[clang] [AArch64][SME2] Fix SME2 mla/mls tests (PR #76711)
Matthew Devereau via cfe-commits
cfe-commits at lists.llvm.org
Wed Jan 10 03:46:27 PST 2024
https://github.com/MDevereau updated https://github.com/llvm/llvm-project/pull/76711
>From 908da224bd01e4758392a98ba2191185d7296c6a Mon Sep 17 00:00:00 2001
From: Matt Devereau <matthew.devereau at arm.com>
Date: Tue, 2 Jan 2024 11:36:33 +0000
Subject: [PATCH 1/4] [AArch64][SME2] Fix SME2 mla/mls tests
The ACLE defines these builtins as svmla[_single]_za32[_f32]_vg1x2,
which means the SVE_ACLE_FUNC macro should test the overloaded forms as
SVE_ACLE_FUNC(svmla,_single,_za32,_f32,_vg1x2)
https://github.com/ARM-software/acle/blob/b88cbf7e9c104100bb5016c848763171494dee44/main/acle.md?plain=1#L10170-L10205
---
.../test/CodeGen/aarch64-sme2-intrinsics/acle_sme2_mla.c | 8 ++++----
.../test/CodeGen/aarch64-sme2-intrinsics/acle_sme2_mls.c | 8 ++++----
2 files changed, 8 insertions(+), 8 deletions(-)
diff --git a/clang/test/CodeGen/aarch64-sme2-intrinsics/acle_sme2_mla.c b/clang/test/CodeGen/aarch64-sme2-intrinsics/acle_sme2_mla.c
index f52edd9888daa9..2679f9cc8dfd0c 100644
--- a/clang/test/CodeGen/aarch64-sme2-intrinsics/acle_sme2_mla.c
+++ b/clang/test/CodeGen/aarch64-sme2-intrinsics/acle_sme2_mla.c
@@ -86,7 +86,7 @@ void test_svmla4_f32(uint32_t slice_base, svfloat32x4_t zn, svfloat32x4_t zm) __
// CPP-CHECK-NEXT: ret void
//
void test_svmla_single2_f32(uint32_t slice_base, svfloat32x2_t zn, svfloat32_t zm) __arm_streaming __arm_shared_za {
- SVE_ACLE_FUNC(svmla_single_za32,,_f32,,_vg1x2)(slice_base, zn, zm);
+ SVE_ACLE_FUNC(svmla,_single,_za32,_f32,_vg1x2)(slice_base, zn, zm);
}
// CHECK-LABEL: @test_svmla_single4_f32(
@@ -108,7 +108,7 @@ void test_svmla_single2_f32(uint32_t slice_base, svfloat32x2_t zn, svfloat32_t z
// CPP-CHECK-NEXT: ret void
//
void test_svmla_single4_f32(uint32_t slice_base, svfloat32x4_t zn, svfloat32_t zm) __arm_streaming __arm_shared_za {
- SVE_ACLE_FUNC(svmla_single_za32,,_f32,,_vg1x4)(slice_base, zn, zm);
+ SVE_ACLE_FUNC(svmla,_single,_za32,_f32,_vg1x4)(slice_base, zn, zm);
}
//
@@ -224,7 +224,7 @@ void test_svmla4_f64(uint32_t slice_base, svfloat64x4_t zn, svfloat64x4_t zm) __
// CPP-CHECK-NEXT: ret void
//
void test_svmla_single2_f64(uint32_t slice_base, svfloat64x2_t zn, svfloat64_t zm) __arm_streaming __arm_shared_za {
- SVE_ACLE_FUNC(svmla_single_za64,,_f64,,_vg1x2)(slice_base, zn, zm);
+ SVE_ACLE_FUNC(svmla,_single,_za64,_f64,_vg1x2)(slice_base, zn, zm);
}
// CHECK-LABEL: @test_svmla_single4_f64(
@@ -246,7 +246,7 @@ void test_svmla_single2_f64(uint32_t slice_base, svfloat64x2_t zn, svfloat64_t z
// CPP-CHECK-NEXT: ret void
//
void test_svmla_single4_f64(uint32_t slice_base, svfloat64x4_t zn, svfloat64_t zm) __arm_streaming __arm_shared_za {
- SVE_ACLE_FUNC(svmla_single_za64,,_f64,,_vg1x4)(slice_base, zn, zm);
+ SVE_ACLE_FUNC(svmla,_single,_za64,_f64,_vg1x4)(slice_base, zn, zm);
}
//
diff --git a/clang/test/CodeGen/aarch64-sme2-intrinsics/acle_sme2_mls.c b/clang/test/CodeGen/aarch64-sme2-intrinsics/acle_sme2_mls.c
index 6830a399e91d64..bb4cbbd5308c06 100644
--- a/clang/test/CodeGen/aarch64-sme2-intrinsics/acle_sme2_mls.c
+++ b/clang/test/CodeGen/aarch64-sme2-intrinsics/acle_sme2_mls.c
@@ -86,7 +86,7 @@ void test_svmls4_f32(uint32_t slice_base, svfloat32x4_t zn, svfloat32x4_t zm) __
// CPP-CHECK-NEXT: ret void
//
void test_svmls_single2_f32(uint32_t slice_base, svfloat32x2_t zn, svfloat32_t zm) __arm_streaming __arm_shared_za {
- SVE_ACLE_FUNC(svmls_single_za32,,_f32,,_vg1x2)(slice_base, zn, zm);
+ SVE_ACLE_FUNC(svmls,_single,_za32,_f32,_vg1x2)(slice_base, zn, zm);
}
// CHECK-LABEL: @test_svmls_single4_f32(
@@ -108,7 +108,7 @@ void test_svmls_single2_f32(uint32_t slice_base, svfloat32x2_t zn, svfloat32_t z
// CPP-CHECK-NEXT: ret void
//
void test_svmls_single4_f32(uint32_t slice_base, svfloat32x4_t zn, svfloat32_t zm) __arm_streaming __arm_shared_za {
- SVE_ACLE_FUNC(svmls_single_za32,,_f32,,_vg1x4)(slice_base, zn, zm);
+ SVE_ACLE_FUNC(svmls,_single,_za32,_f32,_vg1x4)(slice_base, zn, zm);
}
//
@@ -224,7 +224,7 @@ void test_svmls4_f64(uint32_t slice_base, svfloat64x4_t zn, svfloat64x4_t zm) __
// CPP-CHECK-NEXT: ret void
//
void test_svmls_single2_f64(uint32_t slice_base, svfloat64x2_t zn, svfloat64_t zm) __arm_streaming __arm_shared_za {
- SVE_ACLE_FUNC(svmls_single_za64,,_f64,,_vg1x2)(slice_base, zn, zm);
+ SVE_ACLE_FUNC(svmls,_single,_za64,_f64,_vg1x2)(slice_base, zn, zm);
}
// CHECK-LABEL: @test_svmls_single4_f64(
@@ -246,7 +246,7 @@ void test_svmls_single2_f64(uint32_t slice_base, svfloat64x2_t zn, svfloat64_t z
// CPP-CHECK-NEXT: ret void
//
void test_svmls_single4_f64(uint32_t slice_base, svfloat64x4_t zn, svfloat64_t zm) __arm_streaming __arm_shared_za {
- SVE_ACLE_FUNC(svmls_single_za64,,_f64,,_vg1x4)(slice_base, zn, zm);
+ SVE_ACLE_FUNC(svmls,_single,_za64,_f64,_vg1x4)(slice_base, zn, zm);
}
//
>From e09efe2659db596a60b61f984648fd9fb37b2a9c Mon Sep 17 00:00:00 2001
From: Matt Devereau <matthew.devereau at arm.com>
Date: Thu, 4 Jan 2024 16:47:17 +0000
Subject: [PATCH 2/4] Update acle_sme2_mlal.c, acle_sme2_mlall.c &
acle_sme2_mlsl.c
---
.../aarch64-sme2-intrinsics/acle_sme2_mlal.c | 16 ++++----
.../aarch64-sme2-intrinsics/acle_sme2_mlall.c | 40 +++++++++----------
.../aarch64-sme2-intrinsics/acle_sme2_mlsl.c | 16 ++++----
3 files changed, 36 insertions(+), 36 deletions(-)
diff --git a/clang/test/CodeGen/aarch64-sme2-intrinsics/acle_sme2_mlal.c b/clang/test/CodeGen/aarch64-sme2-intrinsics/acle_sme2_mlal.c
index 834ade75350767..313ce4418682b2 100644
--- a/clang/test/CodeGen/aarch64-sme2-intrinsics/acle_sme2_mlal.c
+++ b/clang/test/CodeGen/aarch64-sme2-intrinsics/acle_sme2_mlal.c
@@ -311,7 +311,7 @@ void test_svmla_single1_s16(uint32_t slice_base, svint16_t zn, svint16_t zm) __a
//
void test_svmla_single2_f16(uint32_t slice_base, svfloat16x2_t zn, svfloat16_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmla_single_za32,,_f16,,_vg2x2)(slice_base, zn, zm);
+ SVE_ACLE_FUNC(svmla,_single,_za32,_f16,_vg2x2)(slice_base, zn, zm);
}
// CHECK-LABEL: @test_svmla_single2_bf16(
@@ -330,7 +330,7 @@ void test_svmla_single2_f16(uint32_t slice_base, svfloat16x2_t zn, svfloat16_t z
//
void test_svmla_single2_bf16(uint32_t slice_base, svbfloat16x2_t zn, svbfloat16_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmla_single_za32,,_bf16,,_vg2x2)(slice_base, zn, zm);
+ SVE_ACLE_FUNC(svmla,_single,_za32,_bf16,_vg2x2)(slice_base, zn, zm);
}
// CHECK-LABEL: @test_svmla_single2_u16(
@@ -349,7 +349,7 @@ void test_svmla_single2_bf16(uint32_t slice_base, svbfloat16x2_t zn, svbfloat16_
//
void test_svmla_single2_u16(uint32_t slice_base, svuint16x2_t zn, svuint16_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmla_single_za32,,_u16,,_vg2x2)(slice_base, zn, zm);
+ SVE_ACLE_FUNC(svmla,_single,_za32,_u16,_vg2x2)(slice_base, zn, zm);
}
// CHECK-LABEL: @test_svmla_single2_s16(
@@ -368,7 +368,7 @@ void test_svmla_single2_u16(uint32_t slice_base, svuint16x2_t zn, svuint16_t zm)
//
void test_svmla_single2_s16(uint32_t slice_base, svint16x2_t zn, svint16_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmla_single_za32,,_s16,,_vg2x2)(slice_base, zn, zm);
+ SVE_ACLE_FUNC(svmla,_single,_za32,_s16,_vg2x2)(slice_base, zn, zm);
}
// CHECK-LABEL: @test_svmla_single4_f16(
@@ -391,7 +391,7 @@ void test_svmla_single2_s16(uint32_t slice_base, svint16x2_t zn, svint16_t zm) _
//
void test_svmla_single4_f16(uint32_t slice_base, svfloat16x4_t zn, svfloat16_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmla_single_za32,,_f16,,_vg2x4)(slice_base, zn, zm);
+ SVE_ACLE_FUNC(svmla,_single,_za32,_f16,_vg2x4)(slice_base, zn, zm);
}
// CHECK-LABEL: @test_svmla_single4_bf16(
@@ -414,7 +414,7 @@ void test_svmla_single4_f16(uint32_t slice_base, svfloat16x4_t zn, svfloat16_t z
//
void test_svmla_single4_bf16(uint32_t slice_base, svbfloat16x4_t zn, svbfloat16_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmla_single_za32,,_bf16,,_vg2x4)(slice_base, zn, zm);
+ SVE_ACLE_FUNC(svmla,_single,_za32,_bf16,_vg2x4)(slice_base, zn, zm);
}
// CHECK-LABEL: @test_svmla_single4_u16(
@@ -437,7 +437,7 @@ void test_svmla_single4_bf16(uint32_t slice_base, svbfloat16x4_t zn, svbfloat16_
//
void test_svmla_single4_u16(uint32_t slice_base, svuint16x4_t zn, svuint16_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmla_single_za32,,_u16,,_vg2x4)(slice_base, zn, zm);
+ SVE_ACLE_FUNC(svmla,_single,_za32,_u16,_vg2x4)(slice_base, zn, zm);
}
// CHECK-LABEL: @test_svmla_single4_s16(
@@ -460,7 +460,7 @@ void test_svmla_single4_u16(uint32_t slice_base, svuint16x4_t zn, svuint16_t zm)
//
void test_svmla_single4_s16(uint32_t slice_base, svint16x4_t zn, svint16_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmla_single_za32,,_s16,,_vg2x4)(slice_base, zn, zm);
+ SVE_ACLE_FUNC(svmla,_single,_za32,_s16,_vg2x4)(slice_base, zn, zm);
}
//
diff --git a/clang/test/CodeGen/aarch64-sme2-intrinsics/acle_sme2_mlall.c b/clang/test/CodeGen/aarch64-sme2-intrinsics/acle_sme2_mlall.c
index fceb16a482600d..44e715e9581b7c 100644
--- a/clang/test/CodeGen/aarch64-sme2-intrinsics/acle_sme2_mlall.c
+++ b/clang/test/CodeGen/aarch64-sme2-intrinsics/acle_sme2_mlall.c
@@ -199,7 +199,7 @@ void test_usmlall_single_x1_u8(uint32_t slice_base, svuint8_t zn, svint8_t zm) _
//
void test_svmla_single_x2_s8(uint32_t slice_base, svint8x2_t zn, svint8_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmla_single_za32,,_s8,,_vg4x2)(slice_base, zn, zm);
+ SVE_ACLE_FUNC(svmla,_single,_za32,_s8,_vg4x2)(slice_base, zn, zm);
}
// CHECK-LABEL: @test_svmla_single_x2_s16(
@@ -218,7 +218,7 @@ void test_svmla_single_x2_s8(uint32_t slice_base, svint8x2_t zn, svint8_t zm) __
//
void test_svmla_single_x2_s16(uint32_t slice_base, svint16x2_t zn, svint16_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmla_single_za64,,_s16,,_vg4x2)(slice_base, zn, zm);
+ SVE_ACLE_FUNC(svmla,_single,_za64,_s16,_vg4x2)(slice_base, zn, zm);
}
// CHECK-LABEL: @test_svmla_single_x2_u8(
@@ -237,7 +237,7 @@ void test_svmla_single_x2_s16(uint32_t slice_base, svint16x2_t zn, svint16_t zm)
//
void test_svmla_single_x2_u8(uint32_t slice_base, svuint8x2_t zn, svuint8_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmla_single_za32,,_u8,,_vg4x2)(slice_base, zn, zm);
+ SVE_ACLE_FUNC(svmla,_single,_za32,_u8,_vg4x2)(slice_base, zn, zm);
}
// CHECK-LABEL: @test_svmla_single_x2_u16(
@@ -256,7 +256,7 @@ void test_svmla_single_x2_u8(uint32_t slice_base, svuint8x2_t zn, svuint8_t zm)
//
void test_svmla_single_x2_u16(uint32_t slice_base, svuint16x2_t zn, svuint16_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmla_single_za64,,_u16,,_vg4x2)(slice_base, zn, zm);
+ SVE_ACLE_FUNC(svmla,_single,_za64,_u16,_vg4x2)(slice_base, zn, zm);
}
// MLSL
@@ -277,7 +277,7 @@ void test_svmla_single_x2_u16(uint32_t slice_base, svuint16x2_t zn, svuint16_t z
//
void test_svmls_single_x2_s8(uint32_t slice_base, svint8x2_t zn, svint8_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmls_single_za32,,_s8,,_vg4x2)(slice_base, zn, zm);
+ SVE_ACLE_FUNC(svmls,_single,_za32,_s8,_vg4x2)(slice_base, zn, zm);
}
// CHECK-LABEL: @test_svmls_single_x2_s16(
@@ -296,7 +296,7 @@ void test_svmls_single_x2_s8(uint32_t slice_base, svint8x2_t zn, svint8_t zm) __
//
void test_svmls_single_x2_s16(uint32_t slice_base, svint16x2_t zn, svint16_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmls_single_za64,,_s16,,_vg4x2)(slice_base, zn, zm);
+ SVE_ACLE_FUNC(svmls,_single,_za64,_s16,_vg4x2)(slice_base, zn, zm);
}
// CHECK-LABEL: @test_svmls_single_x2_u8(
@@ -315,7 +315,7 @@ void test_svmls_single_x2_s16(uint32_t slice_base, svint16x2_t zn, svint16_t zm)
//
void test_svmls_single_x2_u8(uint32_t slice_base, svuint8x2_t zn, svuint8_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmls_single_za32,,_u8,,_vg4x2)(slice_base, zn, zm);
+ SVE_ACLE_FUNC(svmls,_single,_za32,_u8,_vg4x2)(slice_base, zn, zm);
}
// CHECK-LABEL: @test_svmls_single_x2_u16(
@@ -334,7 +334,7 @@ void test_svmls_single_x2_u8(uint32_t slice_base, svuint8x2_t zn, svuint8_t zm)
//
void test_svmls_single_x2_u16(uint32_t slice_base, svuint16x2_t zn, svuint16_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmls_single_za64,,_u16,,_vg4x2)(slice_base, zn, zm);
+ SVE_ACLE_FUNC(svmls,_single,_za64,_u16,_vg4x2)(slice_base, zn, zm);
}
// SUMLALL
@@ -355,7 +355,7 @@ void test_svmls_single_x2_u16(uint32_t slice_base, svuint16x2_t zn, svuint16_t z
//
void test_svsumla_single_x2_s8(uint32_t slice_base, svint8x2_t zn, svuint8_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svsumla_single_za32,,_s8,,_vg4x2)(slice_base, zn, zm);
+ SVE_ACLE_FUNC(svsumla,_single,_za32,_s8,_vg4x2)(slice_base, zn, zm);
}
// USMLALL
@@ -376,7 +376,7 @@ void test_svsumla_single_x2_s8(uint32_t slice_base, svint8x2_t zn, svuint8_t zm)
//
void test_usmlall_single_x2_u8(uint32_t slice_base, svuint8x2_t zn, svint8_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svusmla_single_za32,,_u8,,_vg4x2)(slice_base, zn, zm);
+ SVE_ACLE_FUNC(svusmla,_single,_za32,_u8,_vg4x2)(slice_base, zn, zm);
}
//
@@ -405,7 +405,7 @@ void test_usmlall_single_x2_u8(uint32_t slice_base, svuint8x2_t zn, svint8_t zm)
//
void test_svmla_single_x4_s8(uint32_t slice_base, svint8x4_t zn, svint8_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmla_single_za32,,_s8,,_vg4x4)(slice_base, zn, zm);
+ SVE_ACLE_FUNC(svmla,_single,_za32,_s8,_vg4x4)(slice_base, zn, zm);
}
// CHECK-LABEL: @test_svmla_single_x4_s16(
@@ -428,7 +428,7 @@ void test_svmla_single_x4_s8(uint32_t slice_base, svint8x4_t zn, svint8_t zm) __
//
void test_svmla_single_x4_s16(uint32_t slice_base, svint16x4_t zn, svint16_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmla_single_za64,,_s16,,_vg4x4)(slice_base, zn, zm);
+ SVE_ACLE_FUNC(svmla,_single,_za64,_s16,_vg4x4)(slice_base, zn, zm);
}
// CHECK-LABEL: @test_svmla_single_x4_u8(
@@ -451,7 +451,7 @@ void test_svmla_single_x4_s16(uint32_t slice_base, svint16x4_t zn, svint16_t zm)
//
void test_svmla_single_x4_u8(uint32_t slice_base, svuint8x4_t zn, svuint8_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmla_single_za32,,_u8,,_vg4x4)(slice_base, zn, zm);
+ SVE_ACLE_FUNC(svmla,_single,_za32,_u8,_vg4x4)(slice_base, zn, zm);
}
// CHECK-LABEL: @test_svmla_single_x4_u16(
@@ -474,7 +474,7 @@ void test_svmla_single_x4_u8(uint32_t slice_base, svuint8x4_t zn, svuint8_t zm)
//
void test_svmla_single_x4_u16(uint32_t slice_base, svuint16x4_t zn, svuint16_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmla_single_za64,,_u16,,_vg4x4)(slice_base, zn, zm);
+ SVE_ACLE_FUNC(svmla,_single,_za64,_u16,_vg4x4)(slice_base, zn, zm);
}
// MLSL
@@ -499,7 +499,7 @@ void test_svmla_single_x4_u16(uint32_t slice_base, svuint16x4_t zn, svuint16_t z
//
void test_svmls_single_x4_s8(uint32_t slice_base, svint8x4_t zn, svint8_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmls_single_za32,,_s8,,_vg4x4)(slice_base, zn, zm);
+ SVE_ACLE_FUNC(svmls,_single,_za32,_s8,_vg4x4)(slice_base, zn, zm);
}
// CHECK-LABEL: @test_svmls_single_x4_s16(
@@ -522,7 +522,7 @@ void test_svmls_single_x4_s8(uint32_t slice_base, svint8x4_t zn, svint8_t zm) __
//
void test_svmls_single_x4_s16(uint32_t slice_base, svint16x4_t zn, svint16_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmls_single_za64,,_s16,,_vg4x4)(slice_base, zn, zm);
+ SVE_ACLE_FUNC(svmls,_single,_za64,_s16,_vg4x4)(slice_base, zn, zm);
}
// CHECK-LABEL: @test_svmls_single_x4_u8(
@@ -545,7 +545,7 @@ void test_svmls_single_x4_s16(uint32_t slice_base, svint16x4_t zn, svint16_t zm)
//
void test_svmls_single_x4_u8(uint32_t slice_base, svuint8x4_t zn, svuint8_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmls_single_za32,,_u8,,_vg4x4)(slice_base, zn, zm);
+ SVE_ACLE_FUNC(svmls,_single,_za32,_u8,_vg4x4)(slice_base, zn, zm);
}
// CHECK-LABEL: @test_svmls_single_x4_u16(
@@ -568,7 +568,7 @@ void test_svmls_single_x4_u8(uint32_t slice_base, svuint8x4_t zn, svuint8_t zm)
//
void test_svmls_single_x4_u16(uint32_t slice_base, svuint16x4_t zn, svuint16_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmls_single_za64,,_u16,,_vg4x4)(slice_base, zn, zm);
+ SVE_ACLE_FUNC(svmls,_single,_za64,_u16,_vg4x4)(slice_base, zn, zm);
}
// SUMLALL
@@ -593,7 +593,7 @@ void test_svmls_single_x4_u16(uint32_t slice_base, svuint16x4_t zn, svuint16_t z
//
void test_svsumla_single_x4_s8(uint32_t slice_base, svint8x4_t zn, svuint8_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svsumla_single_za32,,_s8,,_vg4x4)(slice_base, zn, zm);
+ SVE_ACLE_FUNC(svsumla,_single,_za32,_s8,_vg4x4)(slice_base, zn, zm);
}
// USMLALL
@@ -618,7 +618,7 @@ void test_svsumla_single_x4_s8(uint32_t slice_base, svint8x4_t zn, svuint8_t zm)
//
void test_usmlall_single_x4_u8(uint32_t slice_base, svuint8x4_t zn, svint8_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svusmla_single_za32,,_u8,,_vg4x4)(slice_base, zn, zm);
+ SVE_ACLE_FUNC(svusmla,_single,_za32,_u8,_vg4x4)(slice_base, zn, zm);
}
//
diff --git a/clang/test/CodeGen/aarch64-sme2-intrinsics/acle_sme2_mlsl.c b/clang/test/CodeGen/aarch64-sme2-intrinsics/acle_sme2_mlsl.c
index 0a87d97f649c90..2c996eef9d2de4 100644
--- a/clang/test/CodeGen/aarch64-sme2-intrinsics/acle_sme2_mlsl.c
+++ b/clang/test/CodeGen/aarch64-sme2-intrinsics/acle_sme2_mlsl.c
@@ -311,7 +311,7 @@ void test_svmls_single1_s16(uint32_t slice_base, svint16_t zn, svint16_t zm) __a
//
void test_svmls_single2_f16(uint32_t slice_base, svfloat16x2_t zn, svfloat16_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmls_single_za32,,_f16,,_vg2x2)(slice_base, zn, zm);
+ SVE_ACLE_FUNC(svmls,_single,_za32,_f16,_vg2x2)(slice_base, zn, zm);
}
// CHECK-LABEL: @test_svmls_single2_bf16(
@@ -330,7 +330,7 @@ void test_svmls_single2_f16(uint32_t slice_base, svfloat16x2_t zn, svfloat16_t z
//
void test_svmls_single2_bf16(uint32_t slice_base, svbfloat16x2_t zn, svbfloat16_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmls_single_za32,,_bf16,,_vg2x2)(slice_base, zn, zm);
+ SVE_ACLE_FUNC(svmls,_single,_za32,_bf16,_vg2x2)(slice_base, zn, zm);
}
// CHECK-LABEL: @test_svmls_single2_u16(
@@ -349,7 +349,7 @@ void test_svmls_single2_bf16(uint32_t slice_base, svbfloat16x2_t zn, svbfloat16_
//
void test_svmls_single2_u16(uint32_t slice_base, svuint16x2_t zn, svuint16_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmls_single_za32,,_u16,,_vg2x2)(slice_base, zn, zm);
+ SVE_ACLE_FUNC(svmls,_single,_za32,_u16,_vg2x2)(slice_base, zn, zm);
}
// CHECK-LABEL: @test_svmls_single2_s16(
@@ -368,7 +368,7 @@ void test_svmls_single2_u16(uint32_t slice_base, svuint16x2_t zn, svuint16_t zm)
//
void test_svmls_single2_s16(uint32_t slice_base, svint16x2_t zn, svint16_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmls_single_za32,,_s16,,_vg2x2)(slice_base, zn, zm);
+ SVE_ACLE_FUNC(svmls,_single,_za32,_s16,_vg2x2)(slice_base, zn, zm);
}
// CHECK-LABEL: @test_svmls_single4_f16(
@@ -391,7 +391,7 @@ void test_svmls_single2_s16(uint32_t slice_base, svint16x2_t zn, svint16_t zm) _
//
void test_svmls_single4_f16(uint32_t slice_base, svfloat16x4_t zn, svfloat16_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmls_single_za32,,_f16,,_vg2x4)(slice_base, zn, zm);
+ SVE_ACLE_FUNC(svmls,_single,_za32,_f16,_vg2x4)(slice_base, zn, zm);
}
// CHECK-LABEL: @test_svmls_single4_bf16(
@@ -414,7 +414,7 @@ void test_svmls_single4_f16(uint32_t slice_base, svfloat16x4_t zn, svfloat16_t z
//
void test_svmls_single4_bf16(uint32_t slice_base, svbfloat16x4_t zn, svbfloat16_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmls_single_za32,,_bf16,,_vg2x4)(slice_base, zn, zm);
+ SVE_ACLE_FUNC(svmls,_single,_za32,_bf16,_vg2x4)(slice_base, zn, zm);
}
// CHECK-LABEL: @test_svmls_single4_u16(
@@ -437,7 +437,7 @@ void test_svmls_single4_bf16(uint32_t slice_base, svbfloat16x4_t zn, svbfloat16_
//
void test_svmls_single4_u16(uint32_t slice_base, svuint16x4_t zn, svuint16_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmls_single_za32,,_u16,,_vg2x4)(slice_base, zn, zm);
+ SVE_ACLE_FUNC(svmls,_single,_za32,_u16,_vg2x4)(slice_base, zn, zm);
}
// CHECK-LABEL: @test_svmls_single4_s16(
@@ -460,7 +460,7 @@ void test_svmls_single4_u16(uint32_t slice_base, svuint16x4_t zn, svuint16_t zm)
//
void test_svmls_single4_s16(uint32_t slice_base, svint16x4_t zn, svint16_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmls_single_za32,,_s16,,_vg2x4)(slice_base, zn, zm);
+ SVE_ACLE_FUNC(svmls,_single,_za32,_s16,_vg2x4)(slice_base, zn, zm);
}
//
>From fd9e0bbe2688752f1a516fb93e9684af36aa184f Mon Sep 17 00:00:00 2001
From: Matt Devereau <matthew.devereau at arm.com>
Date: Fri, 5 Jan 2024 17:14:47 +0000
Subject: [PATCH 3/4] Fix more overloaded tests in acle_sme2_mls.c
and acle_sme2_mlsl.c
---
.../aarch64-sme2-intrinsics/acle_sme2_mls.c | 16 +++++++-------
.../aarch64-sme2-intrinsics/acle_sme2_mlsl.c | 22 +++++++++----------
2 files changed, 19 insertions(+), 19 deletions(-)
diff --git a/clang/test/CodeGen/aarch64-sme2-intrinsics/acle_sme2_mls.c b/clang/test/CodeGen/aarch64-sme2-intrinsics/acle_sme2_mls.c
index bb4cbbd5308c06..fd4bf390b8232e 100644
--- a/clang/test/CodeGen/aarch64-sme2-intrinsics/acle_sme2_mls.c
+++ b/clang/test/CodeGen/aarch64-sme2-intrinsics/acle_sme2_mls.c
@@ -36,7 +36,7 @@
// CPP-CHECK-NEXT: ret void
//
void test_svmls2_f32(uint32_t slice_base, svfloat32x2_t zn, svfloat32x2_t zm) __arm_streaming __arm_shared_za {
- SVE_ACLE_FUNC(svmls_za32,,_f32,,_vg1x2)(slice_base, zn, zm);
+ SVE_ACLE_FUNC(svmls_za32,_f32,_vg1x2,,)(slice_base, zn, zm);
}
// CHECK-LABEL: @test_svmls4_f32(
@@ -66,7 +66,7 @@ void test_svmls2_f32(uint32_t slice_base, svfloat32x2_t zn, svfloat32x2_t zm) __
// CPP-CHECK-NEXT: ret void
//
void test_svmls4_f32(uint32_t slice_base, svfloat32x4_t zn, svfloat32x4_t zm) __arm_streaming __arm_shared_za {
- SVE_ACLE_FUNC(svmls_za32,,_f32,,_vg1x4)(slice_base, zn, zm);
+ SVE_ACLE_FUNC(svmls_za32,_f32,_vg1x4,,)(slice_base, zn, zm);
}
//
@@ -128,7 +128,7 @@ void test_svmls_single4_f32(uint32_t slice_base, svfloat32x4_t zn, svfloat32_t z
// CPP-CHECK-NEXT: ret void
//
void test_svmls_lane2_f32(uint32_t slice_base, svfloat32x2_t zn, svfloat32_t zm) __arm_streaming __arm_shared_za {
- SVE_ACLE_FUNC(svmls_lane_za32,,_f32,,_vg1x2)(slice_base, zn, zm, 3);
+ SVE_ACLE_FUNC(svmls_lane_za32,_f32,_vg1x2,,)(slice_base, zn, zm, 3);
}
// CHECK-LABEL: @test_svmls_lane4_f32(
@@ -150,7 +150,7 @@ void test_svmls_lane2_f32(uint32_t slice_base, svfloat32x2_t zn, svfloat32_t zm)
// CPP-CHECK-NEXT: ret void
//
void test_svmls_lane4_f32(uint32_t slice_base, svfloat32x4_t zn, svfloat32_t zm) __arm_streaming __arm_shared_za {
- SVE_ACLE_FUNC(svmls_lane_za32,,_f32,,_vg1x4)(slice_base, zn, zm, 3);
+ SVE_ACLE_FUNC(svmls_lane_za32,_f32,_vg1x4,,)(slice_base, zn, zm, 3);
}
//
@@ -174,7 +174,7 @@ void test_svmls_lane4_f32(uint32_t slice_base, svfloat32x4_t zn, svfloat32_t zm)
// CPP-CHECK-NEXT: ret void
//
void test_svmls2_f64(uint32_t slice_base, svfloat64x2_t zn, svfloat64x2_t zm) __arm_streaming __arm_shared_za {
- SVE_ACLE_FUNC(svmls_za64,,_f64,,_vg1x2)(slice_base, zn, zm);
+ SVE_ACLE_FUNC(svmls_za64,_f64,_vg1x2,,)(slice_base, zn, zm);
}
// CHECK-LABEL: @test_svmls4_f64(
@@ -204,7 +204,7 @@ void test_svmls2_f64(uint32_t slice_base, svfloat64x2_t zn, svfloat64x2_t zm) __
// CPP-CHECK-NEXT: ret void
//
void test_svmls4_f64(uint32_t slice_base, svfloat64x4_t zn, svfloat64x4_t zm) __arm_streaming __arm_shared_za {
- SVE_ACLE_FUNC(svmls_za64,,_f64,,_vg1x4)(slice_base, zn, zm);
+ SVE_ACLE_FUNC(svmls_za64,_f64,_vg1x4,,)(slice_base, zn, zm);
}
//
@@ -266,7 +266,7 @@ void test_svmls_single4_f64(uint32_t slice_base, svfloat64x4_t zn, svfloat64_t z
// CPP-CHECK-NEXT: ret void
//
void test_svmls_lane2_f64(uint32_t slice_base, svfloat64x2_t zn, svfloat64_t zm) __arm_streaming __arm_shared_za {
- SVE_ACLE_FUNC(svmls_lane_za64,,_f64,,_vg1x2)(slice_base, zn, zm, 1);
+ SVE_ACLE_FUNC(svmls_lane_za64,_f64,_vg1x2,,)(slice_base, zn, zm, 1);
}
// CHECK-LABEL: @test_svmls_lane4_f64(
@@ -288,5 +288,5 @@ void test_svmls_lane2_f64(uint32_t slice_base, svfloat64x2_t zn, svfloat64_t zm)
// CPP-CHECK-NEXT: ret void
//
void test_svmls_lane4_f64(uint32_t slice_base, svfloat64x4_t zn, svfloat64_t zm) __arm_streaming __arm_shared_za {
- SVE_ACLE_FUNC(svmls_lane_za64,,_f64,,_vg1x4)(slice_base, zn, zm, 1);
+ SVE_ACLE_FUNC(svmls_lane_za64,_f64,_vg1x4,,)(slice_base, zn, zm, 1);
}
diff --git a/clang/test/CodeGen/aarch64-sme2-intrinsics/acle_sme2_mlsl.c b/clang/test/CodeGen/aarch64-sme2-intrinsics/acle_sme2_mlsl.c
index 2c996eef9d2de4..f30ad70aed0f77 100644
--- a/clang/test/CodeGen/aarch64-sme2-intrinsics/acle_sme2_mlsl.c
+++ b/clang/test/CodeGen/aarch64-sme2-intrinsics/acle_sme2_mlsl.c
@@ -494,7 +494,7 @@ void test_svmls_lane1_f16(uint32_t slice_base, svfloat16_t zn, svfloat16_t zm) _
//
void test_svmls_lane1_bf16(uint32_t slice_base, svbfloat16_t zn, svbfloat16_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmls_lane_za32,,_bf16,,_vg2x1)(slice_base, zn, zm, 7);
+ SVE_ACLE_FUNC(svmls_lane_za32,_bf16,_vg2x1,,)(slice_base, zn, zm, 7);
}
// CHECK-LABEL: @test_svmls_lane1_u16(
@@ -509,7 +509,7 @@ void test_svmls_lane1_bf16(uint32_t slice_base, svbfloat16_t zn, svbfloat16_t zm
//
void test_svmls_lane1_u16(uint32_t slice_base, svuint16_t zn, svuint16_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmls_lane_za32,,_u16,,_vg2x1)(slice_base, zn, zm, 7);
+ SVE_ACLE_FUNC(svmls_lane_za32,_u16,_vg2x1,,)(slice_base, zn, zm, 7);
}
// CHECK-LABEL: @test_svmls_lane1_s16(
@@ -524,7 +524,7 @@ void test_svmls_lane1_u16(uint32_t slice_base, svuint16_t zn, svuint16_t zm) __a
//
void test_svmls_lane1_s16(uint32_t slice_base, svint16_t zn, svint16_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmls_lane_za32,,_s16,,_vg2x1)(slice_base, zn, zm, 7);
+ SVE_ACLE_FUNC(svmls_lane_za32,_s16,_vg2x1,,)(slice_base, zn, zm, 7);
}
// CHECK-LABEL: @test_svmls_lane2_f16(
@@ -543,7 +543,7 @@ void test_svmls_lane1_s16(uint32_t slice_base, svint16_t zn, svint16_t zm) __arm
//
void test_svmls_lane2_f16(uint32_t slice_base, svfloat16x2_t zn, svfloat16_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmls_lane_za32,,_f16,,_vg2x2)(slice_base, zn, zm, 7);
+ SVE_ACLE_FUNC(svmls_lane_za32,_f16,_vg2x2,,)(slice_base, zn, zm, 7);
}
// CHECK-LABEL: @test_svmls_lane2_bf16(
@@ -562,7 +562,7 @@ void test_svmls_lane2_f16(uint32_t slice_base, svfloat16x2_t zn, svfloat16_t zm)
//
void test_svmls_lane2_bf16(uint32_t slice_base, svbfloat16x2_t zn, svbfloat16_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmls_lane_za32,,_bf16,,_vg2x2)(slice_base, zn, zm, 7);
+ SVE_ACLE_FUNC(svmls_lane_za32,_bf16,_vg2x2,,)(slice_base, zn, zm, 7);
}
// CHECK-LABEL: @test_svmls_lane2_u16(
@@ -581,7 +581,7 @@ void test_svmls_lane2_bf16(uint32_t slice_base, svbfloat16x2_t zn, svbfloat16_t
//
void test_svmls_lane2_u16(uint32_t slice_base, svuint16x2_t zn, svuint16_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmls_lane_za32,,_u16,,_vg2x2)(slice_base, zn, zm, 7);
+ SVE_ACLE_FUNC(svmls_lane_za32,_u16,_vg2x2,,)(slice_base, zn, zm, 7);
}
// CHECK-LABEL: @test_svmls_lane2_s16(
@@ -600,7 +600,7 @@ void test_svmls_lane2_u16(uint32_t slice_base, svuint16x2_t zn, svuint16_t zm) _
//
void test_svmls_lane2_s16(uint32_t slice_base, svint16x2_t zn, svint16_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmls_lane_za32,,_s16,,_vg2x2)(slice_base, zn, zm, 7);
+ SVE_ACLE_FUNC(svmls_lane_za32,_s16,_vg2x2,,)(slice_base, zn, zm, 7);
}
// CHECK-LABEL: @test_svmls_lane4_f16(
@@ -623,7 +623,7 @@ void test_svmls_lane2_s16(uint32_t slice_base, svint16x2_t zn, svint16_t zm) __a
//
void test_svmls_lane4_f16(uint32_t slice_base, svfloat16x4_t zn, svfloat16_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmls_lane_za32,,_f16,,_vg2x4)(slice_base, zn, zm, 7);
+ SVE_ACLE_FUNC(svmls_lane_za32,_f16,_vg2x4,,)(slice_base, zn, zm, 7);
}
// CHECK-LABEL: @test_svmls_lane4_bf16(
@@ -646,7 +646,7 @@ void test_svmls_lane4_f16(uint32_t slice_base, svfloat16x4_t zn, svfloat16_t zm)
//
void test_svmls_lane4_bf16(uint32_t slice_base, svbfloat16x4_t zn, svbfloat16_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmls_lane_za32,,_bf16,,_vg2x4)(slice_base, zn, zm, 7);
+ SVE_ACLE_FUNC(svmls_lane_za32,_bf16,_vg2x4,,)(slice_base, zn, zm, 7);
}
// CHECK-LABEL: @test_svmls_lane4_u16(
@@ -669,7 +669,7 @@ void test_svmls_lane4_bf16(uint32_t slice_base, svbfloat16x4_t zn, svbfloat16_t
//
void test_svmls_lane4_u16(uint32_t slice_base, svuint16x4_t zn, svuint16_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmls_lane_za32,,_u16,,_vg2x4)(slice_base, zn, zm, 7);
+ SVE_ACLE_FUNC(svmls_lane_za32,_u16,_vg2x4,,)(slice_base, zn, zm, 7);
}
// CHECK-LABEL: @test_svmls_lane4_s16(
@@ -692,5 +692,5 @@ void test_svmls_lane4_u16(uint32_t slice_base, svuint16x4_t zn, svuint16_t zm) _
//
void test_svmls_lane4_s16(uint32_t slice_base, svint16x4_t zn, svint16_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmls_lane_za32,,_s16,,_vg2x4)(slice_base, zn, zm, 7);
+ SVE_ACLE_FUNC(svmls_lane_za32,_s16,_vg2x4,,)(slice_base, zn, zm, 7);
}
>From 752a35c4ea086d66275ad884f039b732c4fa4315 Mon Sep 17 00:00:00 2001
From: Matt Devereau <matthew.devereau at arm.com>
Date: Wed, 10 Jan 2024 11:43:05 +0000
Subject: [PATCH 4/4] Update missed svmls_lane test and svmla_lane builtins
---
.../aarch64-sme2-intrinsics/acle_sme2_mlal.c | 24 +++++++++----------
.../aarch64-sme2-intrinsics/acle_sme2_mlsl.c | 2 +-
2 files changed, 13 insertions(+), 13 deletions(-)
diff --git a/clang/test/CodeGen/aarch64-sme2-intrinsics/acle_sme2_mlal.c b/clang/test/CodeGen/aarch64-sme2-intrinsics/acle_sme2_mlal.c
index 313ce4418682b2..032830512987af 100644
--- a/clang/test/CodeGen/aarch64-sme2-intrinsics/acle_sme2_mlal.c
+++ b/clang/test/CodeGen/aarch64-sme2-intrinsics/acle_sme2_mlal.c
@@ -479,7 +479,7 @@ void test_svmla_single4_s16(uint32_t slice_base, svint16x4_t zn, svint16_t zm) _
//
void test_svmla_lane1_f16(uint32_t slice_base, svfloat16_t zn, svfloat16_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmla_lane_za32,,,_f16,_vg2x1)(slice_base, zn, zm, 7);
+ SVE_ACLE_FUNC(svmla_lane_za32,_f16,_vg2x1,,)(slice_base, zn, zm, 7);
}
// CHECK-LABEL: @test_svmla_lane1_bf16(
@@ -494,7 +494,7 @@ void test_svmla_lane1_f16(uint32_t slice_base, svfloat16_t zn, svfloat16_t zm) _
//
void test_svmla_lane1_bf16(uint32_t slice_base, svbfloat16_t zn, svbfloat16_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmla_lane_za32,,,_bf16,_vg2x1)(slice_base, zn, zm, 7);
+ SVE_ACLE_FUNC(svmla_lane_za32,_bf16,_vg2x1,,)(slice_base, zn, zm, 7);
}
// CHECK-LABEL: @test_svmla_lane1_u16(
@@ -509,7 +509,7 @@ void test_svmla_lane1_bf16(uint32_t slice_base, svbfloat16_t zn, svbfloat16_t zm
//
void test_svmla_lane1_u16(uint32_t slice_base, svuint16_t zn, svuint16_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmla_lane_za32,,,_u16,_vg2x1)(slice_base, zn, zm, 7);
+ SVE_ACLE_FUNC(svmla_lane_za32,_u16,_vg2x1,,)(slice_base, zn, zm, 7);
}
// CHECK-LABEL: @test_svmla_lane1_s16(
@@ -524,7 +524,7 @@ void test_svmla_lane1_u16(uint32_t slice_base, svuint16_t zn, svuint16_t zm) __a
//
void test_svmla_lane1_s16(uint32_t slice_base, svint16_t zn, svint16_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmla_lane_za32,,,_s16,_vg2x1)(slice_base, zn, zm, 7);
+ SVE_ACLE_FUNC(svmla_lane_za32,_s16,_vg2x1,,)(slice_base, zn, zm, 7);
}
// CHECK-LABEL: @test_svmla_lane2_f16(
@@ -543,7 +543,7 @@ void test_svmla_lane1_s16(uint32_t slice_base, svint16_t zn, svint16_t zm) __arm
//
void test_svmla_lane2_f16(uint32_t slice_base, svfloat16x2_t zn, svfloat16_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmla_lane_za32,,,_f16,_vg2x2)(slice_base, zn, zm, 7);
+ SVE_ACLE_FUNC(svmla_lane_za32,_f16,_vg2x2,,)(slice_base, zn, zm, 7);
}
// CHECK-LABEL: @test_svmla_lane2_bf16(
@@ -562,7 +562,7 @@ void test_svmla_lane2_f16(uint32_t slice_base, svfloat16x2_t zn, svfloat16_t zm)
//
void test_svmla_lane2_bf16(uint32_t slice_base, svbfloat16x2_t zn, svbfloat16_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmla_lane_za32,,,_bf16,_vg2x2)(slice_base, zn, zm, 7);
+ SVE_ACLE_FUNC(svmla_lane_za32,_bf16,_vg2x2,,)(slice_base, zn, zm, 7);
}
// CHECK-LABEL: @test_svmla_lane2_u16(
@@ -581,7 +581,7 @@ void test_svmla_lane2_bf16(uint32_t slice_base, svbfloat16x2_t zn, svbfloat16_t
//
void test_svmla_lane2_u16(uint32_t slice_base, svuint16x2_t zn, svuint16_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmla_lane_za32,,,_u16,_vg2x2)(slice_base, zn, zm, 7);
+ SVE_ACLE_FUNC(svmla_lane_za32,_u16,_vg2x2,,)(slice_base, zn, zm, 7);
}
// CHECK-LABEL: @test_svmla_lane2_s16(
@@ -600,7 +600,7 @@ void test_svmla_lane2_u16(uint32_t slice_base, svuint16x2_t zn, svuint16_t zm) _
//
void test_svmla_lane2_s16(uint32_t slice_base, svint16x2_t zn, svint16_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmla_lane_za32,,,_s16,_vg2x2)(slice_base, zn, zm, 7);
+ SVE_ACLE_FUNC(svmla_lane_za32,_s16,_vg2x2,,)(slice_base, zn, zm, 7);
}
// CHECK-LABEL: @test_svmla_lane4_f16(
@@ -623,7 +623,7 @@ void test_svmla_lane2_s16(uint32_t slice_base, svint16x2_t zn, svint16_t zm) __a
//
void test_svmla_lane4_f16(uint32_t slice_base, svfloat16x4_t zn, svfloat16_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmla_lane_za32,,,_f16,_vg2x4)(slice_base, zn, zm, 7);
+ SVE_ACLE_FUNC(svmla_lane_za32,_f16,_vg2x4,,)(slice_base, zn, zm, 7);
}
// CHECK-LABEL: @test_svmla_lane4_bf16(
@@ -646,7 +646,7 @@ void test_svmla_lane4_f16(uint32_t slice_base, svfloat16x4_t zn, svfloat16_t zm)
//
void test_svmla_lane4_bf16(uint32_t slice_base, svbfloat16x4_t zn, svbfloat16_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmla_lane_za32,,,_bf16,_vg2x4)(slice_base, zn, zm, 7);
+ SVE_ACLE_FUNC(svmla_lane_za32,_bf16,_vg2x4,,)(slice_base, zn, zm, 7);
}
// CHECK-LABEL: @test_svmla_lane4_u16(
@@ -669,7 +669,7 @@ void test_svmla_lane4_bf16(uint32_t slice_base, svbfloat16x4_t zn, svbfloat16_t
//
void test_svmla_lane4_u16(uint32_t slice_base, svuint16x4_t zn, svuint16_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmla_lane_za32,,,_u16,_vg2x4)(slice_base, zn, zm, 7);
+ SVE_ACLE_FUNC(svmla_lane_za32,_u16,_vg2x4,,)(slice_base, zn, zm, 7);
}
// CHECK-LABEL: @test_svmla_lane4_s16(
@@ -692,5 +692,5 @@ void test_svmla_lane4_u16(uint32_t slice_base, svuint16x4_t zn, svuint16_t zm) _
//
void test_svmla_lane4_s16(uint32_t slice_base, svint16x4_t zn, svint16_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmla_lane_za32,,,_s16,_vg2x4)(slice_base, zn, zm, 7);
+ SVE_ACLE_FUNC(svmla_lane_za32,_s16,_vg2x4,,)(slice_base, zn, zm, 7);
}
diff --git a/clang/test/CodeGen/aarch64-sme2-intrinsics/acle_sme2_mlsl.c b/clang/test/CodeGen/aarch64-sme2-intrinsics/acle_sme2_mlsl.c
index f30ad70aed0f77..afc7db797c9dfc 100644
--- a/clang/test/CodeGen/aarch64-sme2-intrinsics/acle_sme2_mlsl.c
+++ b/clang/test/CodeGen/aarch64-sme2-intrinsics/acle_sme2_mlsl.c
@@ -479,7 +479,7 @@ void test_svmls_single4_s16(uint32_t slice_base, svint16x4_t zn, svint16_t zm) _
//
void test_svmls_lane1_f16(uint32_t slice_base, svfloat16_t zn, svfloat16_t zm) __arm_streaming __arm_shared_za
{
- SVE_ACLE_FUNC(svmls_lane_za32,,_f16,,_vg2x1)(slice_base, zn, zm, 7);
+ SVE_ACLE_FUNC(svmls_lane_za32,_f16,_vg2x1,,)(slice_base, zn, zm, 7);
}
// CHECK-LABEL: @test_svmls_lane1_bf16(
More information about the cfe-commits
mailing list