[LLVMbugs] [Bug 14361] New: optimizer produces incorrect code for simple arithmetic with i64 types
bugzilla-daemon at llvm.org
bugzilla-daemon at llvm.org
Fri Nov 16 07:24:27 PST 2012
http://llvm.org/bugs/show_bug.cgi?id=14361
Bug #: 14361
Summary: optimizer produces incorrect code for simple
arithmetic with i64 types
Product: new-bugs
Version: 3.1
Platform: All
OS/Version: All
Status: NEW
Severity: release blocker
Priority: P
Component: new bugs
AssignedTo: unassignedbugs at nondot.org
ReportedBy: vtjnash at gmail.com
CC: llvmbugs at cs.uiuc.edu
Classification: Unclassified
The LLVM optimizer (createInstructionCombiningPass?) incorrectly transforms
a+b == b+a
into
a == b
when a and b are i64 types. I am on a x86_64 machine.
(it correctly transforms that statement into true with smaller types)
---------------
$ clang -O0 opt-test.c
$ ./a.out
commutative? true
$ clang -O1 opt-test.c
$ ./a.out
commutative? false
$ cat opt-test.c
#include <stdio.h>
#include <stdint.h>
int __attribute__ ((noinline)) f(long long a, long long b) {
return a+b == b+a;
}
int main() {
printf("commutative? %s\n", (f(2,3) ? "true" : "false"));
}
$ clang --version
Apple clang version 3.1 (tags/Apple/clang-318.0.61) (based on LLVM 3.1svn)
Target: x86_64-apple-darwin11.3.0
Thread model: posix
--------------
(as mentioned in
https://github.com/JuliaLang/julia/issues/1555#issuecomment-10438267)
Before optimization:
define i1 @f(i64, i64) {
top:
%a = alloca i64, !dbg !5697
%b = alloca i64, !dbg !5697
store i64 %0, i64* %a, !dbg !5697
store i64 %1, i64* %b, !dbg !5697
%2 = load i64* %a, !dbg !5704
%3 = load i64* %b, !dbg !5704
%4 = add i64 %2, %3, !dbg !5704
%5 = load i64* %b, !dbg !5704
%6 = load i64* %a, !dbg !5704
%7 = add i64 %5, %6, !dbg !5704
%8 = icmp eq i64 %4, %7, !dbg !5704
ret i1 %8, !dbg !5704
}
After optimization:
define i1 @f(i64, i64) {
top:
%2 = icmp eq i64 %1, %0, !dbg !5697
ret i1 %2, !dbg !5697
}
--------------
--
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
More information about the llvm-bugs
mailing list