[brlcad-commits] SF.net SVN: brlcad:[46360] brlcad/trunk/TODO

[<br...@us...>]

Revision: 46360
          http://brlcad.svn.sourceforge.net/brlcad/?rev=46360&view=rev
Author:   brlcad
Date:     2011-08-23 23:04:24 +0000 (Tue, 23 Aug 2011)
Log Message:
-----------
potential lead, -B works with parallel enabled so something fishy is going on.  possibly conversion from float to double ... but random numbers should not be involved!

Modified Paths:
--------------
    brlcad/trunk/TODO

Modified: brlcad/trunk/TODO
===================================================================
--- brlcad/trunk/TODO	2011-08-23 22:33:28 UTC (rev 46359)
+++ brlcad/trunk/TODO	2011-08-23 23:04:24 UTC (rev 46360)
@@ -58,9 +58,11 @@
 THESE ARE UNSCHEDULED BACKLOG TASKS
 -----------------------------------
 
-* investigate off-by-one errors from identical rt invocations.
-  running with -P1 results in no errors, parallel (on 64-bit Linux
-  RHEL) is giving many floating-point related errors.
+* investigate off-by-one AND off-by-many errors from identical rt
+  invocations.  running with -P1 results in no errors, parallel (on
+  64-bit Linux RHEL) is giving many floating-point related errors.
+  running with -B also results in no errors indicating maybe something
+  related to libbn random numbers.
 
 * investigate why epa primitive is so sensitive to coarse absolute
   value changes (run csgbrep, tol abs=1, facetize and get 321k poly)

This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.