<div class="gmail_quote">2011/4/7 Siarhei Siamashka <span dir="ltr"><a href="mailto:siarhei.siamashka@gmail.com">siarhei.siamashka@gmail.com</a></span></div>
<div class="gmail_quote"><span dir="ltr"></span><br> </div>
<blockquote style="BORDER-LEFT: #ccc 1px solid; MARGIN: 0px 0px 0px 0.8ex; PADDING-LEFT: 1ex" class="gmail_quote">I could not come up with any better solution other than just using<br>different indentation for logically independent code blocks. Something<br>
like converting:<br></blockquote>
<div> </div>
<div>Yeah, I've seen that in your recent neon patches for bilinear.</div>
<div>I think it is a reasonable way to distinguish between logical code blocks.</div>
<div> </div>
<blockquote style="BORDER-LEFT: #ccc 1px solid; MARGIN: 0px 0px 0px 0.8ex; PADDING-LEFT: 1ex" class="gmail_quote">
<div class="im">I think it might be interesting for you. I also have the following<br>experimental branch:<br><a href="http://cgit.freedesktop.org/~siamashka/pixman/log/?h=playground/slow-path-reporter" target="_blank">http://cgit.freedesktop.org/~siamashka/pixman/log/?h=playground/slow-path-reporter</a><br>
<br>It collects statistics about what operations do not have optimized<br>fast paths, along with the number of uses of these operations, total<br>number of pixels processed, average number of pixels per operation and<br>average scanline length. The code is currently linux specific and<br>
writes results to syslog. These results can be converted into a more<br>human readable form by a script. I'm using it quite successfully and<br>it revealed some of the missing optimizations which would be hard to<br>identify in some other way.</div>
</blockquote>
<div> </div>
<div>This is actually a great help for me!!</div>
<div>Thanks :-)</div>
<div><br>-- <br>Best Regards, </div>
<div>Taekyun Kim</div>