[Clang][AArch64]Remove _single from vg2x1 and vg4x1 svmla #166799
+38
−34
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
llvmbot
added
clang
Member
|
@llvm/pr-subscribers-llvm-ir @llvm/pr-subscribers-clang Author: None (CarolineConcatto) ChangesThis patch follows the PR#421[1] from the ACLE These 2 FP8 intrinsics had single removed from them: from Full diff: https://github.com/llvm/llvm-project/pull/166799.diff 2 Files Affected:
diff --git a/clang/include/clang/Basic/arm_sme.td b/clang/include/clang/Basic/arm_sme.td
index 5f6a6eaab80a3..8fdb93519e8ec 100644
--- a/clang/include/clang/Basic/arm_sme.td
+++ b/clang/include/clang/Basic/arm_sme.td
@@ -992,7 +992,7 @@ let SMETargetGuard = "sme-f8f32" in {
def SVMLA_FP8_LANE_ZA16_VG4x4 : Inst<"svmla_lane_za32[_mf8]_vg4x4", "vm4di>", "m", MergeNone, "aarch64_sme_fp8_fmlall_lane_za32_vg4x4",
[IsStreaming, IsInOutZA, IsOverloadNone], [ImmCheck<3, ImmCheck0_15>]>;
// FMLALL (single)
- def SVMLA_FP8_SINGLE_ZA32_VG4x1 : Inst<"svmla[_single]_za32[_mf8]_vg4x1", "vmdd>", "m", MergeNone, "aarch64_sme_fp8_fmlall_single_za32_vg4x1",
+ def SVMLA_FP8_SINGLE_ZA32_VG4x1 : Inst<"svmla_za32[_mf8]_vg4x1", "vmdd>", "m", MergeNone, "aarch64_sme_fp8_fmlall_single_za32_vg4x1",
[IsStreaming, IsInOutZA, IsOverloadNone], []>;
def SVMLA_FP8_SINGLE_ZA32_VG4x2 : Inst<"svmla[_single]_za32[_mf8]_vg4x2", "vm2d>", "m", MergeNone, "aarch64_sme_fp8_fmlall_single_za32_vg4x2",
[IsStreaming, IsInOutZA, IsOverloadNone], []>;
@@ -1016,7 +1016,7 @@ let SMETargetGuard = "sme-f8f16" in {
def SVMLA_FP8_LANE_ZA16_VG2x4 : Inst<"svmla_lane_za16[_mf8]_vg2x4", "vm4di>", "m", MergeNone, "aarch64_sme_fp8_fmlal_lane_za16_vg2x4",
[IsStreaming, IsInOutZA, IsOverloadNone], [ImmCheck<3, ImmCheck0_15>]>;
// FMLAL (single)
- def SVMLA_FP8_SINGLE_ZA16_VG2x1 : Inst<"svmla[_single]_za16[_mf8]_vg2x1", "vmdd>", "m", MergeNone, "aarch64_sme_fp8_fmlal_single_za16_vg2x1",
+ def SVMLA_FP8_SINGLE_ZA16_VG2x1 : Inst<"svmla_za16[_mf8]_vg2x1", "vmdd>", "m", MergeNone, "aarch64_sme_fp8_fmlal_single_za16_vg2x1",
[IsStreaming, IsInOutZA, IsOverloadNone], []>;
def SVMLA_FP8_SINGLE_ZA16_VG2x2 : Inst<"svmla[_single]_za16[_mf8]_vg2x2", "vm2d>", "m", MergeNone, "aarch64_sme_fp8_fmlal_single_za16_vg2x2",
[IsStreaming, IsInOutZA, IsOverloadNone], []>;
diff --git a/clang/test/CodeGen/AArch64/fp8-intrinsics/acle_sme2_fp8_mla.c b/clang/test/CodeGen/AArch64/fp8-intrinsics/acle_sme2_fp8_mla.c
index d603045edf282..601b66710f6ff 100644
--- a/clang/test/CodeGen/AArch64/fp8-intrinsics/acle_sme2_fp8_mla.c
+++ b/clang/test/CodeGen/AArch64/fp8-intrinsics/acle_sme2_fp8_mla.c
@@ -144,7 +144,7 @@ void test_svmla_lane_za32_vg4x4(uint32_t slice, svmfloat8x4_t zn, svmfloat8_t zm
// CPP-CHECK-NEXT: ret void
//
void test_svmla_single_za16_vg2x1(uint32_t slice, svmfloat8_t zn, svmfloat8_t zm, fpm_t fpm) __arm_streaming __arm_inout("za") {
- SME_ACLE_FUNC(svmla,_single,_za16,_mf8,_vg2x1_fpm)(slice, zn, zm, fpm);
+ SME_ACLE_FUNC(svmla_za16,_mf8,_vg2x1_fpm,,)(slice, zn, zm, fpm);
}
// CHECK-LABEL: define dso_local void @test_svmla_single_za16_vg2x2(
@@ -200,7 +200,7 @@ void test_svmla_single_za16_vg2x4(uint32_t slice, svmfloat8x4_t zn, svmfloat8_t
// CPP-CHECK-NEXT: ret void
//
void test_svmla_single_za32_vg4x1(uint32_t slice, svmfloat8_t zn, svmfloat8_t zm, fpm_t fpm) __arm_streaming __arm_inout("za") {
- SME_ACLE_FUNC(svmla,_single,_za32,_mf8,_vg4x1_fpm)(slice, zn, zm, fpm);
+ SME_ACLE_FUNC(svmla_za32,_mf8,_vg4x1_fpm,,)(slice, zn, zm, fpm);
}
// CHECK-LABEL: define dso_local void @test_svmla_single_za32_vg4x2(
|
SpencerAbson
reviewed
Nov 6, 2025
Contributor
SpencerAbson
left a comment
SpencerAbson
left a comment
There was a problem hiding this comment.
It looks like there are references to the '_single' names in clang/test/Sema. Aside from that this LGTM :).
This patch follows the PR#421[1] from the ACLE These 2 FP8 intrinsics had single removed from them: from ``svmla[_single]_za16[_mf8]_vg2x1_fpm`` to ``svmla_za16[_mf8]_vg2x1_fpm`` and from ``svmla[_single]_za32[_mf8]_vg4x1_fpm`` to ``svmla_za32[_mf8]_vg4x1_fpm`` [1]ARM-software/acle#421
CarolineConcatto
force-pushed
the
fp8_mla_fix_clang
branch
from
6954a1f to
306d31e
Compare
SpencerAbson
approved these changes
Nov 13, 2025
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This patch follows the PR#421[1] from the ACLE
These 2 FP8 intrinsics had single removed from them: from
svmla[_single]_za16[_mf8]_vg2x1_fpmtosvmla_za16[_mf8]_vg2x1_fpmand fromsvmla[_single]_za32[_mf8]_vg4x1_fpmtosvmla_za32[_mf8]_vg4x1_fpm[1]ARM-software/acle#421