<html><head><meta http-equiv="Content-Type" content="text/html charset=utf-8"></head><body style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space;" class=""><div></div><div><br class=""></div><div>Michael<br class=""><blockquote type="cite" class=""><div class=""><div style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space;" class=""><div class=""><div class=""><br class=""></div><div class=""><br class=""></div><div class=""><br class=""></div><div class="">Michael</div><div class=""><br class=""></div><blockquote type="cite" class=""><div class=""><div class="gmail_extra"><br class=""><div class="gmail_quote">On Thu, Jan 29, 2015 at 4:12 PM, Michael Zolotukhin <span dir="ltr" class=""><<a href="mailto:mzolotukhin@apple.com" target="_blank" class="">mzolotukhin@apple.com</a>></span> wrote:<br class=""><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">Hi,<br class="">
<br class="">
Currently InlineCost pass fails to use fast-math flags when analyzing profitability of inlining. Here is a small example to demonstrate that:<br class="">
<br class="">
float foo(float *a, float b) {<br class="">
float r;<br class="">
if (a[1] * b)<br class="">
r = a[1]+a[2]/a[3]-a[4]*a[5]+a[6]-a[7]+a[8]*a[9]+<br class="">
a[11]+a[12]/a[13]-a[14]*a[15]+a[16]-a[17]+a[18]*a[19]+<br class="">
a[21]+a[22]/a[23]-a[24]*a[25]+a[26]-a[27]+a[28]*a[29]+<br class="">
a[31]+a[32]/a[33]-a[34]*a[35]+a[36]-a[37]+a[38]*a[39]+<br class="">
a[41]+a[42]/a[43]-a[44]*a[45]+a[46]-a[47]+a[48]*a[49]+<br class="">
a[51]+a[52]/a[53]-a[54]*a[55]+a[56]-a[57]+a[58]*a[59]+<br class="">
a[61]+a[62]/a[63]-a[64]*a[65]+a[66]-a[67]+a[68]*a[69]+<br class="">
a[0]*a[0]+b;<br class="">
else<br class="">
r = 1;<br class="">
return r;<br class="">
}<br class="">
float boo(float *a) {<br class="">
return foo(a, 0.0);<br class="">
}<br class="">
<br class="">
Here inliner can’t figure out that a[1]*b can be folded to 0 if the call is inlined, and decides not to inline.<br class="">
<br class="">
That happens because inliner drops fast-math flags in SimplifyBinOp. The attached patch fixes it by looking into the instruction math flags. Does it look good?<br class="">
<br class="">
Without patch:<br class="">
> clang inline.c -O3 -S -Rpass=inline -ffast-math<br class="">
// Empty output<br class="">
<br class="">
With patch:<br class="">
> clang inline.c -O3 -S -Rpass=inline -ffast-math<br class="">
inline.c:17:10: remark: foo inlined into boo [-Rpass=inline]<br class="">
return foo(a, 0.0);<br class="">
^<br class="">
<br class="">
<br class=""><br class="">
<br class="">
Thanks,<br class="">
Michael<br class="">_______________________________________________<br class="">
llvm-commits mailing list<br class="">
<a href="mailto:llvm-commits@cs.uiuc.edu" class="">llvm-commits@cs.uiuc.edu</a><br class="">
<a href="http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits" target="_blank" class="">http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits</a><br class="">
<br class=""></blockquote></div><br class=""></div>
</div></blockquote></div><br class=""></div></div></blockquote></div><br class=""></body></html>