[llvm] [AArch64] Add Neoverse V3AE scheduling info for v264 frecpe (PR #219115)
via llvm-commits
llvm-commits at lists.llvm.org
Wed Aug 26 22:28:34 PDT 2026
llvmorg-github-actions[bot] wrote:
<!--LLVM PR SUMMARY COMMENT-->
@llvm/pr-subscribers-backend-aarch64
Author: David Green (davemgreen)
<details>
<summary>Changes</summary>
This was missing, falling back to the default information. It should be like the others.
---
Full diff: https://github.com/llvm/llvm-project/pull/219115.diff
2 Files Affected:
- (modified) llvm/lib/Target/AArch64/AArch64SchedNeoverseV3AE.td (+1)
- (modified) llvm/test/tools/llvm-mca/AArch64/Neoverse/V3AE-neon-instructions.s (+5-5)
``````````diff
diff --git a/llvm/lib/Target/AArch64/AArch64SchedNeoverseV3AE.td b/llvm/lib/Target/AArch64/AArch64SchedNeoverseV3AE.td
index 29f8fa783467a..39d92f6c4e8a9 100644
--- a/llvm/lib/Target/AArch64/AArch64SchedNeoverseV3AE.td
+++ b/llvm/lib/Target/AArch64/AArch64SchedNeoverseV3AE.td
@@ -1576,6 +1576,7 @@ def : InstRW<[V3AEWrite_4c_2V0], (instrs URECPEv4i32, URSQRTEv4i32)>;
// ASIMD reciprocal and square root estimate, D-form F32 and scalar forms
def : InstRW<[V3AEWrite_3c_1V0], (instrs FRECPEv1f16, FRECPEv1i32,
FRECPEv1i64, FRECPEv2f32,
+ FRECPEv2f64, FRSQRTEv2f64,
FRSQRTEv1f16, FRSQRTEv1i32,
FRSQRTEv1i64, FRSQRTEv2f32)>;
diff --git a/llvm/test/tools/llvm-mca/AArch64/Neoverse/V3AE-neon-instructions.s b/llvm/test/tools/llvm-mca/AArch64/Neoverse/V3AE-neon-instructions.s
index b2d3f604013dc..b5fc1d0302cae 100644
--- a/llvm/test/tools/llvm-mca/AArch64/Neoverse/V3AE-neon-instructions.s
+++ b/llvm/test/tools/llvm-mca/AArch64/Neoverse/V3AE-neon-instructions.s
@@ -398,7 +398,7 @@
# CHECK-NEXT: 1 3 1.00 frecpe h20, h8
# CHECK-NEXT: 1 3 1.00 frecpe d13, d13
# CHECK-NEXT: 1 3 1.00 frecpe s19, s14
-# CHECK-NEXT: 1 2 0.50 frecpe v0.2d, v0.2d
+# CHECK-NEXT: 1 3 1.00 frecpe v0.2d, v0.2d
# CHECK-NEXT: 1 3 1.00 frecpe v0.2s, v0.2s
# CHECK-NEXT: 2 4 2.00 frecpe v0.4h, v0.4h
# CHECK-NEXT: 2 4 2.00 frecpe v0.4s, v0.4s
@@ -450,7 +450,7 @@
# CHECK-NEXT: 1 3 1.00 frsqrte h23, h26
# CHECK-NEXT: 1 3 1.00 frsqrte d21, d12
# CHECK-NEXT: 1 3 1.00 frsqrte s22, s13
-# CHECK-NEXT: 1 2 0.50 frsqrte v0.2d, v0.2d
+# CHECK-NEXT: 1 3 1.00 frsqrte v0.2d, v0.2d
# CHECK-NEXT: 1 3 1.00 frsqrte v0.2s, v0.2s
# CHECK-NEXT: 2 4 2.00 frsqrte v0.4h, v0.4h
# CHECK-NEXT: 2 4 2.00 frsqrte v0.4s, v0.4s
@@ -1599,7 +1599,7 @@
# CHECK: Resource pressure per iteration:
# CHECK-NEXT: [0.0] [0.1] [0.2] [1.0] [1.1] [2.0] [2.1] [2.2] [2.3] [3.0] [3.1] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15]
-# CHECK-NEXT: - - - - - - - - - 106.00 106.00 189.00 19.13 6.13 6.13 6.13 6.13 6.13 6.13 6.13 83.00 1341.50 859.50
+# CHECK-NEXT: - - - - - - - - - 106.00 106.00 189.00 19.13 6.13 6.13 6.13 6.13 6.13 6.13 6.13 83.00 1342.50 858.50
# CHECK: Resource pressure by instruction:
# CHECK-NEXT: [0.0] [0.1] [0.2] [1.0] [1.1] [2.0] [2.1] [2.2] [2.3] [3.0] [3.1] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] Instructions:
@@ -1991,7 +1991,7 @@
# CHECK-NEXT: - - - - - - - - - - - - - - - - - - - - - 1.00 - frecpe h20, h8
# CHECK-NEXT: - - - - - - - - - - - - - - - - - - - - - 1.00 - frecpe d13, d13
# CHECK-NEXT: - - - - - - - - - - - - - - - - - - - - - 1.00 - frecpe s19, s14
-# CHECK-NEXT: - - - - - - - - - - - - - - - - - - - - - 0.50 0.50 frecpe v0.2d, v0.2d
+# CHECK-NEXT: - - - - - - - - - - - - - - - - - - - - - 1.00 - frecpe v0.2d, v0.2d
# CHECK-NEXT: - - - - - - - - - - - - - - - - - - - - - 1.00 - frecpe v0.2s, v0.2s
# CHECK-NEXT: - - - - - - - - - - - - - - - - - - - - - 2.00 - frecpe v0.4h, v0.4h
# CHECK-NEXT: - - - - - - - - - - - - - - - - - - - - - 2.00 - frecpe v0.4s, v0.4s
@@ -2043,7 +2043,7 @@
# CHECK-NEXT: - - - - - - - - - - - - - - - - - - - - - 1.00 - frsqrte h23, h26
# CHECK-NEXT: - - - - - - - - - - - - - - - - - - - - - 1.00 - frsqrte d21, d12
# CHECK-NEXT: - - - - - - - - - - - - - - - - - - - - - 1.00 - frsqrte s22, s13
-# CHECK-NEXT: - - - - - - - - - - - - - - - - - - - - - 0.50 0.50 frsqrte v0.2d, v0.2d
+# CHECK-NEXT: - - - - - - - - - - - - - - - - - - - - - 1.00 - frsqrte v0.2d, v0.2d
# CHECK-NEXT: - - - - - - - - - - - - - - - - - - - - - 1.00 - frsqrte v0.2s, v0.2s
# CHECK-NEXT: - - - - - - - - - - - - - - - - - - - - - 2.00 - frsqrte v0.4h, v0.4h
# CHECK-NEXT: - - - - - - - - - - - - - - - - - - - - - 2.00 - frsqrte v0.4s, v0.4s
``````````
</details>
https://github.com/llvm/llvm-project/pull/219115
More information about the llvm-commits
mailing list