[clang] db79974 - [SveEmitter] Add builtins for svld1rq
Sander de Smalen via cfe-commits
cfe-commits at lists.llvm.org
Fri Apr 24 03:21:00 PDT 2020
Author: Sander de Smalen
Date: 2020-04-24T11:10:28+01:00
New Revision: db7997472bbd01645f058b23a4ca9e991d5b8144
URL: https://github.com/llvm/llvm-project/commit/db7997472bbd01645f058b23a4ca9e991d5b8144
DIFF: https://github.com/llvm/llvm-project/commit/db7997472bbd01645f058b23a4ca9e991d5b8144.diff
LOG: [SveEmitter] Add builtins for svld1rq
Reviewed By: efriedma
Tags: #clang
Differential Revision: https://reviews.llvm.org/D78748
Added:
clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1rq.c
Modified:
clang/include/clang/Basic/arm_sve.td
Removed:
################################################################################
diff --git a/clang/include/clang/Basic/arm_sve.td b/clang/include/clang/Basic/arm_sve.td
index f3dfe2c8bb1d..4b69cdcb6468 100644
--- a/clang/include/clang/Basic/arm_sve.td
+++ b/clang/include/clang/Basic/arm_sve.td
@@ -444,6 +444,8 @@ def SVLDNT1 : MInst<"svldnt1[_{2}]", "dPc", "csilUcUsUiUlhfd", [IsLoad], MemEltT
// Load one vector, unextended load, non-temporal (scalar base, VL displacement)
def SVLDNT1_VNUM : MInst<"svldnt1_vnum[_{2}]", "dPcl", "csilUcUsUiUlhfd", [IsLoad], MemEltTyDefault, "aarch64_sve_ldnt1">;
+// Load one quadword and replicate (scalar base)
+def SVLD1RQ : SInst<"svld1rq[_{2}]", "dPc", "csilUcUsUiUlhfd", MergeNone, "aarch64_sve_ld1rq">;
////////////////////////////////////////////////////////////////////////////////
// Stores
diff --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1rq.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1rq.c
new file mode 100644
index 000000000000..f0a3d5c1a62a
--- /dev/null
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1rq.c
@@ -0,0 +1,108 @@
+// RUN: %clang_cc1 -D__ARM_FEATURE_SVE -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
+// RUN: %clang_cc1 -D__ARM_FEATURE_SVE -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
+
+#include <arm_sve.h>
+
+#ifdef SVE_OVERLOADED_FORMS
+// A simple used,unused... macro, long enough to represent any SVE builtin.
+#define SVE_ACLE_FUNC(A1,A2_UNUSED,A3,A4_UNUSED) A1##A3
+#else
+#define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
+#endif
+
+svint8_t test_svld1rq_s8(svbool_t pg, const int8_t *base)
+{
+ // CHECK-LABEL: test_svld1rq_s8
+ // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ld1rq.nxv16i8(<vscale x 16 x i1> %pg, i8* %base)
+ // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
+ return SVE_ACLE_FUNC(svld1rq,_s8,,)(pg, base);
+}
+
+svint16_t test_svld1rq_s16(svbool_t pg, const int16_t *base)
+{
+ // CHECK-LABEL: test_svld1rq_s16
+ // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
+ // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ld1rq.nxv8i16(<vscale x 8 x i1> %[[PG]], i16* %base)
+ // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
+ return SVE_ACLE_FUNC(svld1rq,_s16,,)(pg, base);
+}
+
+svint32_t test_svld1rq_s32(svbool_t pg, const int32_t *base)
+{
+ // CHECK-LABEL: test_svld1rq_s32
+ // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
+ // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1rq.nxv4i32(<vscale x 4 x i1> %[[PG]], i32* %base)
+ // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
+ return SVE_ACLE_FUNC(svld1rq,_s32,,)(pg, base);
+}
+
+svint64_t test_svld1rq_s64(svbool_t pg, const int64_t *base)
+{
+ // CHECK-LABEL: test_svld1rq_s64
+ // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
+ // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1rq.nxv2i64(<vscale x 2 x i1> %[[PG]], i64* %base)
+ // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
+ return SVE_ACLE_FUNC(svld1rq,_s64,,)(pg, base);
+}
+
+svuint8_t test_svld1rq_u8(svbool_t pg, const uint8_t *base)
+{
+ // CHECK-LABEL: test_svld1rq_u8
+ // CHECK: %[[INTRINSIC:.*]] = call <vscale x 16 x i8> @llvm.aarch64.sve.ld1rq.nxv16i8(<vscale x 16 x i1> %pg, i8* %base)
+ // CHECK: ret <vscale x 16 x i8> %[[INTRINSIC]]
+ return SVE_ACLE_FUNC(svld1rq,_u8,,)(pg, base);
+}
+
+svuint16_t test_svld1rq_u16(svbool_t pg, const uint16_t *base)
+{
+ // CHECK-LABEL: test_svld1rq_u16
+ // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
+ // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x i16> @llvm.aarch64.sve.ld1rq.nxv8i16(<vscale x 8 x i1> %[[PG]], i16* %base)
+ // CHECK: ret <vscale x 8 x i16> %[[INTRINSIC]]
+ return SVE_ACLE_FUNC(svld1rq,_u16,,)(pg, base);
+}
+
+svuint32_t test_svld1rq_u32(svbool_t pg, const uint32_t *base)
+{
+ // CHECK-LABEL: test_svld1rq_u32
+ // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
+ // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x i32> @llvm.aarch64.sve.ld1rq.nxv4i32(<vscale x 4 x i1> %[[PG]], i32* %base)
+ // CHECK: ret <vscale x 4 x i32> %[[INTRINSIC]]
+ return SVE_ACLE_FUNC(svld1rq,_u32,,)(pg, base);
+}
+
+svuint64_t test_svld1rq_u64(svbool_t pg, const uint64_t *base)
+{
+ // CHECK-LABEL: test_svld1rq_u64
+ // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
+ // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x i64> @llvm.aarch64.sve.ld1rq.nxv2i64(<vscale x 2 x i1> %[[PG]], i64* %base)
+ // CHECK: ret <vscale x 2 x i64> %[[INTRINSIC]]
+ return SVE_ACLE_FUNC(svld1rq,_u64,,)(pg, base);
+}
+
+svfloat16_t test_svld1rq_f16(svbool_t pg, const float16_t *base)
+{
+ // CHECK-LABEL: test_svld1rq_f16
+ // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
+ // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x half> @llvm.aarch64.sve.ld1rq.nxv8f16(<vscale x 8 x i1> %[[PG]], half* %base)
+ // CHECK: ret <vscale x 8 x half> %[[INTRINSIC]]
+ return SVE_ACLE_FUNC(svld1rq,_f16,,)(pg, base);
+}
+
+svfloat32_t test_svld1rq_f32(svbool_t pg, const float32_t *base)
+{
+ // CHECK-LABEL: test_svld1rq_f32
+ // CHECK: %[[PG:.*]] = call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %pg)
+ // CHECK: %[[INTRINSIC:.*]] = call <vscale x 4 x float> @llvm.aarch64.sve.ld1rq.nxv4f32(<vscale x 4 x i1> %[[PG]], float* %base)
+ // CHECK: ret <vscale x 4 x float> %[[INTRINSIC]]
+ return SVE_ACLE_FUNC(svld1rq,_f32,,)(pg, base);
+}
+
+svfloat64_t test_svld1rq_f64(svbool_t pg, const float64_t *base)
+{
+ // CHECK-LABEL: test_svld1rq_f64
+ // CHECK: %[[PG:.*]] = call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> %pg)
+ // CHECK: %[[INTRINSIC:.*]] = call <vscale x 2 x double> @llvm.aarch64.sve.ld1rq.nxv2f64(<vscale x 2 x i1> %[[PG]], double* %base)
+ // CHECK: ret <vscale x 2 x double> %[[INTRINSIC]]
+ return SVE_ACLE_FUNC(svld1rq,_f64,,)(pg, base);
+}
More information about the cfe-commits
mailing list