.c
diff --git a/clang/include/clang/Basic/arm_sve.td b/clang/include/clang/Basic/arm_sve.td
index 7f80fb0386cc77..ae1a154d63355a 100644
--- a/clang/include/clang/Basic/arm_sve.td
+++ b/clang/include/clang/Basic/arm_sve.td
@@ -1952,10 +1952,6 @@ def SVPSEL_B : SInst<"svpsel_lane_b8", "PPPm", "Pc", MergeNone, "", [IsStreamin
def SVPSEL_H : SInst<"svpsel_lane_b16", "PPPm", "Ps", MergeNone, "", [IsStreamingCompatible], []>;
def SVPSEL_S : SInst<"svpsel_lane_b32", "PPPm", "Pi", MergeNone, "", [IsStreamingCompatible], []>;
def SVPSEL_D : SInst<"svpsel_lane_b64", "PPPm", "Pl", MergeNone, "", [IsStreamingCompatible], []>;
-def SVPSEL_COUNT_ALIAS_B : SInst<"svpsel_lane_c8", "}}Pm", "Pc", MergeNone, "", [IsStreamingCompatible], []>;
-def SVPSEL_COUNT_ALIAS_H : SInst<"svpsel_lane_c16", "}}Pm", "Ps", MergeNone, "", [IsStreamingCompatible], []>;
-def SVPSEL_COUNT_ALIAS_S : SInst<"svpsel_lane_c32", "}}Pm", "Pi", MergeNone, "", [IsStreamingCompatible], []>;
-def SVPSEL_COUNT_ALIAS_D : SInst<"svpsel_lane_c64", "}}Pm", "Pl", MergeNone, "", [IsStreamingCompatible], []>;
}
// Standalone sve2.1 builtins
@@ -1979,6 +1975,11 @@ let TargetGuard = "sve2p1|sme2" in {
def SVPEXT_SINGLE : SInst<"svpext_lane_{d}", "P}i", "QcQsQiQl", MergeNone, "aarch64_sve_pext", [IsStreamingOrSVE2p1], [ImmCheck<1, ImmCheck0_3>]>;
def SVPEXT_X2 : SInst<"svpext_lane_{d}_x2", "2.P}i", "QcQsQiQl", MergeNone, "aarch64_sve_pext_x2", [IsStreamingOrSVE2p1], [ImmCheck<1, ImmCheck0_1>]>;
+def SVPSEL_COUNT_ALIAS_B : SInst<"svpsel_lane_c8", "}}Pm", "Pc", MergeNone, "", [IsStreamingCompatible], []>;
+def SVPSEL_COUNT_ALIAS_H : SInst<"svpsel_lane_c16", "}}Pm", "Ps", MergeNone, "", [IsStreamingCompatible], []>;
+def SVPSEL_COUNT_ALIAS_S : SInst<"svpsel_lane_c32", "}}Pm", "Pi", MergeNone, "", [IsStreamingCompatible], []>;
+def SVPSEL_COUNT_ALIAS_D : SInst<"svpsel_lane_c64", "}}Pm", "Pl", MergeNone, "", [IsStreamingCompatible], []>;
+
def SVWHILEGE_COUNT : SInst<"svwhilege_{d}[_{1}]", "}lli", "QcQsQiQl", MergeNone, "aarch64_sve_whilege_{d}", [IsOverloadNone, IsStreamingOrSVE2p1], [ImmCheck<2, ImmCheck2_4_Mul2>]>;
def SVWHILEGT_COUNT : SInst<"svwhilegt_{d}[_{1}]", "}lli", "QcQsQiQl", MergeNone, "aarch64_sve_whilegt_{d}", [IsOverloadNone, IsStreamingOrSVE2p1], [ImmCheck<2, ImmCheck2_4_Mul2>]>;
def SVWHILELE_COUNT : SInst<"svwhilele_{d}[_{1}]", "}lli", "QcQsQiQl", MergeNone, "aarch64_sve_whilele_{d}", [IsOverloadNone, IsStreamingOrSVE2p1], [ImmCheck<2, ImmCheck2_4_Mul2>]>;
diff --git a/clang/test/CodeGen/aarch64-sve2p1-intrinsics/acle_sve2p1_psel.c b/clang/test/CodeGen/aarch64-sve2p1-intrinsics/acle_sve2p1_psel.c
index de3f6a9a57bfeb..7cf3d94ccd55d4 100644
--- a/clang/test/CodeGen/aarch64-sve2p1-intrinsics/acle_sve2p1_psel.c
+++ b/clang/test/CodeGen/aarch64-sve2p1-intrinsics/acle_sve2p1_psel.c
@@ -11,13 +11,13 @@
// RUN: -target-feature +sve2p1 -S -O1 -Werror -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
// RUN: %clang_cc1 -fclang-abi-compat=latest -triple aarch64-none-linux-gnu -target-feature +sve2p1 -S -disable-O0-optnone -Werror -Wall -o /dev/null %s
// RUN: %clang_cc1 -triple aarch64-none-linux-gnu \
-// RUN: -target-feature +sme2 -S -DTEST_SME2 -O1 -Werror -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
-// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sme2 -S -DTEST_SME2 -disable-O0-optnone -Werror -Wall -o /dev/null %s
+// RUN: -target-feature +sme -S -DTEST_SME -O1 -Werror -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sme -S -DTEST_SME -disable-O0-optnone -Werror -Wall -o /dev/null %s
// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2p1 -S -disable-O0-optnone -Werror -Wall -o /dev/null %s
#include <arm_sve.h>
-#ifndef TEST_SME2
+#ifndef TEST_SME
#define ATTR
#else
#define ATTR __arm_streaming_compatible
@@ -92,89 +92,3 @@ svbool_t test_svpsel_lane_b32(svbool_t p1, svbool_t p2, uint32_t idx) ATTR {
svbool_t test_svpsel_lane_b64(svbool_t p1, svbool_t p2, uint32_t idx) ATTR {
return svpsel_lane_b64(p1, p2, idx + 1);
}
-
-// CHECK-LABEL: @test_svpsel_lane_c8(
-// CHECK-NEXT: entry:
-// CHECK-NEXT: [[ADD:%.*]] = add i32 [[IDX:%.*]], 15
-// CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.taarch64.svcountt(target("aarch64.svcount") [[P1:%.*]])
-// CHECK-NEXT: [[TMP1:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.psel.nxv16i1(<vscale x 16 x i1> [[TMP0]], <vscale x 16 x i1> [[P2:%.*]], i32 [[ADD]])
-// CHECK-NEXT: [[TMP2:%.*]] = tail call target("aarch64.svcount") @llvm.aarch64.sve.convert.from.svbool.taarch64.svcountt(<vscale x 16 x i1> [[TMP1]])
-// CHECK-NEXT: ret target("aarch64.svcount") [[TMP2]]
-//
-// CPP-CHECK-LABEL: @_Z19test_svpsel_lane_c8u11__SVCount_tu10__SVBool_tj(
-// CPP-CHECK-NEXT: entry:
-// CPP-CHECK-NEXT: [[ADD:%.*]] = add i32 [[IDX:%.*]], 15
-// CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.taarch64.svcountt(target("aarch64.svcount") [[P1:%.*]])
-// CPP-CHECK-NEXT: [[TMP1:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.psel.nxv16i1(<vscale x 16 x i1> [[TMP0]], <vscale x 16 x i1> [[P2:%.*]], i32 [[ADD]])
-// CPP-CHECK-NEXT: [[TMP2:%.*]] = tail call target("aarch64.svcount") @llvm.aarch64.sve.convert.from.svbool.taarch64.svcountt(<vscale x 16 x i1> [[TMP1]])
-// CPP-CHECK-NEXT: ret target("aarch64.svcount") [[TMP2]]
-//
-svcount_t test_svpsel_lane_c8(svcount_t p1, svbool_t p2, uint32_t idx) ATTR {
- return svpsel_lane_c8(p1, p2, idx + 15);
-}
-
-// CHECK-LABEL: @test_svpsel_lane_c16(
-// CHECK-NEXT: entry:
-// CHECK-NEXT: [[ADD:%.*]] = add i32 [[IDX:%.*]], 7
-// CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.taarch64.svcountt(target("aarch64.svcount") [[P1:%.*]])
-// CHECK-NEXT: [[TMP1:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[P2:%.*]])
-// CHECK-NEXT: [[TMP2:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.psel.nxv8i1(<vscale x 16 x i1> [[TMP0]], <vscale x 8 x i1> [[TMP1]], i32 [[ADD]])
-// CHECK-NEXT: [[TMP3:%.*]] = tail call target("aarch64.svcount") @llvm.aarch64.sve.convert.from.svbool.taarch64.svcountt(<vscale x 16 x i1> [[TMP2]])
-// CHECK-NEXT: ret target("aarch64.svcount") [[TMP3]]
-//
-// CPP-CHECK-LABEL: @_Z20test_svpsel_lane_c16u11__SVCount_tu10__SVBool_tj(
-// CPP-CHECK-NEXT: entry:
-// CPP-CHECK-NEXT: [[ADD:%.*]] = add i32 [[IDX:%.*]], 7
-// CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.taarch64.svcountt(target("aarch64.svcount") [[P1:%.*]])
-// CPP-CHECK-NEXT: [[TMP1:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[P2:%.*]])
-// CPP-CHECK-NEXT: [[TMP2:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.psel.nxv8i1(<vscale x 16 x i1> [[TMP0]], <vscale x 8 x i1> [[TMP1]], i32 [[ADD]])
-// CPP-CHECK-NEXT: [[TMP3:%.*]] = tail call target("aarch64.svcount") @llvm.aarch64.sve.convert.from.svbool.taarch64.svcountt(<vscale x 16 x i1> [[TMP2]])
-// CPP-CHECK-NEXT: ret target("aarch64.svcount") [[TMP3]]
-//
-svcount_t test_svpsel_lane_c16(svcount_t p1, svbool_t p2, uint32_t idx) ATTR {
- return svpsel_lane_c16(p1, p2, idx + 7);
-}
-
-// CHECK-LABEL: @test_svpsel_lane_c32(
-// CHECK-NEXT: entry:
-// CHECK-NEXT: [[ADD:%.*]] = add i32 [[IDX:%.*]], 3
-// CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.taarch64.svcountt(target("aarch64.svcount") [[P1:%.*]])
-// CHECK-NEXT: [[TMP1:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[P2:%.*]])
-// CHECK-NEXT: [[TMP2:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.psel.nxv4i1(<vscale x 16 x i1> [[TMP0]], <vscale x 4 x i1> [[TMP1]], i32 [[ADD]])
-// CHECK-NEXT: [[TMP3:%.*]] = tail call target("aarch64.svcount") @llvm.aarch64.sve.convert.from.svbool.taarch64.svcountt(<vscale x 16 x i1> [[TMP2]])
-// CHECK-NEXT: ret target("aarch64.svcount") [[TMP3]]
-//
-// CPP-CHECK-LABEL: @_Z20test_svpsel_lane_c32u11__SVCount_tu10__SVBool_tj(
-// CPP-CHECK-NEXT: entry:
-// CPP-CHECK-NEXT: [[ADD:%.*]] = add i32 [[IDX:%.*]], 3
-// CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.taarch64.svcountt(target("aarch64.svcount") [[P1:%.*]])
-// CPP-CHECK-NEXT: [[TMP1:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[P2:%.*]])
-// CPP-CHECK-NEXT: [[TMP2:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.psel.nxv4i1(<vscale x 16 x i1> [[TMP0]], <vscale x 4 x i1> [[TMP1]], i32 [[ADD]])
-// CPP-CHECK-NEXT: [[TMP3:%.*]] = tail call target("aarch64.svcount") @llvm.aarch64.sve.convert.from.svbool.taarch64.svcountt(<vscale x 16 x i1> [[TMP2]])
-// CPP-CHECK-NEXT: ret target("aarch64.svcount") [[TMP3]]
-//
-svcount_t test_svpsel_lane_c32(svcount_t p1, svbool_t p2, uint32_t idx) ATTR {
- return svpsel_lane_c32(p1, p2, idx + 3);
-}
-
-// CHECK-LABEL: @test_svpsel_lane_c64(
-// CHECK-NEXT: entry:
-// CHECK-NEXT: [[ADD:%.*]] = add i32 [[IDX:%.*]], 1
-// CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.taarch64.svcountt(target("aarch64.svcount") [[P1:%.*]])
-// CHECK-NEXT: [[TMP1:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[P2:%.*]])
-// CHECK-NEXT: [[TMP2:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.psel.nxv2i1(<vscale x 16 x i1> [[TMP0]], <vscale x 2 x i1> [[TMP1]], i32 [[ADD]])
-// CHECK-NEXT: [[TMP3:%.*]] = tail call target("aarch64.svcount") @llvm.aarch64.sve.convert.from.svbool.taarch64.svcountt(<vscale x 16 x i1> [[TMP2]])
-// CHECK-NEXT: ret target("aarch64.svcount") [[TMP3]]
-//
-// CPP-CHECK-LABEL: @_Z20test_svpsel_lane_c64u11__SVCount_tu10__SVBool_tj(
-// CPP-CHECK-NEXT: entry:
-// CPP-CHECK-NEXT: [[ADD:%.*]] = add i32 [[IDX:%.*]], 1
-// CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.taarch64.svcountt(target("aarch64.svcount") [[P1:%.*]])
-// CPP-CHECK-NEXT: [[TMP1:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[P2:%.*]])
-// CPP-CHECK-NEXT: [[TMP2:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.psel.nxv2i1(<vscale x 16 x i1> [[TMP0]], <vscale x 2 x i1> [[TMP1]], i32 [[ADD]])
-// CPP-CHECK-NEXT: [[TMP3:%.*]] = tail call target("aarch64.svcount") @llvm.aarch64.sve.convert.from.svbool.taarch64.svcountt(<vscale x 16 x i1> [[TMP2]])
-// CPP-CHECK-NEXT: ret target("aarch64.svcount") [[TMP3]]
-//
-svcount_t test_svpsel_lane_c64(svcount_t p1, svbool_t p2, uint32_t idx) ATTR {
- return svpsel_lane_c64(p1, p2, idx + 1);
-}
diff --git a/clang/test/CodeGen/aarch64-sve2p1-intrinsics/acle_sve2p1_psel_svcount.c b/clang/test/CodeGen/aarch64-sve2p1-intrinsics/acle_sve2p1_psel_svcount.c
new file mode 100644
index 00000000000000..b7280a79d9d9ea
--- /dev/null
+++ b/clang/test/CodeGen/aarch64-sve2p1-intrinsics/acle_sve2p1_psel_svcount.c
@@ -0,0 +1,110 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
+// REQUIRES: aarch64-registered-target
+// RUN: %clang_cc1 -fclang-abi-compat=latest -triple aarch64-none-linux-gnu \
+// RUN: -target-feature +sve2p1 -S -O1 -Werror -emit-llvm -o - %s | FileCheck %s
+// RUN: %clang_cc1 -fclang-abi-compat=latest -triple aarch64-none-linux-gnu \
+// RUN: -target-feature +sve2p1 -S -O1 -Werror -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -fclang-abi-compat=latest -triple aarch64-none-linux-gnu -target-feature +sve2p1 -S -disable-O0-optnone -Werror -Wall -o /dev/null %s
+// RUN: %clang_cc1 -fclang-abi-compat=latest -triple aarch64-none-linux-gnu \
+// RUN: -target-feature +sve2p1 -S -O1 -Werror -emit-llvm -o - %s | FileCheck %s
+// RUN: %clang_cc1 -fclang-abi-compat=latest -triple aarch64-none-linux-gnu \
+// RUN: -target-feature +sve2p1 -S -O1 -Werror -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -fclang-abi-compat=latest -triple aarch64-none-linux-gnu -target-feature +sve2p1 -S -disable-O0-optnone -Werror -Wall -o /dev/null %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu \
+// RUN: -target-feature +sme2 -S -DTEST_SME2 -O1 -Werror -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sme2 -S -DTEST_SME2 -disable-O0-optnone -Werror -Wall -o /dev/null %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2p1 -S -disable-O0-optnone -Werror -Wall -o /dev/null %s
+
+#include <arm_sve.h>
+
+#ifndef TEST_SME2
+#define ATTR
+#else
+#define ATTR __arm_streaming_compatible
+#endif
+
+// CHECK-LABEL: @test_svpsel_lane_c8(
+// CHECK-NEXT: entry:
+// CHECK-NEXT: [[ADD:%.*]] = add i32 [[IDX:%.*]], 15
+// CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.taarch64.svcountt(target("aarch64.svcount") [[P1:%.*]])
+// CHECK-NEXT: [[TMP1:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.psel.nxv16i1(<vscale x 16 x i1> [[TMP0]], <vscale x 16 x i1> [[P2:%.*]], i32 [[ADD]])
+// CHECK-NEXT: [[TMP2:%.*]] = tail call target("aarch64.svcount") @llvm.aarch64.sve.convert.from.svbool.taarch64.svcountt(<vscale x 16 x i1> [[TMP1]])
+// CHECK-NEXT: ret target("aarch64.svcount") [[TMP2]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svpsel_lane_c8u11__SVCount_tu10__SVBool_tj(
+// CPP-CHECK-NEXT: entry:
+// CPP-CHECK-NEXT: [[ADD:%.*]] = add i32 [[IDX:%.*]], 15
+// CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.taarch64.svcountt(target("aarch64.svcount") [[P1:%.*]])
+// CPP-CHECK-NEXT: [[TMP1:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.psel.nxv16i1(<vscale x 16 x i1> [[TMP0]], <vscale x 16 x i1> [[P2:%.*]], i32 [[ADD]])
+// CPP-CHECK-NEXT: [[TMP2:%.*]] = tail call target("aarch64.svcount") @llvm.aarch64.sve.convert.from.svbool.taarch64.svcountt(<vscale x 16 x i1> [[TMP1]])
+// CPP-CHECK-NEXT: ret target("aarch64.svcount") [[TMP2]]
+//
+svcount_t test_svpsel_lane_c8(svcount_t p1, svbool_t p2, uint32_t idx) ATTR {
+ return svpsel_lane_c8(p1, p2, idx + 15);
+}
+
+// CHECK-LABEL: @test_svpsel_lane_c16(
+// CHECK-NEXT: entry:
+// CHECK-NEXT: [[ADD:%.*]] = add i32 [[IDX:%.*]], 7
+// CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.taarch64.svcountt(target("aarch64.svcount") [[P1:%.*]])
+// CHECK-NEXT: [[TMP1:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[P2:%.*]])
+// CHECK-NEXT: [[TMP2:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.psel.nxv8i1(<vscale x 16 x i1> [[TMP0]], <vscale x 8 x i1> [[TMP1]], i32 [[ADD]])
+// CHECK-NEXT: [[TMP3:%.*]] = tail call target("aarch64.svcount") @llvm.aarch64.sve.convert.from.svbool.taarch64.svcountt(<vscale x 16 x i1> [[TMP2]])
+// CHECK-NEXT: ret target("aarch64.svcount") [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svpsel_lane_c16u11__SVCount_tu10__SVBool_tj(
+// CPP-CHECK-NEXT: entry:
+// CPP-CHECK-NEXT: [[ADD:%.*]] = add i32 [[IDX:%.*]], 7
+// CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.taarch64.svcountt(target("aarch64.svcount") [[P1:%.*]])
+// CPP-CHECK-NEXT: [[TMP1:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[P2:%.*]])
+// CPP-CHECK-NEXT: [[TMP2:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.psel.nxv8i1(<vscale x 16 x i1> [[TMP0]], <vscale x 8 x i1> [[TMP1]], i32 [[ADD]])
+// CPP-CHECK-NEXT: [[TMP3:%.*]] = tail call target("aarch64.svcount") @llvm.aarch64.sve.convert.from.svbool.taarch64.svcountt(<vscale x 16 x i1> [[TMP2]])
+// CPP-CHECK-NEXT: ret target("aarch64.svcount") [[TMP3]]
+//
+svcount_t test_svpsel_lane_c16(svcount_t p1, svbool_t p2, uint32_t idx) ATTR {
+ return svpsel_lane_c16(p1, p2, idx + 7);
+}
+
+// CHECK-LABEL: @test_svpsel_lane_c32(
+// CHECK-NEXT: entry:
+// CHECK-NEXT: [[ADD:%.*]] = add i32 [[IDX:%.*]], 3
+// CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.taarch64.svcountt(target("aarch64.svcount") [[P1:%.*]])
+// CHECK-NEXT: [[TMP1:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[P2:%.*]])
+// CHECK-NEXT: [[TMP2:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.psel.nxv4i1(<vscale x 16 x i1> [[TMP0]], <vscale x 4 x i1> [[TMP1]], i32 [[ADD]])
+// CHECK-NEXT: [[TMP3:%.*]] = tail call target("aarch64.svcount") @llvm.aarch64.sve.convert.from.svbool.taarch64.svcountt(<vscale x 16 x i1> [[TMP2]])
+// CHECK-NEXT: ret target("aarch64.svcount") [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svpsel_lane_c32u11__SVCount_tu10__SVBool_tj(
+// CPP-CHECK-NEXT: entry:
+// CPP-CHECK-NEXT: [[ADD:%.*]] = add i32 [[IDX:%.*]], 3
+// CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.taarch64.svcountt(target("aarch64.svcount") [[P1:%.*]])
+// CPP-CHECK-NEXT: [[TMP1:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[P2:%.*]])
+// CPP-CHECK-NEXT: [[TMP2:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.psel.nxv4i1(<vscale x 16 x i1> [[TMP0]], <vscale x 4 x i1> [[TMP1]], i32 [[ADD]])
+// CPP-CHECK-NEXT: [[TMP3:%.*]] = tail call target("aarch64.svcount") @llvm.aarch64.sve.convert.from.svbool.taarch64.svcountt(<vscale x 16 x i1> [[TMP2]])
+// CPP-CHECK-NEXT: ret target("aarch64.svcount") [[TMP3]]
+//
+svcount_t test_svpsel_lane_c32(svcount_t p1, svbool_t p2, uint32_t idx) ATTR {
+ return svpsel_lane_c32(p1, p2, idx + 3);
+}
+
+// CHECK-LABEL: @test_svpsel_lane_c64(
+// CHECK-NEXT: entry:
+// CHECK-NEXT: [[ADD:%.*]] = add i32 [[IDX:%.*]], 1
+// CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.taarch64.svcountt(target("aarch64.svcount") [[P1:%.*]])
+// CHECK-NEXT: [[TMP1:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[P2:%.*]])
+// CHECK-NEXT: [[TMP2:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.psel.nxv2i1(<vscale x 16 x i1> [[TMP0]], <vscale x 2 x i1> [[TMP1]], i32 [[ADD]])
+// CHECK-NEXT: [[TMP3:%.*]] = tail call target("aarch64.svcount") @llvm.aarch64.sve.convert.from.svbool.taarch64.svcountt(<vscale x 16 x i1> [[TMP2]])
+// CHECK-NEXT: ret target("aarch64.svcount") [[TMP3]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svpsel_lane_c64u11__SVCount_tu10__SVBool_tj(
+// CPP-CHECK-NEXT: entry:
+// CPP-CHECK-NEXT: [[ADD:%.*]] = add i32 [[IDX:%.*]], 1
+// CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.taarch64.svcountt(target("aarch64.svcount") [[P1:%.*]])
+// CPP-CHECK-NEXT: [[TMP1:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[P2:%.*]])
+// CPP-CHECK-NEXT: [[TMP2:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.psel.nxv2i1(<vscale x 16 x i1> [[TMP0]], <vscale x 2 x i1> [[TMP1]], i32 [[ADD]])
+// CPP-CHECK-NEXT: [[TMP3:%.*]] = tail call target("aarch64.svcount") @llvm.aarch64.sve.convert.from.svbool.taarch64.svcountt(<vscale x 16 x i1> [[TMP2]])
+// CPP-CHECK-NEXT: ret target("aarch64.svcount") [[TMP3]]
+//
+svcount_t test_svpsel_lane_c64(svcount_t p1, svbool_t p2, uint32_t idx) ATTR {
+ return svpsel_lane_c64(p1, p2, idx + 1);
+}
More information about the cfe-commits
mailing list