[CalcSpillWeights] Avoid x87 excess precision influencing weight result #100165
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
|
@llvm/pr-subscribers-llvm-support @llvm/pr-subscribers-backend-x86 Author: Dimitry Andric (DimitryAndric) ChangesFixes #99396 The result of Similar to 7af3432, add a Full diff: https://github.com/llvm/llvm-project/pull/100165.diff 2 Files Affected:
diff --git a/llvm/lib/CodeGen/CalcSpillWeights.cpp b/llvm/lib/CodeGen/CalcSpillWeights.cpp
index 1d767a3484bca..60055e6ba0cd3 100644
--- a/llvm/lib/CodeGen/CalcSpillWeights.cpp
+++ b/llvm/lib/CodeGen/CalcSpillWeights.cpp
@@ -257,7 +257,9 @@ float VirtRegAuxInfo::weightCalcHelper(LiveInterval &LI, SlotIndex *Start,
return -1.0f;
}
- float Weight = 1.0f;
+ // Force Weight onto the stack so that x86 doesn't add hidden precision,
+ // similar to HWeight below.
+ volatile float Weight = 1.0f;
if (IsSpillable) {
// Get loop info for mi.
if (MI->getParent() != MBB) {
diff --git a/llvm/test/CodeGen/X86/pr99396.ll b/llvm/test/CodeGen/X86/pr99396.ll
new file mode 100644
index 0000000000000..7bc1153b57990
--- /dev/null
+++ b/llvm/test/CodeGen/X86/pr99396.ll
@@ -0,0 +1,59 @@
+; RUN: llc < %s -mtriple=i386-unknown-freebsd -enable-misched -relocation-model=pic | FileCheck %s
+
+target datalayout = "e-m:e-p:32:32-p270:32:32-p271:32:32-p272:64:64-i128:128-f64:32:64-f80:32-n8:16:32-S128"
+target triple = "i386-unknown-freebsd15.0"
+
+@c = external local_unnamed_addr global ptr
+
+declare i32 @fn2() local_unnamed_addr
+
+declare i32 @fn3() local_unnamed_addr
+
+define noundef i32 @fn4() #0 {
+entry:
+ %0 = load i32, ptr @fn4, align 4
+; CHECK: movl fn4@GOT(%ebx), %edi
+; CHECK-NEXT: movl (%edi), %edx
+ %1 = load ptr, ptr @c, align 4
+; CHECK: movl c@GOT(%ebx), %eax
+; CHECK-NEXT: movl (%eax), %esi
+; CHECK-NEXT: testl %esi, %esi
+ %cmp.g = icmp eq ptr %1, null
+ br i1 %cmp.g, label %if.then.g, label %if.end3.g
+
+if.then.g: ; preds = %entry
+ %2 = load i32, ptr inttoptr (i32 1 to ptr), align 4
+ %cmp1.g = icmp slt i32 %2, 0
+ br i1 %cmp1.g, label %if.then2.g, label %if.end3.g
+
+if.then2.g: ; preds = %if.then.g
+ %.g = load volatile i32, ptr null, align 2147483648
+ br label %f.exit
+
+if.end3.g: ; preds = %if.then.g, %entry
+ %h.i.g = icmp eq i32 %0, 0
+ br i1 %h.i.g, label %f.exit, label %while.body.g
+
+while.body.g: ; preds = %if.end3.g, %if.end8.g
+ %buff.addr.019.g = phi ptr [ %incdec.ptr.g, %if.end8.g ], [ @fn4, %if.end3.g ]
+ %g.addr.018.g = phi i32 [ %dec.g, %if.end8.g ], [ %0, %if.end3.g ]
+ %call4.g = tail call i32 @fn3(ptr %1, ptr %buff.addr.019.g, i32 %g.addr.018.g)
+ %cmp5.g = icmp slt i32 %call4.g, 0
+ br i1 %cmp5.g, label %if.then6.g, label %if.end8.g
+
+if.then6.g: ; preds = %while.body.g
+ %call7.g = tail call i32 @fn2(ptr null)
+ br label %f.exit
+
+if.end8.g: ; preds = %while.body.g
+ %dec.g = add i32 %g.addr.018.g, 1
+ %incdec.ptr.g = getelementptr i32, ptr %buff.addr.019.g, i32 1
+ store i64 0, ptr %1, align 4
+ %h.not.g = icmp eq i32 %dec.g, 0
+ br i1 %h.not.g, label %f.exit, label %while.body.g
+
+f.exit: ; preds = %if.end8.g, %if.then6.g, %if.end3.g, %if.then2.g
+ ret i32 0
+}
+
+attributes #0 = { "frame-pointer"="all" "tune-cpu"="generic" }
|
arsenm
reviewed
Jul 23, 2024
llvm/test/CodeGen/X86/pr99396.ll
Outdated
| %0 = load i32, ptr @fn4, align 4 | ||
| ; CHECK: movl fn4@GOT(%ebx), %edi | ||
| ; CHECK-NEXT: movl (%edi), %edx | ||
| %1 = load ptr, ptr @c, align 4 |
There was a problem hiding this comment.
Sorry, I haven't got much experience with writing these tests, what does this mean? I'm intending here to check the exact registers used.
There was a problem hiding this comment.
%0 should be %load or some other name etc.
| float Weight = 1.0f; | ||
| // Force Weight onto the stack so that x86 doesn't add hidden precision, | ||
| // similar to HWeight below. | ||
| volatile float Weight = 1.0f; |
There was a problem hiding this comment.
Can we wrap this in a macro in Support somewhere to only do this if using degenerate x87 handling?
There was a problem hiding this comment.
Sure, maybe a specific stack_float_t type?
DimitryAndric
force-pushed
the
fix-pr99396-1
branch
from
c969817 to
5071ae3
Compare
arsenm
reviewed
Jul 25, 2024
llvm/test/CodeGen/X86/pr99396.ll
Outdated
Comment on lines
3
to
4
| target datalayout = "e-m:e-p:32:32-p270:32:32-p271:32:32-p272:64:64-i128:128-f64:32:64-f80:32-n8:16:32-S128" | ||
| target triple = "i386-unknown-freebsd15.0" |
There was a problem hiding this comment.
These are redundant with the command line arguments
| @@ -770,6 +770,10 @@ std::enable_if_t<std::is_signed_v<T>, T> MulOverflow(T X, T Y, T &Result) { | |||
| #endif | |||
| } | |||
|
|
|||
| /// Type to force float point values onto the stack, so that x86 doesn't add | |||
| /// hidden precision, avoiding rounding differences on various platforms. | |||
| using stack_float_t = volatile float; | |||
There was a problem hiding this comment.
The only reason this would go in support would be to conditionalize this on using the volatile only when x87 is in use
Fixes llvm#99396 The result of `VirtRegAuxInfo::weightCalcHelper` can be influenced by x87 excess precision, which can result in slightly different register choices when the compiler is hosted on x86_64 or i386. This leads to different object file output when cross-compiling to i386, or native. Similar to 7af3432, we need to add a `volatile` qualifier to the local `Weight` variable to force it onto the stack, and avoid the excess precision. Define `stack_float_t` in `MathExtras.h` for this purpose, and use it.
DimitryAndric
force-pushed
the
fix-pr99396-1
branch
from
5071ae3 to
2899738
Compare
arsenm
approved these changes
Jul 25, 2024
| @@ -770,6 +770,14 @@ std::enable_if_t<std::is_signed_v<T>, T> MulOverflow(T X, T Y, T &Result) { | |||
| #endif | |||
| } | |||
|
|
|||
| /// Type to force float point values onto the stack, so that x86 doesn't add | |||
| /// hidden precision, avoiding rounding differences on various platforms. | |||
| #if defined(__i386__) || defined(_M_IX86) | |||
There was a problem hiding this comment.
Not sure this is the exact condition but ok
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
Fixes #99396
The result of
VirtRegAuxInfo::weightCalcHelpercan be influenced by x87 excess precision, which can result in slightly different register choices when the compiler is hosted on x86_64 or i386. This leads to different object file output when cross-compiling to i386, or native.Similar to 7af3432, add a
volatilequalifier to the localWeightvariable to force it onto the stack, and avoid the excess precision.