<br><br><div class="gmail_quote">2009/6/22 Jason Garrett-Glaser <span dir="ltr"><<a href="mailto:darkshikari@gmail.com">darkshikari@gmail.com</a>></span><br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
<div><div></div><div class="h5">On Mon, Jun 22, 2009 at 12:42 PM, Igor Mozolevsky<<a href="mailto:igor@hybrid-lab.co.uk">igor@hybrid-lab.co.uk</a>> wrote:<br>
> 2009/6/22 Jason Garrett-Glaser <<a href="mailto:darkshikari@gmail.com">darkshikari@gmail.com</a>>:<br>
>> On Mon, Jun 22, 2009 at 12:33 PM, Igor Mozolevsky<<a href="mailto:igor@hybrid-lab.co.uk">igor@hybrid-lab.co.uk</a>> wrote:<br>
>>> If i_satd <= i_best<br>
>><br>
>> But i_satd can't be less than i_best, since i_best is initialized to<br>
>> COST_MAX64, which is a dozen orders of magnitude larger than the<br>
>> largest possible FIX8 i8x8 RD cost.<br>
><br>
> Can it be equal? If not, is there a need for `if( i_best > i_satd )' tests then?<br>
<br>
</div></div>No, it can't be equal either, but it isn't as if >= is faster than >.</blockquote><div><br></div><div>So are you saying that i_best > i_satd will always be true? Why have the `if' in the first place then? Also, i_nnz is initialised to zero for I_4x4 in the existing codebase.</div>
<div><br></div><div><br></div><div>Cheers,</div><div>--</div><div>Igor</div></div>