Skip to content

[WebAssembly] Partial SMLA with relaxed dot #163529

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Merged
merged 1 commit into from
Oct 16, 2025
Merged

[WebAssembly] Partial SMLA with relaxed dot #163529

merged 1 commit into from
Oct 16, 2025

Conversation

@sparker-arm
Copy link
Contributor

@sparker-arm sparker-arm commented Oct 15, 2025

Lower v16i8 to v4i32 partial_smla to relaxed_dot. I'm still unsure whether we could/should take advantage of the unknown signedness of the rhs, and also lower the partial_sumla operation too.

@sparker-arm
[WebAssembly] Partial SMLA with relaxed dot
d82e97e 
Lower v16i8 to v4i32 partial_smla to relaxed_dot. I'm still unsure
whether we could/should take advantage of the unknown signedness of
the rhs, and also lower the partial_sumla operation too.
@sparker-arm sparker-arm self-assigned this Oct 15, 2025
@llvmbot
Copy link
Member

llvmbot commented Oct 15, 2025

@llvm/pr-subscribers-backend-webassembly

Author: Sam Parker (sparker-arm)

Changes

Lower v16i8 to v4i32 partial_smla to relaxed_dot. I'm still unsure whether we could/should take advantage of the unknown signedness of the rhs, and also lower the partial_sumla operation too.

Full diff: https://github.com/llvm/llvm-project/pull/163529.diff

2 Files Affected:

  • (modified) llvm/lib/Target/WebAssembly/WebAssemblyInstrSIMD.td (+4)
  • (modified) llvm/test/CodeGen/WebAssembly/int-mac-reduction-loops.ll (+208)
diff --git a/llvm/lib/Target/WebAssembly/WebAssemblyInstrSIMD.td b/llvm/lib/Target/WebAssembly/WebAssemblyInstrSIMD.td
index 49af78bce68c3..857c4b187d668 100644
--- a/llvm/lib/Target/WebAssembly/WebAssemblyInstrSIMD.td
+++ b/llvm/lib/Target/WebAssembly/WebAssemblyInstrSIMD.td
@@ -1719,6 +1719,10 @@ defm RELAXED_DOT_ADD :
             "i32x4.relaxed_dot_i8x16_i7x16_add_s\t$dst, $lhs, $rhs, $acc",
             "i32x4.relaxed_dot_i8x16_i7x16_add_s", 0x113>;
 
+def : Pat<(v4i32 (partial_reduce_smla (v4i32 V128:$acc), (v16i8 V128:$lhs),
+                                                         (v16i8 V128:$rhs))),
+          (RELAXED_DOT_ADD $lhs, $rhs, $acc)>, Requires<[HasRelaxedSIMD]>;
+
 //===----------------------------------------------------------------------===//
 // Relaxed BFloat16 dot product
 //===----------------------------------------------------------------------===//
diff --git a/llvm/test/CodeGen/WebAssembly/int-mac-reduction-loops.ll b/llvm/test/CodeGen/WebAssembly/int-mac-reduction-loops.ll
index 04a2268db1755..314e1b4fc69a1 100644
--- a/llvm/test/CodeGen/WebAssembly/int-mac-reduction-loops.ll
+++ b/llvm/test/CodeGen/WebAssembly/int-mac-reduction-loops.ll
@@ -1,5 +1,6 @@
 ; RUN: opt -mattr=+simd128 -passes=loop-vectorize %s | llc -mtriple=wasm32 -mattr=+simd128 -verify-machineinstrs -o - | FileCheck %s
 ; RUN: opt -mattr=+simd128 -passes=loop-vectorize -vectorizer-maximize-bandwidth %s | llc -mtriple=wasm32 -mattr=+simd128 -verify-machineinstrs -o - | FileCheck %s --check-prefix=MAX-BANDWIDTH
+; RUN: opt -mattr=+simd128,+relaxed-simd -passes=loop-vectorize -vectorizer-maximize-bandwidth %s | llc -mtriple=wasm32 -mattr=+simd128,+relaxed-simd -verify-machineinstrs -o - | FileCheck %s --check-prefix=RELAXED-MAX-BANDWIDTH
 
 target triple = "wasm32"
 
@@ -23,6 +24,10 @@ define hidden i32 @i32_mac_s8(ptr nocapture noundef readonly %a, ptr nocapture n
 ; MAX-BANDWIDTH: i32x4.add
 ; MAX-BANDWIDTH: i32x4.add
 
+; RELAXED-MAX-BANDWIDTH: v128.load
+; RELAXED-MAX-BANDWIDTH: v128.load
+; RELAXED-MAX-BANDWIDTH: i32x4.relaxed_dot_i8x16_i7x16_add_s
+
 entry:
   %cmp7.not = icmp eq i32 %N, 0
   br i1 %cmp7.not, label %for.cond.cleanup, label %for.body
@@ -47,6 +52,109 @@ for.body:                                         ; preds = %entry, %for.body
   br i1 %exitcond.not, label %for.cond.cleanup, label %for.body
 }
 
+define hidden i32 @i32_mac_u8_s8(ptr nocapture noundef readonly %a, ptr nocapture noundef readonly %b, i32 noundef %N) {
+; CHECK-LABEL: i32_mac_u8_s8:
+; CHECK: loop
+; CHECK: v128.load32_zero
+; CHECK: i16x8.extend_low_i8x16_u
+; CHECK: i32x4.extend_low_i16x8_u
+; CHECK: v128.load32_zero
+; CHECK: i16x8.extend_low_i8x16_s
+; CHECK: i32x4.extend_low_i16x8_s
+; CHECK: i32x4.mul
+; CHECK: i32x4.add
+
+; MAX-BANDWIDTH: loop
+; MAX-BANDWIDTH: v128.load
+; MAX-BANDWIDTH: i8x16.shuffle   12, 13, 14, 15, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
+; MAX-BANDWIDTH: i16x8.extend_low_i8x16_u
+; MAX-BANDWIDTH: i32x4.extend_low_i16x8_u
+; MAX-BANDWIDTH: v128.load
+; MAX-BANDWIDTH: i8x16.shuffle   12, 13, 14, 15, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
+; MAX-BANDWIDTH: i16x8.extend_low_i8x16_s
+; MAX-BANDWIDTH: i32x4.extend_low_i16x8_s
+; MAX-BANDWIDTH: i32x4.mul
+; MAX-BANDWIDTH: i32x4.add
+; MAX-BANDWIDTH: i8x16.shuffle   8, 9, 10, 11, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
+; MAX-BANDWIDTH: i16x8.extend_low_i8x16_u
+; MAX-BANDWIDTH: i32x4.extend_low_i16x8_u
+; MAX-BANDWIDTH: i8x16.shuffle   8, 9, 10, 11, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
+; MAX-BANDWIDTH: i16x8.extend_low_i8x16_s
+; MAX-BANDWIDTH: i32x4.extend_low_i16x8_s
+; MAX-BANDWIDTH: i32x4.mul
+; MAX-BANDWIDTH: i32x4.add
+; MAX-BANDWIDTH: i8x16.shuffle   4, 5, 6, 7, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
+; MAX-BANDWIDTH: i16x8.extend_low_i8x16_u
+; MAX-BANDWIDTH: i32x4.extend_low_i16x8_u
+; MAX-BANDWIDTH: i8x16.shuffle   4, 5, 6, 7, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
+; MAX-BANDWIDTH: i16x8.extend_low_i8x16_s
+; MAX-BANDWIDTH: i32x4.extend_low_i16x8_s
+; MAX-BANDWIDTH: i32x4.mul
+; MAX-BANDWIDTH: i32x4.add
+; MAX-BANDWIDTH: i16x8.extend_low_i8x16_u
+; MAX-BANDWIDTH: i32x4.extend_low_i16x8_u
+; MAX-BANDWIDTH: i16x8.extend_low_i8x16_s
+; MAX-BANDWIDTH: i32x4.extend_low_i16x8_s
+; MAX-BANDWIDTH: i32x4.mul
+; MAX-BANDWIDTH: i32x4.add
+
+; RELAXED-MAX-BANDWIDTH: loop
+; RELAXED-MAX-BANDWIDTH: v128.load
+; RELAXED-MAX-BANDWIDTH: i8x16.shuffle   12, 13, 14, 15, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
+; RELAXED-MAX-BANDWIDTH: i16x8.extend_low_i8x16_u
+; RELAXED-MAX-BANDWIDTH: i32x4.extend_low_i16x8_u
+; RELAXED-MAX-BANDWIDTH: v128.load
+; RELAXED-MAX-BANDWIDTH: i8x16.shuffle   12, 13, 14, 15, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
+; RELAXED-MAX-BANDWIDTH: i16x8.extend_low_i8x16_s
+; RELAXED-MAX-BANDWIDTH: i32x4.extend_low_i16x8_s
+; RELAXED-MAX-BANDWIDTH: i32x4.mul
+; RELAXED-MAX-BANDWIDTH: i32x4.add
+; RELAXED-MAX-BANDWIDTH: i8x16.shuffle   8, 9, 10, 11, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
+; RELAXED-MAX-BANDWIDTH: i16x8.extend_low_i8x16_u
+; RELAXED-MAX-BANDWIDTH: i32x4.extend_low_i16x8_u
+; RELAXED-MAX-BANDWIDTH: i8x16.shuffle   8, 9, 10, 11, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
+; RELAXED-MAX-BANDWIDTH: i16x8.extend_low_i8x16_s
+; RELAXED-MAX-BANDWIDTH: i32x4.extend_low_i16x8_s
+; RELAXED-MAX-BANDWIDTH: i32x4.mul
+; RELAXED-MAX-BANDWIDTH: i32x4.add
+; RELAXED-MAX-BANDWIDTH: i8x16.shuffle   4, 5, 6, 7, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
+; RELAXED-MAX-BANDWIDTH: i16x8.extend_low_i8x16_u
+; RELAXED-MAX-BANDWIDTH: i32x4.extend_low_i16x8_u
+; RELAXED-MAX-BANDWIDTH: i8x16.shuffle   4, 5, 6, 7, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
+; RELAXED-MAX-BANDWIDTH: i16x8.extend_low_i8x16_s
+; RELAXED-MAX-BANDWIDTH: i32x4.extend_low_i16x8_s
+; RELAXED-MAX-BANDWIDTH: i32x4.mul
+; RELAXED-MAX-BANDWIDTH: i32x4.add
+; RELAXED-MAX-BANDWIDTH: i16x8.extend_low_i8x16_u
+; RELAXED-MAX-BANDWIDTH: i32x4.extend_low_i16x8_u
+; RELAXED-MAX-BANDWIDTH: i16x8.extend_low_i8x16_s
+; RELAXED-MAX-BANDWIDTH: i32x4.extend_low_i16x8_s
+; RELAXED-MAX-BANDWIDTH: i32x4.mul
+; RELAXED-MAX-BANDWIDTH: i32x4.add
+entry:
+  %cmp7.not = icmp eq i32 %N, 0
+  br i1 %cmp7.not, label %for.cond.cleanup, label %for.body
+
+for.cond.cleanup:                                 ; preds = %for.body, %entry
+  %res.0.lcssa = phi i32 [ 0, %entry ], [ %add, %for.body ]
+  ret i32 %res.0.lcssa
+
+for.body:                                         ; preds = %entry, %for.body
+  %i.09 = phi i32 [ %inc, %for.body ], [ 0, %entry ]
+  %res.08 = phi i32 [ %add, %for.body ], [ 0, %entry ]
+  %arrayidx = getelementptr inbounds i8, ptr %a, i32 %i.09
+  %0 = load i8, ptr %arrayidx, align 1
+  %conv = sext i8 %0 to i32
+  %arrayidx1 = getelementptr inbounds i8, ptr %b, i32 %i.09
+  %1 = load i8, ptr %arrayidx1, align 1
+  %conv2 = zext i8 %1 to i32
+  %mul = mul nsw i32 %conv2, %conv
+  %add = add nsw i32 %mul, %res.08
+  %inc = add nuw i32 %i.09, 1
+  %exitcond.not = icmp eq i32 %inc, %N
+  br i1 %exitcond.not, label %for.cond.cleanup, label %for.body
+}
+
 define hidden i32 @i32_mac_s16(ptr nocapture noundef readonly %a, ptr nocapture noundef readonly %b, i32 noundef %N) {
 ; CHECK-LABEL: i32_mac_s16:
 ; CHECK:    i32x4.load16x4_s 0:p2align=1
@@ -57,6 +165,12 @@ define hidden i32 @i32_mac_s16(ptr nocapture noundef readonly %a, ptr nocapture
 ; MAX-BANDWIDTH: v128.load
 ; MAX-BANDWIDTH: v128.load
 ; MAX-BANDWIDTH: i32x4.dot_i16x8_s
+; MAX-BANDWIDTH: i32x4.add
+
+; RELAXED-MAX-BANDWIDTH: v128.load
+; RELAXED-MAX-BANDWIDTH: v128.load
+; RELAXED-MAX-BANDWIDTH: i32x4.dot_i16x8_s
+; RELAXED-MAX-BANDWIDTH: i32x4.add
 
 entry:
   %cmp7.not = icmp eq i32 %N, 0
@@ -116,6 +230,31 @@ define hidden i64 @i64_mac_s16(ptr nocapture noundef readonly %a, ptr nocapture
 ; MAX-BANDWIDTH: i64x2.extmul_low_i32x4_s
 ; MAX-BANDWIDTH: i64x2.add
 
+; RELAXED-MAX-BANDWIDTH: v128.load
+; RELAXED-MAX-BANDWIDTH: i8x16.shuffle	12, 13, 14, 15, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1
+; RELAXED-MAX-BANDWIDTH: i32x4.extend_low_i16x8_s
+; RELAXED-MAX-BANDWIDTH: v128.load
+; RELAXED-MAX-BANDWIDTH: i8x16.shuffle	12, 13, 14, 15, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1
+; RELAXED-MAX-BANDWIDTH: i32x4.extend_low_i16x8_s
+; RELAXED-MAX-BANDWIDTH: i64x2.extmul_low_i32x4_s
+; RELAXED-MAX-BANDWIDTH: i64x2.add
+; RELAXED-MAX-BANDWIDTH: i8x16.shuffle	8, 9, 10, 11, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1
+; RELAXED-MAX-BANDWIDTH: i32x4.extend_low_i16x8_s
+; RELAXED-MAX-BANDWIDTH: i8x16.shuffle	8, 9, 10, 11, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1
+; RELAXED-MAX-BANDWIDTH: i32x4.extend_low_i16x8_s
+; RELAXED-MAX-BANDWIDTH: i64x2.extmul_low_i32x4_s
+; RELAXED-MAX-BANDWIDTH: i64x2.add
+; RELAXED-MAX-BANDWIDTH: i8x16.shuffle	4, 5, 6, 7, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1
+; RELAXED-MAX-BANDWIDTH: i32x4.extend_low_i16x8_s
+; RELAXED-MAX-BANDWIDTH: i8x16.shuffle	4, 5, 6, 7, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1
+; RELAXED-MAX-BANDWIDTH: i32x4.extend_low_i16x8_s
+; RELAXED-MAX-BANDWIDTH: i64x2.extmul_low_i32x4_s
+; RELAXED-MAX-BANDWIDTH: i64x2.add
+; RELAXED-MAX-BANDWIDTH: i32x4.extend_low_i16x8_s
+; RELAXED-MAX-BANDWIDTH: i32x4.extend_low_i16x8_s
+; RELAXED-MAX-BANDWIDTH: i64x2.extmul_low_i32x4_s
+; RELAXED-MAX-BANDWIDTH: i64x2.add
+
 entry:
   %cmp7.not = icmp eq i32 %N, 0
   br i1 %cmp7.not, label %for.cond.cleanup, label %for.body
@@ -156,6 +295,14 @@ define hidden i64 @i64_mac_s32(ptr nocapture noundef readonly %a, ptr nocapture
 ; MAX-BANDWIDTH: i64x2.extend_low_i32x4_s
 ; MAX-BANDWIDTH: i64x2.add
 
+; RELAXED-MAX-BANDWIDTH: v128.load
+; RELAXED-MAX-BANDWIDTH: v128.load
+; RELAXED-MAX-BANDWIDTH: i32x4.mul
+; RELAXED-MAX-BANDWIDTH: i64x2.extend_high_i32x4_s
+; RELAXED-MAX-BANDWIDTH: i64x2.add
+; RELAXED-MAX-BANDWIDTH: i64x2.extend_low_i32x4_s
+; RELAXED-MAX-BANDWIDTH: i64x2.add
+
 entry:
   %cmp6.not = icmp eq i32 %N, 0
   br i1 %cmp6.not, label %for.cond.cleanup, label %for.body
@@ -197,6 +344,15 @@ define hidden i32 @i32_mac_u8(ptr nocapture noundef readonly %a, ptr nocapture n
 ; MAX-BANDWIDTH: i32x4.add
 ; MAX-BANDWIDTH: i32x4.add
 
+; RELAXED-MAX-BANDWIDTH: v128.load
+; RELAXED-MAX-BANDWIDTH: v128.load
+; RELAXED-MAX-BANDWIDTH: i16x8.extmul_low_i8x16_u
+; RELAXED-MAX-BANDWIDTH: i32x4.extadd_pairwise_i16x8_u
+; RELAXED-MAX-BANDWIDTH: i16x8.extmul_high_i8x16_u
+; RELAXED-MAX-BANDWIDTH: i32x4.extadd_pairwise_i16x8_u
+; RELAXED-MAX-BANDWIDTH: i32x4.add
+; RELAXED-MAX-BANDWIDTH: i32x4.add
+
 entry:
   %cmp7.not = icmp eq i32 %N, 0
   br i1 %cmp7.not, label %for.cond.cleanup, label %for.body
@@ -235,6 +391,13 @@ define hidden i32 @i32_mac_u16(ptr nocapture noundef readonly %a, ptr nocapture
 ; MAX-BANDWIDTH: i32x4.add
 ; MAX-BANDWIDTH: i32x4.add
 
+; RELAXED-MAX-BANDWIDTH: v128.load
+; RELAXED-MAX-BANDWIDTH: v128.load
+; RELAXED-MAX-BANDWIDTH: i32x4.extmul_low_i16x8_u
+; RELAXED-MAX-BANDWIDTH: i32x4.extmul_high_i16x8_u
+; RELAXED-MAX-BANDWIDTH: i32x4.add
+; RELAXED-MAX-BANDWIDTH: i32x4.add
+
 entry:
   %cmp7.not = icmp eq i32 %N, 0
   br i1 %cmp7.not, label %for.cond.cleanup, label %for.body
@@ -277,6 +440,17 @@ define hidden i32 @i32_mac_u16_s16(ptr nocapture noundef readonly %a, ptr nocapt
 ; MAX-BANDWIDTH: i32x4.add
 ; MAX-BANDWIDTH: i32x4.add
 
+; RELAXED-MAX-BANDWIDTH: v128.load
+; RELAXED-MAX-BANDWIDTH: i32x4.extend_high_i16x8_s
+; RELAXED-MAX-BANDWIDTH: v128.load
+; RELAXED-MAX-BANDWIDTH: i32x4.extend_high_i16x8_u
+; RELAXED-MAX-BANDWIDTH: i32x4.mul
+; RELAXED-MAX-BANDWIDTH: i32x4.extend_low_i16x8_s
+; RELAXED-MAX-BANDWIDTH: i32x4.extend_low_i16x8_u
+; RELAXED-MAX-BANDWIDTH: i32x4.mul
+; RELAXED-MAX-BANDWIDTH: i32x4.add
+; RELAXED-MAX-BANDWIDTH: i32x4.add
+
 entry:
   %cmp7.not = icmp eq i32 %N, 0
   br i1 %cmp7.not, label %for.cond.cleanup, label %for.body
@@ -335,6 +509,32 @@ define hidden i64 @i64_mac_u16(ptr nocapture noundef readonly %a, ptr nocapture
 ; MAX-BANDWIDTH: i64x2.extmul_low_i32x4_u
 ; MAX-BANDWIDTH: i64x2.add
 
+; RELAXED-MAX-BANDWIDTH: v128.load
+; RELAXED-MAX-BANDWIDTH: i8x16.shuffle	12, 13, 14, 15, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1
+; RELAXED-MAX-BANDWIDTH: i32x4.extend_low_i16x8_u
+; RELAXED-MAX-BANDWIDTH: v128.load
+; RELAXED-MAX-BANDWIDTH: i8x16.shuffle	12, 13, 14, 15, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1
+; RELAXED-MAX-BANDWIDTH: i32x4.extend_low_i16x8_u
+; RELAXED-MAX-BANDWIDTH: i64x2.extmul_low_i32x4_u
+; RELAXED-MAX-BANDWIDTH: i64x2.add
+; RELAXED-MAX-BANDWIDTH: i8x16.shuffle	8, 9, 10, 11, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1
+; RELAXED-MAX-BANDWIDTH: i32x4.extend_low_i16x8_u
+; RELAXED-MAX-BANDWIDTH: i8x16.shuffle	8, 9, 10, 11, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1
+; RELAXED-MAX-BANDWIDTH: i32x4.extend_low_i16x8_u
+; RELAXED-MAX-BANDWIDTH: i64x2.extmul_low_i32x4_u
+; RELAXED-MAX-BANDWIDTH: i64x2.add
+; RELAXED-MAX-BANDWIDTH: i8x16.shuffle	4, 5, 6, 7, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1
+; RELAXED-MAX-BANDWIDTH: i32x4.extend_low_i16x8_u
+; RELAXED-MAX-BANDWIDTH: i8x16.shuffle	4, 5, 6, 7, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1
+; RELAXED-MAX-BANDWIDTH: i32x4.extend_low_i16x8_u
+; RELAXED-MAX-BANDWIDTH: i64x2.extmul_low_i32x4_u
+; RELAXED-MAX-BANDWIDTH: i64x2.add
+; RELAXED-MAX-BANDWIDTH: i32x4.extend_low_i16x8_u
+; RELAXED-MAX-BANDWIDTH: i32x4.extend_low_i16x8_u
+; RELAXED-MAX-BANDWIDTH: i64x2.extmul_low_i32x4_u
+; RELAXED-MAX-BANDWIDTH: i64x2.add
+
+
 entry:
   %cmp8.not = icmp eq i32 %N, 0
   br i1 %cmp8.not, label %for.cond.cleanup, label %for.body
@@ -375,6 +575,14 @@ define hidden i64 @i64_mac_u32(ptr nocapture noundef readonly %a, ptr nocapture
 ; MAX-BANDWIDTH: i64x2.extend_low_i32x4_u
 ; MAX-BANDWIDTH: i64x2.add
 
+; RELAXED-MAX-BANDWIDTH: v128.load
+; RELAXED-MAX-BANDWIDTH: v128.load
+; RELAXED-MAX-BANDWIDTH: i32x4.mul
+; RELAXED-MAX-BANDWIDTH: i64x2.extend_high_i32x4_u
+; RELAXED-MAX-BANDWIDTH: i64x2.add
+; RELAXED-MAX-BANDWIDTH: i64x2.extend_low_i32x4_u
+; RELAXED-MAX-BANDWIDTH: i64x2.add
+
 entry:
   %cmp6.not = icmp eq i32 %N, 0
   br i1 %cmp6.not, label %for.cond.cleanup, label %for.body

badumbatish
badumbatish approved these changes Oct 15, 2025
View reviewed changes
Copy link
Contributor

@badumbatish badumbatish left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM, ty

@sparker-arm sparker-arm merged commit 65363e6 into llvm:main Oct 16, 2025
12 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@dschuff dschuff dschuff approved these changes

@badumbatish badumbatish badumbatish approved these changes

Assignees

@sparker-arm sparker-arm

Labels

backend:WebAssembly

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

4 participants

@sparker-arm @llvmbot @dschuff @badumbatish