[brlcad-tracker] [ brlcad-Patches-2690863 ] improved performance for the extrude primitive
Open Source Solid Modeling CAD
Brought to you by:
brlcad
From: SourceForge.net <no...@so...> - 2009-03-19 06:15:51
|
Patches item #2690863, was opened at 2009-03-17 21:22 Message generated for change (Comment added) made by ddreeves70 You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=640804&aid=2690863&group_id=105292 Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: Raytrace Group: Incomplete Status: Pending Resolution: Rejected Priority: 7 Private: No Submitted By: David Reeves (ddreeves70) Assigned to: Sean Morrison (brlcad) Summary: improved performance for the extrude primitive Initial Comment: Mainly just focused on eliminating the allocs while raytracing. Moved most of the dynamic stuff to the prep function. ---------------------------------------------------------------------- Comment By: David Reeves (ddreeves70) Date: 2009-03-19 01:15 Message: Ok I have attached a new patch with the issues outlined below addressed. We discussed the MAXDEPTH today but for documentation purposes that settings is controlling how many times to subdivide the bezier in order to get down to a single segment (this would support up to 2^64 control points). Note though this was a const that was there before in the bezier evaluator. I only modified 3 files nurbs.h, bezier_2d_isect.c and extude.c. I have also modified my editor settings to be compliant with the HACKING standards. Please let me know if there is anything else so I can correct in my settings. ---------------------------------------------------------------------- Comment By: Sean Morrison (brlcad) Date: 2009-03-18 11:34 Message: Did a quick review and there are a few problems with the patch, some that need fixing. Looks like the patch has some debugging statements left in it that shouldn't be applied. (brep.cpp, ebm.cpp, maybe more) There is mixed style in the edits. Per our dev guide (HACKING file), the code should be using k&r/knf/stroustrup style instead of bsd/allman/java style for consistency sake. What's the implication of the MAXDEPTH 30 constant? Didn't get through all of the logic to infer what that impacts -- does that mean only 30 segments on a given shotline through an extrusion? There are a few stray edits that shouldn't be in there -- whitespace changes to a few random files. (src/rt/main.c, src/librt/shoot.c) Those are pretty innocuous and would normally be ignored, but given there are other issues, an update could leave them out. Often easiest to just run svn diff on the specific files you want included. ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=640804&aid=2690863&group_id=105292 |