<html>
<head>
<base href="http://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 --- - big vector causes code explosion (no looping with vector attributes?)"
href="http://llvm.org/bugs/show_bug.cgi?id=17417">17417</a>
</td>
</tr>
<tr>
<th>Summary</th>
<td>big vector causes code explosion (no looping with vector attributes?)
</td>
</tr>
<tr>
<th>Product</th>
<td>new-bugs
</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>new bugs
</td>
</tr>
<tr>
<th>Assignee</th>
<td>unassignedbugs@nondot.org
</td>
</tr>
<tr>
<th>Reporter</th>
<td>kkhoo@perfwizard.com
</td>
</tr>
<tr>
<th>CC</th>
<td>llvmbugs@cs.uiuc.edu
</td>
</tr>
<tr>
<th>Classification</th>
<td>Unclassified
</td>
</tr></table>
<p>
<div>
<pre>Apologies if I've misunderstood the intent of the 'ext_vector_type' attribute,
but I did not expect the following source code to generate over 4KB of machine
code:
$ cat vecadd.c
typedef float big_float_vector __attribute__((ext_vector_type(800)));
big_float_vector foo(big_float_vector x, big_float_vector y) {
return x + y;
}
$ ./clang -Os -S -fomit-frame-pointer vecadd.c -o - -emit-llvm
; ModuleID = 'vecadd.c'
target datalayout =
"e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64-S128"
target triple = "x86_64-apple-macosx10.7.0"
; Function Attrs: nounwind optsize readonly ssp uwtable
define <800 x float> @foo(<800 x float>* byval nocapture readonly align 32,
<800 x float>* byval nocapture readonly align 32) #0 {
entry:
%x = load <800 x float>* %0, align 32
%y = load <800 x float>* %1, align 32
%add = fadd <800 x float> %x, %y
ret <800 x float> %add
}
attributes #0 = { nounwind optsize readonly ssp uwtable
"less-precise-fpmad"="false" "no-frame-pointer-elim"="false"
"no-infs-fp-math"="false" "no-nans-fp-math"="false"
"stack-protector-buffer-size"="8" "unsafe-fp-math"="false"
"use-soft-float"="false" }
$ ./clang -Os -S -fomit-frame-pointer vecadd.c -o -
.section __TEXT,__text,regular,pure_instructions
.globl _foo
_foo: ## @foo
.cfi_startproc
## BB#0: ## %entry
pushq %rbp
Ltmp2:
.cfi_def_cfa_offset 16
Ltmp3:
.cfi_offset %rbp, -16
movq %rsp, %rbp
Ltmp4:
.cfi_def_cfa_register %rbp
andq $-32, %rsp
subq $32, %rsp
movaps 3200(%rbp), %xmm0
movaps 3184(%rbp), %xmm1
movaps 3152(%rbp), %xmm2
addps 7248(%rbp), %xmm2
addps 7280(%rbp), %xmm1
addps 7296(%rbp), %xmm0
movaps 16(%rbp), %xmm8
movaps 32(%rbp), %xmm9
movaps 48(%rbp), %xmm5
movaps 64(%rbp), %xmm6
movaps 3072(%rbp), %xmm7
movaps 3136(%rbp), %xmm3
addps 7232(%rbp), %xmm3
addps 7168(%rbp), %xmm7
addps 4160(%rbp), %xmm6
movaps 3088(%rbp), %xmm4
addps 7184(%rbp), %xmm4
movaps %xmm0, 3184(%rdi)
...
(and so on - no loop for an 800 element vector)
This is with:
$ ./clang -v
clang version 3.4 (trunk 191675) (llvm/trunk 191675:191681M)
Target: x86_64-apple-darwin11.4.2
Thread model: posix</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>