<div dir="ltr"><div>Hello,<br><br></div><div>->Based on Olly's extremely good advice, I made changes to Inl2 and Ifb2 by calculating some complex sub expressions in Init() whhich depend on constant values. They are a part of the pull requests.<br>
<br></div><div>-> I also optimized IneB2 weighting scheme by :-<br></div><div><br></div><div>a.) Simplifying expressions and cancelling out common factors.<br></div><div>b.) Caclulating common sub-expressions in Init()<br>
c.) Tightening the upper bound.<br><br></div><div>I have opened a pull request for IneB2 too.<br></div><div><br></div><div>-> I am now working on DPH. I think I have some new ideas after discussing the upper bound optimizations with Olly today. <br>
<br></div><div>-> I will also work on timing the automated tests for the old and new code for each scheme and also profile the new scheme with kcachegrind and document the output.<br></div><div><br></div><div>Regards<br>
</div></div>