Skip to content

[llvm][AArch64] Improve the cost model for i128 div's #107306

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 5 commits into from
Sep 5, 2024
Merged

[llvm][AArch64] Improve the cost model for i128 div's #107306

merged 5 commits into from
Sep 5, 2024

Conversation

jroelofs
Copy link
Contributor

@jroelofs jroelofs commented Sep 4, 2024 •
edited
Loading

No description provided.

@jroelofs
[llvm][AArch64] Pre-commit tests for [us]div i128 cost model improvem…
7bf4600 
…ents
@jroelofs
[llvm][AArch64] Improve the cost model for i128 div's
750ad33 
They're implemented via libcalls, which I've microbenchmarked to take about 14
and 16 cycles for udiv and sdiv i128 respectively on this Apple M2 Max.
@jroelofs jroelofs requested review from fhahn and davemgreen September 4, 2024 20:28
@llvmbot
Copy link
Member

llvmbot commented Sep 4, 2024 •
edited
Loading

@llvm/pr-subscribers-backend-aarch64

@llvm/pr-subscribers-llvm-analysis

Author: Jon Roelofs (jroelofs)

Changes

They're implemented via libcalls, which I've microbenchmarked to take about 14 and 16 cycles for udiv and sdiv i128 respectively on this Apple M2 Max.

Patch is 20.12 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/107306.diff

2 Files Affected:

  • (modified) llvm/lib/Target/AArch64/AArch64TargetTransformInfo.cpp (+6)
  • (modified) llvm/test/Analysis/CostModel/AArch64/div.ll (+43)
diff --git a/llvm/lib/Target/AArch64/AArch64TargetTransformInfo.cpp b/llvm/lib/Target/AArch64/AArch64TargetTransformInfo.cpp
index 58c267f1ce4bd6..a94bba6150494f 100644
--- a/llvm/lib/Target/AArch64/AArch64TargetTransformInfo.cpp
+++ b/llvm/lib/Target/AArch64/AArch64TargetTransformInfo.cpp
@@ -3222,6 +3222,8 @@ InstructionCost AArch64TTIImpl::getArithmeticInstrCost(
                                      Op1Info.getNoProps(), Op2Info.getNoProps());
       return Cost;
     }
+    if (TLI->getValueType(DL, Ty) == MVT::i128)
+      return 16;
     [[fallthrough]];
   case ISD::UDIV: {
     if (Op2Info.isConstant() && Op2Info.isUniform()) {
@@ -3240,6 +3242,10 @@ InstructionCost AArch64TTIImpl::getArithmeticInstrCost(
       }
     }
 
+    auto VT = TLI->getValueType(DL, Ty);
+    if (VT == MVT::i128)
+      return 14;
+
     InstructionCost Cost = BaseT::getArithmeticInstrCost(
         Opcode, Ty, CostKind, Op1Info, Op2Info);
     if (Ty->isVectorTy()) {
diff --git a/llvm/test/Analysis/CostModel/AArch64/div.ll b/llvm/test/Analysis/CostModel/AArch64/div.ll
index 7b52e2dc18aa52..9c23f5f53f1028 100644
--- a/llvm/test/Analysis/CostModel/AArch64/div.ll
+++ b/llvm/test/Analysis/CostModel/AArch64/div.ll
@@ -5,6 +5,7 @@ target datalayout = "e-m:e-i8:8:32-i16:16:32-i64:64-i128:128-n32:64-S128"
 
 define i32 @sdiv() {
 ; CHECK-LABEL: 'sdiv'
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %I128 = sdiv i128 undef, undef
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = sdiv i64 undef, undef
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %V2i64 = sdiv <2 x i64> undef, undef
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 56 for instruction: %V4i64 = sdiv <4 x i64> undef, undef
@@ -23,6 +24,8 @@ define i32 @sdiv() {
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 672 for instruction: %V64i8 = sdiv <64 x i8> undef, undef
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
+  %I128 = sdiv i128 undef, undef
+
   %I64 = sdiv i64 undef, undef
   %V2i64 = sdiv <2 x i64> undef, undef
   %V4i64 = sdiv <4 x i64> undef, undef
@@ -48,6 +51,7 @@ define i32 @sdiv() {
 
 define i32 @udiv() {
 ; CHECK-LABEL: 'udiv'
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %I128 = udiv i128 undef, undef
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %I64 = udiv i64 undef, undef
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %V2i64 = udiv <2 x i64> undef, undef
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 56 for instruction: %V4i64 = udiv <4 x i64> undef, undef
@@ -66,6 +70,8 @@ define i32 @udiv() {
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 672 for instruction: %V64i8 = udiv <64 x i8> undef, undef
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
+  %I128 = udiv i128 undef, undef
+
   %I64 = udiv i64 undef, undef
   %V2i64 = udiv <2 x i64> undef, undef
   %V4i64 = udiv <4 x i64> undef, undef
@@ -91,6 +97,7 @@ define i32 @udiv() {
 
 define i32 @sdiv_const() {
 ; CHECK-LABEL: 'sdiv_const'
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %I128 = sdiv i128 undef, 7
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %I64 = sdiv i64 undef, 7
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %V2i64 = sdiv <2 x i64> undef, <i64 6, i64 7>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 56 for instruction: %V4i64 = sdiv <4 x i64> undef, <i64 4, i64 5, i64 6, i64 7>
@@ -109,6 +116,8 @@ define i32 @sdiv_const() {
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 672 for instruction: %V64i8 = sdiv <64 x i8> undef, <i8 4, i8 5, i8 6, i8 7, i8 8, i8 9, i8 10, i8 11, i8 12, i8 13, i8 14, i8 15, i8 16, i8 17, i8 18, i8 19, i8 4, i8 5, i8 6, i8 7, i8 8, i8 9, i8 10, i8 11, i8 12, i8 13, i8 14, i8 15, i8 16, i8 17, i8 18, i8 19, i8 4, i8 5, i8 6, i8 7, i8 8, i8 9, i8 10, i8 11, i8 12, i8 13, i8 14, i8 15, i8 16, i8 17, i8 18, i8 19, i8 4, i8 5, i8 6, i8 7, i8 8, i8 9, i8 10, i8 11, i8 12, i8 13, i8 14, i8 15, i8 16, i8 17, i8 18, i8 19>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
+  %I128 = sdiv i128 undef, 7
+
   %I64 = sdiv i64 undef, 7
   %V2i64 = sdiv <2 x i64> undef, <i64 6, i64 7>
   %V4i64 = sdiv <4 x i64> undef, <i64 4, i64 5, i64 6, i64 7>
@@ -134,6 +143,7 @@ define i32 @sdiv_const() {
 
 define i32 @udiv_const() {
 ; CHECK-LABEL: 'udiv_const'
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %I128 = udiv i128 undef, 7
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %I64 = udiv i64 undef, 7
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %V2i64 = udiv <2 x i64> undef, <i64 6, i64 7>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 56 for instruction: %V4i64 = udiv <4 x i64> undef, <i64 4, i64 5, i64 6, i64 7>
@@ -152,6 +162,9 @@ define i32 @udiv_const() {
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 672 for instruction: %V64i8 = udiv <64 x i8> undef, <i8 4, i8 5, i8 6, i8 7, i8 8, i8 9, i8 10, i8 11, i8 12, i8 13, i8 14, i8 15, i8 16, i8 17, i8 18, i8 19, i8 4, i8 5, i8 6, i8 7, i8 8, i8 9, i8 10, i8 11, i8 12, i8 13, i8 14, i8 15, i8 16, i8 17, i8 18, i8 19, i8 4, i8 5, i8 6, i8 7, i8 8, i8 9, i8 10, i8 11, i8 12, i8 13, i8 14, i8 15, i8 16, i8 17, i8 18, i8 19, i8 4, i8 5, i8 6, i8 7, i8 8, i8 9, i8 10, i8 11, i8 12, i8 13, i8 14, i8 15, i8 16, i8 17, i8 18, i8 19>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
+
+  %I128 = udiv i128 undef, 7
+
   %I64 = udiv i64 undef, 7
   %V2i64 = udiv <2 x i64> undef, <i64 6, i64 7>
   %V4i64 = udiv <4 x i64> undef, <i64 4, i64 5, i64 6, i64 7>
@@ -177,6 +190,7 @@ define i32 @udiv_const() {
 
 define i32 @sdiv_uniformconst() {
 ; CHECK-LABEL: 'sdiv_uniformconst'
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %I128 = sdiv i128 undef, 7
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %I64 = sdiv i64 undef, 7
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V2i64 = sdiv <2 x i64> undef, <i64 7, i64 7>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V4i64 = sdiv <4 x i64> undef, <i64 7, i64 7, i64 7, i64 7>
@@ -195,6 +209,8 @@ define i32 @sdiv_uniformconst() {
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 320 for instruction: %V64i8 = sdiv <64 x i8> undef, <i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
+  %I128 = sdiv i128 undef, 7
+
   %I64 = sdiv i64 undef, 7
   %V2i64 = sdiv <2 x i64> undef, <i64 7, i64 7>
   %V4i64 = sdiv <4 x i64> undef, <i64 7, i64 7, i64 7, i64 7>
@@ -220,6 +236,7 @@ define i32 @sdiv_uniformconst() {
 
 define i32 @udiv_uniformconst() {
 ; CHECK-LABEL: 'udiv_uniformconst'
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %I128 = udiv i128 undef, 7
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %I64 = udiv i64 undef, 7
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V2i64 = udiv <2 x i64> undef, <i64 7, i64 7>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V4i64 = udiv <4 x i64> undef, <i64 7, i64 7, i64 7, i64 7>
@@ -238,6 +255,8 @@ define i32 @udiv_uniformconst() {
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 320 for instruction: %V64i8 = udiv <64 x i8> undef, <i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
+  %I128 = udiv i128 undef, 7
+
   %I64 = udiv i64 undef, 7
   %V2i64 = udiv <2 x i64> undef, <i64 7, i64 7>
   %V4i64 = udiv <4 x i64> undef, <i64 7, i64 7, i64 7, i64 7>
@@ -263,6 +282,7 @@ define i32 @udiv_uniformconst() {
 
 define i32 @sdiv_constpow2() {
 ; CHECK-LABEL: 'sdiv_constpow2'
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %I128 = sdiv i128 undef, 16
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %I64 = sdiv i64 undef, 16
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %V2i64 = sdiv <2 x i64> undef, <i64 8, i64 16>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 56 for instruction: %V4i64 = sdiv <4 x i64> undef, <i64 2, i64 4, i64 8, i64 16>
@@ -281,6 +301,8 @@ define i32 @sdiv_constpow2() {
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 672 for instruction: %V64i8 = sdiv <64 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
+  %I128 = sdiv i128 undef, 16
+
   %I64 = sdiv i64 undef, 16
   %V2i64 = sdiv <2 x i64> undef, <i64 8, i64 16>
   %V4i64 = sdiv <4 x i64> undef, <i64 2, i64 4, i64 8, i64 16>
@@ -306,6 +328,7 @@ define i32 @sdiv_constpow2() {
 
 define i32 @udiv_constpow2() {
 ; CHECK-LABEL: 'udiv_constpow2'
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %I128 = udiv i128 undef, 16
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %I64 = udiv i64 undef, 16
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %V2i64 = udiv <2 x i64> undef, <i64 8, i64 16>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 56 for instruction: %V4i64 = udiv <4 x i64> undef, <i64 2, i64 4, i64 8, i64 16>
@@ -324,6 +347,8 @@ define i32 @udiv_constpow2() {
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 672 for instruction: %V64i8 = udiv <64 x i8> undef, <i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16, i8 2, i8 4, i8 8, i8 16>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
+  %I128 = udiv i128 undef, 16
+
   %I64 = udiv i64 undef, 16
   %V2i64 = udiv <2 x i64> undef, <i64 8, i64 16>
   %V4i64 = udiv <4 x i64> undef, <i64 2, i64 4, i64 8, i64 16>
@@ -349,6 +374,7 @@ define i32 @udiv_constpow2() {
 
 define i32 @sdiv_uniformconstpow2() {
 ; CHECK-LABEL: 'sdiv_uniformconstpow2'
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %I128 = sdiv i128 undef, 16
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %I64 = sdiv i64 undef, 16
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %V2i64 = sdiv <2 x i64> undef, <i64 16, i64 16>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 30 for instruction: %V4i64 = sdiv <4 x i64> undef, <i64 16, i64 16, i64 16, i64 16>
@@ -367,6 +393,8 @@ define i32 @sdiv_uniformconstpow2() {
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 396 for instruction: %V64i8 = sdiv <64 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
+  %I128 = sdiv i128 undef, 16
+
   %I64 = sdiv i64 undef, 16
   %V2i64 = sdiv <2 x i64> undef, <i64 16, i64 16>
   %V4i64 = sdiv <4 x i64> undef, <i64 16, i64 16, i64 16, i64 16>
@@ -392,6 +420,7 @@ define i32 @sdiv_uniformconstpow2() {
 
 define i32 @udiv_uniformconstpow2() {
 ; CHECK-LABEL: 'udiv_uniformconstpow2'
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %I128 = udiv i128 undef, 16
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %I64 = udiv i64 undef, 16
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V2i64 = udiv <2 x i64> undef, <i64 16, i64 16>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V4i64 = udiv <4 x i64> undef, <i64 16, i64 16, i64 16, i64 16>
@@ -410,6 +439,8 @@ define i32 @udiv_uniformconstpow2() {
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 320 for instruction: %V64i8 = udiv <64 x i8> undef, <i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16, i8 16>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
+  %I128 = udiv i128 undef, 16
+
   %I64 = udiv i64 undef, 16
   %V2i64 = udiv <2 x i64> undef, <i64 16, i64 16>
   %V4i64 = udiv <4 x i64> undef, <i64 16, i64 16, i64 16, i64 16>
@@ -435,6 +466,7 @@ define i32 @udiv_uniformconstpow2() {
 
 define i32 @sdiv_constnegpow2() {
 ; CHECK-LABEL: 'sdiv_constnegpow2'
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %I128 = sdiv i128 undef, -16
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %I64 = sdiv i64 undef, -16
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %V2i64 = sdiv <2 x i64> undef, <i64 -8, i64 -16>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 56 for instruction: %V4i64 = sdiv <4 x i64> undef, <i64 -2, i64 -4, i64 -8, i64 -16>
@@ -453,6 +485,8 @@ define i32 @sdiv_constnegpow2() {
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 672 for instruction: %V64i8 = sdiv <64 x i8> undef, <i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
+  %I128 = sdiv i128 undef, -16
+
   %I64 = sdiv i64 undef, -16
   %V2i64 = sdiv <2 x i64> undef, <i64 -8, i64 -16>
   %V4i64 = sdiv <4 x i64> undef, <i64 -2, i64 -4, i64 -8, i64 -16>
@@ -478,6 +512,7 @@ define i32 @sdiv_constnegpow2() {
 
 define i32 @udiv_constnegpow2() {
 ; CHECK-LABEL: 'udiv_constnegpow2'
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %I128 = udiv i128 undef, -16
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %I64 = udiv i64 undef, -16
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %V2i64 = udiv <2 x i64> undef, <i64 -8, i64 -16>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 56 for instruction: %V4i64 = udiv <4 x i64> undef, <i64 -2, i64 -4, i64 -8, i64 -16>
@@ -496,6 +531,8 @@ define i32 @udiv_constnegpow2() {
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 672 for instruction: %V64i8 = udiv <64 x i8> undef, <i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16, i8 -2, i8 -4, i8 -8, i8 -16>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
+  %I128 = udiv i128 undef, -16
+
   %I64 = udiv i64 undef, -16
   %V2i64 = udiv <2 x i64> undef, <i64 -8, i64 -16>
   %V4i64 = udiv <4 x i64> undef, <i64 -2, i64 -4, i64 -8, i64 -16>
@@ -521,6 +558,7 @@ define i32 @udiv_constnegpow2() {
 
 define i32 @sdiv_uniformconstnegpow2() {
 ; CHECK-LABEL: 'sdiv_uniformconstnegpow2'
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %I128 = sdiv i128 undef, -16
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %I64 = sdiv i64 undef, -16
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V2i64 = sdiv <2 x i64> undef, <i64 -16, i64 -16>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V4i64 = sdiv <4 x i64> undef, <i64 -16, i64 -16, i64 -16, i64 -16>
@@ -539,6 +577,8 @@ define i32 @sdiv_uniformconstnegpow2() {
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 320 for instruction: %V64i8 = sdiv <64 x i8> undef, <i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
+  %I128 = sdiv i128 undef, -16
+
   %I64 = sdiv i64 undef, -16
   %V2i64 = sdiv <2 x i64> undef, <i64 -16, i64 -16>
   %V4i64 = sdiv <4 x i64> undef, <i64 -16, i64 -16, i64 -16, i64 -16>
@@ -564,6 +604,7 @@ define i32 @sdiv_uniformconstnegpow2() {
 
 define i32 @udiv_uniformconstnegpow2() {
 ; CHECK-LABEL: 'udiv_uniformconstnegpow2'
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 14 for instruction: %I128 = udiv i128 undef, -16
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %I64 = udiv i64 undef, -16
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V2i64 = udiv <2 x i64> undef, <i64 -16, i64 -16>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V4i64 = udiv <4 x i64> undef, <i64 -16, i64 -16, i64 -16, i64 -16>
@@ -582,6 +623,8 @@ define i32 @udiv_uniformconstnegpow2() {
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 320 for instruction: %V64i8 = udiv <64 x i8> undef, <i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16, i8 -16>
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
+  %I128 = udiv i128 undef, -16
+
   %I64 = udiv i64 undef, -1...
[truncated]

@github-actions GitHub Actions
Copy link

github-actions bot commented Sep 4, 2024 •
edited
Loading

✅ With the latest revision this PR passed the C/C++ code formatter.

davemgreen
davemgreen reviewed Sep 4, 2024
View reviewed changes
Copy link
Collaborator

@davemgreen davemgreen left a comment

Choose a reason for hiding this comment

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

We tend to use 10 as the cost of a libcall. Does what you are doing need to be more precise than that?

It might be worth using size > 64. For larger sizes the codegen is horrendous, but giving it a higher cost sounds at least more correct! It could be 10 * type legalization cost or something like it.

Otherwise this sounds good to me. Let me know what you think.

@jroelofs
Copy link
Contributor Author

jroelofs commented Sep 5, 2024

I don't need precision for the case I'm trying to avoid: the vectorizer thought this was way too cheap and emitted one in a runtime check. So... the 10 from getCallInstCost sounds good to me.

davemgreen
davemgreen approved these changes Sep 5, 2024
View reviewed changes
Copy link
Collaborator

@davemgreen davemgreen left a comment

Choose a reason for hiding this comment

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

Thanks, LGTM. Its good to hear that the the call overhead and the call roughly match up in this case.

@jroelofs jroelofs merged commit bded3b3 into llvm:main Sep 5, 2024
8 checks passed
@jroelofs jroelofs deleted the jroelofs/rdar133996027-udiv128-cost branch September 5, 2024 14:42
@llvm-ci
Copy link
Collaborator

llvm-ci commented Sep 5, 2024

LLVM Buildbot has detected a new failure on builder lldb-x86_64-debian running on lldb-x86_64-debian while building llvm at step 6 "test".

Full details are available at: https://lab.llvm.org/buildbot/#/builders/162/builds/5624

Here is the relevant piece of the build log for the reference 
Step 6 (test) failure: build (failure)
...
UNSUPPORTED: lldb-shell :: ScriptInterpreter/Lua/breakpoint_callback.test (2657 of 2666)
UNSUPPORTED: lldb-shell :: ScriptInterpreter/Lua/lua-python.test (2658 of 2666)
UNSUPPORTED: lldb-shell :: ScriptInterpreter/Lua/nested_sessions.test (2659 of 2666)
UNSUPPORTED: lldb-shell :: ScriptInterpreter/Python/Crashlog/interactive_crashlog_legacy.test (2660 of 2666)
PASS: lldb-api :: api/multithreaded/TestMultithreaded.py (2661 of 2666)
PASS: lldb-api :: tools/lldb-server/TestGdbRemoteFork.py (2662 of 2666)
PASS: lldb-api :: tools/lldb-server/TestLldbGdbServer.py (2663 of 2666)
PASS: lldb-api :: commands/process/attach/TestProcessAttach.py (2664 of 2666)
PASS: lldb-api :: repl/clang/TestClangREPL.py (2665 of 2666)
UNRESOLVED: lldb-api :: iohandler/sigint/TestIOHandlerPythonREPLSigint.py (2666 of 2666)
******************** TEST 'lldb-api :: iohandler/sigint/TestIOHandlerPythonREPLSigint.py' FAILED ********************
Script:
--
/usr/bin/python3 /home/worker/2.0.1/lldb-x86_64-debian/llvm-project/lldb/test/API/dotest.py -u CXXFLAGS -u CFLAGS --env ARCHIVER=/usr/bin/ar --env OBJCOPY=/usr/bin/objcopy --env LLVM_LIBS_DIR=/home/worker/2.0.1/lldb-x86_64-debian/build/./lib --env LLVM_INCLUDE_DIR=/home/worker/2.0.1/lldb-x86_64-debian/build/include --env LLVM_TOOLS_DIR=/home/worker/2.0.1/lldb-x86_64-debian/build/./bin --arch x86_64 --build-dir /home/worker/2.0.1/lldb-x86_64-debian/build/lldb-test-build.noindex --lldb-module-cache-dir /home/worker/2.0.1/lldb-x86_64-debian/build/lldb-test-build.noindex/module-cache-lldb/lldb-api --clang-module-cache-dir /home/worker/2.0.1/lldb-x86_64-debian/build/lldb-test-build.noindex/module-cache-clang/lldb-api --executable /home/worker/2.0.1/lldb-x86_64-debian/build/./bin/lldb --compiler /home/worker/2.0.1/lldb-x86_64-debian/build/./bin/clang --dsymutil /home/worker/2.0.1/lldb-x86_64-debian/build/./bin/dsymutil --llvm-tools-dir /home/worker/2.0.1/lldb-x86_64-debian/build/./bin --lldb-obj-root /home/worker/2.0.1/lldb-x86_64-debian/build/tools/lldb --lldb-libs-dir /home/worker/2.0.1/lldb-x86_64-debian/build/./lib -t /home/worker/2.0.1/lldb-x86_64-debian/llvm-project/lldb/test/API/iohandler/sigint -p TestIOHandlerPythonREPLSigint.py
--
Exit Code: 1

Command Output (stdout):
--
lldb version 20.0.0git (https://github.com/llvm/llvm-project.git revision bded3b3ea9f78c5b3edc3d4a6076665af0ea746b)
  clang revision bded3b3ea9f78c5b3edc3d4a6076665af0ea746b
  llvm revision bded3b3ea9f78c5b3edc3d4a6076665af0ea746b
(lldb) settings clear -all
(lldb) settings set symbols.enable-external-lookup false
(lldb) settings set target.inherit-tcc true
(lldb) settings set target.disable-aslr false
(lldb) settings set target.detach-on-error false
(lldb) settings set target.auto-apply-fixits false
(lldb) settings set plugin.process.gdb-remote.packet-timeout 60
(lldb) settings set symbols.clang-modules-cache-path "/home/worker/2.0.1/lldb-x86_64-debian/build/lldb-test-build.noindex/module-cache-lldb/lldb-api"
(lldb) settings set use-color false
(lldb) script -l python --
script -l python --
Python Interactive Interpreter. To exit, type 'quit()', 'exit()' or Ctrl-D.
>>> import time; print('running' + 'now'); time.sleep(10000);
import time; print('running' + 'now'); time.sleep(10000);
runningnow�
^Csettings set interpreter.prompt-on-quit false
quit
settings set interpreter.prompt-on-quit false^Jquit^JSkipping the following test categories: ['libc++', 'dsym', 'gmodules', 'debugserver', 'objc']

--
Command Output (stderr):
--
Change dir to: /home/worker/2.0.1/lldb-x86_64-debian/llvm-project/lldb/test/API/iohandler/sigint
runCmd: settings clear -all

output:

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@davemgreen davemgreen davemgreen approved these changes

@fhahn fhahn Awaiting requested review from fhahn

Assignees
No one assigned
Labels
backend:AArch64 llvm:analysis
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
4 participants
@jroelofs @llvmbot @llvm-ci @davemgreen