<html>
<head>
<base href="https://llvm.org/bugs/" />
</head>
<body><table border="1" cellspacing="0" cellpadding="8">
<tr>
<th>Bug ID</th>
<td><a class="bz_bug_link
bz_status_NEW "
title="NEW --- - Failure to hoist constant out of loop"
href="https://llvm.org/bugs/show_bug.cgi?id=27136">27136</a>
</td>
</tr>
<tr>
<th>Summary</th>
<td>Failure to hoist constant out of loop
</td>
</tr>
<tr>
<th>Product</th>
<td>libraries
</td>
</tr>
<tr>
<th>Version</th>
<td>trunk
</td>
</tr>
<tr>
<th>Hardware</th>
<td>PC
</td>
</tr>
<tr>
<th>OS</th>
<td>All
</td>
</tr>
<tr>
<th>Status</th>
<td>NEW
</td>
</tr>
<tr>
<th>Severity</th>
<td>normal
</td>
</tr>
<tr>
<th>Priority</th>
<td>P
</td>
</tr>
<tr>
<th>Component</th>
<td>Backend: X86
</td>
</tr>
<tr>
<th>Assignee</th>
<td>unassignedbugs@nondot.org
</td>
</tr>
<tr>
<th>Reporter</th>
<td>llvm-dev@redking.me.uk
</td>
</tr>
<tr>
<th>CC</th>
<td>chisophugis@gmail.com, llvm-bugs@lists.llvm.org, spatel+llvm@rotateright.com
</td>
</tr>
<tr>
<th>Classification</th>
<td>Unclassified
</td>
</tr></table>
<p>
<div>
<pre>Poor code gen identified in Andreas Fredriksson's GDC 2016 Talk 'Taming the
Jaguar: x86 Optimization at Insomniac Games':
<a href="http://schedule.gdconf.com/session/taming-the-jaguar-x86-optimization-at-insomniac-games">http://schedule.gdconf.com/session/taming-the-jaguar-x86-optimization-at-insomniac-games</a>
search.c:
#include <stdint.h>
#include <stdbool.h>
bool search(uint32_t needle, const uint32_t *haystack, int count) {
for (int i = 0; i < count; ++i)
if (needle == haystack[i])
return true;
return false;
}
clang -O3 -S -march=btver2 search.c -o -
_search:
.cfi_startproc
## BB#0:
pushq %rbp
Ltmp0:
.cfi_def_cfa_offset 16
Ltmp1:
.cfi_offset %rbp, -16
movq %rsp, %rbp
Ltmp2:
.cfi_def_cfa_register %rbp
testl %edx, %edx
jle LBB0_1
## BB#4:
movslq %edx, %rcx
xorl %edx, %edx
.p2align 4, 0x90
LBB0_5:
cmpl %edi, (%rsi,%rdx,4)
movb $1, %al
je LBB0_6
## BB#2:
incq %rdx
cmpq %rcx, %rdx
jl LBB0_5
## BB#3:
xorl %eax, %eax
popq %rbp
retq
LBB0_1:
xorl %eax, %eax
popq %rbp
retq
LBB0_6:
popq %rbp
retq
.cfi_endproc
As well as failing to pull the "movb $1, %al" out of the loop (and avoid the
partial register move issue), it also fails to merge the 3 epilogs into
something like:
## BB#3:
LBB0_1:
xorl %eax, %eax
LBB0_6:
popq %rbp
retq</pre>
</div>
</p>
<hr>
<span>You are receiving this mail because:</span>
<ul>
<li>You are on the CC list for the bug.</li>
</ul>
</body>
</html>