| 00:22.05 | *** join/#brlcad ibot (i=ibot@rikers.org) |
| 00:22.05 | *** topic/#brlcad is http://brlcad.org/ || BRL-CAD is now Free Software! || BRL-CAD Windows port has been merged to HEAD, not yet stable yet || December/January is a unified iteration to allow time for holidays || Archer is now on head, kick the tires, report any bugs |
| 00:28.34 | *** join/#brlcad ibot (i=ibot@pdpc/supporter/active/TimRiker/bot/apt) |
| 00:28.34 | *** topic/#brlcad is http://brlcad.org/ || BRL-CAD is now Free Software! || BRL-CAD Windows port has been merged to HEAD, not yet stable yet || December/January is a unified iteration to allow time for holidays || Archer is now on head, kick the tires, report any bugs |
| 01:21.07 | PrezKennedy | ill reconsider a gov't job when our fine senators and representatives and GWB have to sit through ethics training like the rest of us |
| 01:41.29 | pra5ad | hear hear |
| 02:04.22 | tegtmeye | yo |
| 02:26.58 | *** join/#brlcad AchiestDragon (n=dave@whipy.demon.co.uk) |
| 02:40.49 | pra5ad | sup mike |
| 02:48.57 | tegtmeye | hows the engine? |
| 03:14.42 | pra5ad | slow progress |
| 03:14.53 | pra5ad | most of the time spent on 'OO' design |
| 03:14.57 | pra5ad | =( |
| 03:28.01 | *** join/#brlcad mahesh (n=mahesh@12-217-228-235.client.mchsi.com) |
| 07:10.53 | *** join/#brlcad clock_ (n=clock@84-72-61-200.dclient.hispeed.ch) |
| 10:51.15 | *** join/#brlcad DTRemenak (n=DTRemena@DHCP-170-143.caltech.edu) |
| 11:30.14 | *** join/#brlcad phcoder (n=phcoder@pcp0011650294pcs.aberdn01.md.comcast.net) |
| 14:16.50 | *** join/#brlcad clock_ (n=clock@84-72-61-200.dclient.hispeed.ch) |
| 14:17.08 | *** join/#brlcad clock_ (n=clock@84-72-61-200.dclient.hispeed.ch) |
| 15:25.10 | *** join/#brlcad pier (n=pier@151.56.220.10) |
| 18:34.21 | *** join/#brlcad learner (n=brlcad@pdpc/supporter/silver/brlcad) |
| 18:34.21 | *** mode/#brlcad [+o learner] by ChanServ |
| 19:09.25 | *** join/#brlcad mahesh (n=mahesh@12-217-228-235.client.mchsi.com) |
| 19:09.44 | mahesh | hi Sean |
| 19:10.01 | brlcad | howdy mahesh |
| 19:10.03 | brlcad | been a while |
| 19:10.09 | mahesh | yeah |
| 19:10.14 | brlcad | at least a while since you've spoken :) |
| 19:10.20 | mahesh | true |
| 19:10.28 | mahesh | as usual, had few questions |
| 19:10.32 | brlcad | how's the vacation? |
| 19:10.44 | mahesh | its good....enjoying |
| 19:11.00 | brlcad | that's good |
| 19:11.25 | mahesh | howz yours? |
| 19:12.27 | clock_ | brlcad: hi |
| 19:12.50 | brlcad | no vacation here yet, but the holidays were nice |
| 19:12.54 | brlcad | hi clock_ |
| 19:14.40 | mahesh | nice |
| 19:16.14 | mahesh | there is rt_shootray defined in shoot.c and vshoot.c |
| 19:16.44 | brlcad | yes, no need to worry about the latter |
| 19:17.03 | brlcad | that's for vectorized computation |
| 19:17.24 | mahesh | oh ok. my code currently breaks in rt_shootray |
| 19:17.35 | mahesh | i am trying to figure out why exactly that is happening |
| 19:17.44 | brlcad | backtrace? |
| 19:18.01 | mahesh | as of now, all the processors gets a certain chunk of pixel |
| 19:18.16 | mahesh | then when everyone tries to call rt_shootray |
| 19:18.33 | brlcad | how do they all get the geometry? |
| 19:18.34 | mahesh | the initiator node works fine but others blow up |
| 19:18.40 | mahesh | how do i backtrace? |
| 19:18.51 | brlcad | ever used gdb? |
| 19:19.11 | mahesh | yeah |
| 19:19.47 | brlcad | can run it in gdb to see where it halts, not sure how that'll work on the remote nodes though |
| 19:20.09 | brlcad | if you can get the remote to generate a core file, you could get a backtrace with that |
| 19:20.38 | brlcad | how'd you determine it was in rt_shootray to begin with? |
| 19:21.40 | mahesh | pretty ugly way but it works. i just used a print statement just before the call to rt_shootray and a statement after the function call. |
| 19:22.09 | mahesh | so all the nodes print the first statement |
| 19:23.10 | brlcad | hehe |
| 19:23.21 | brlcad | so "somewhere" inside of rt_shootray() :) |
| 19:23.59 | mahesh | yeah |
| 19:24.43 | brlcad | how are you invoking rt? |
| 19:24.54 | brlcad | there are debug flags you can pass at runtime |
| 19:25.01 | brlcad | that will cause loads of output |
| 19:26.08 | mahesh | mpirun -np <no. of processors> ./rt model.g <objects> |
| 19:26.22 | brlcad | the -x parameter for starters, try -x 1 |
| 19:27.21 | mahesh | ok |
| 19:27.58 | brlcad | there are bit fields defined in include/raytrace.h for DEBUG_* -- pass that to -x to get various sorts of debug output |
| 19:29.24 | brlcad | 1 is to print debug information about calls to rt_shootray() ;) |
| 19:29.43 | brlcad | 15 should be "everything" related to shotlining (first four bits) |
| 19:29.44 | mahesh | oh great. i wanted exactly that |
| 19:30.03 | brlcad | #define DEBUG_ALLRAYS 0x00000001 /* 1 Print calls to rt_shootray() */ |
| 19:30.04 | brlcad | #define DEBUG_ALLHITS 0x00000002 /* 2 Print partitions passed to a_hit() */ |
| 19:30.06 | brlcad | #define DEBUG_SHOOT 0x00000004 /* 3 Info about rt_shootray() processing */ |
| 19:30.09 | brlcad | #define DEBUG_INSTANCE 0x00000008 /* 4 regionid instance revectoring */ |
| 19:30.48 | brlcad | add up the values you want and that should be what you can provide to -x |
| 19:30.54 | mahesh | got it |
| 19:33.53 | CIA-5 | BRL-CAD: 03brlcad * 10brlcad/src/gtools/g_transfer.c: add an optional geometry argument for specifying what geometry to send (default is all); add a help option and make short options case insensitive |
| 19:46.04 | learner | big legs aren't going to help me bench next month |
| 19:46.29 | learner | neither will writing into the wrong window ;) |
| 21:19.07 | *** part/#brlcad pier (n=pier@151.56.220.10) |
| 21:20.32 | *** join/#brlcad Maloeran (n=maloeran@glvortex.net) |
| 21:54.06 | *** join/#brlcad _AchiestDragon (n=dave@whipy.demon.co.uk) |
| 23:11.01 | *** join/#brlcad mahesh (n=mahesh@12-217-228-235.client.mchsi.com) |