| 00:20.52 | *** join/#brlcad Twingy (n=justin@74.92.144.217) |
| 02:16.37 | *** join/#brlcad yukonbob (i=1000@s142-179-54-198.bc.hsia.telus.net) [NETSPLIT VICTIM] |
| 02:16.37 | *** join/#brlcad vedge (i=vedge@vedge.org) [NETSPLIT VICTIM] |
| 04:19.01 | *** join/#brlcad louipc (n=louipc@206-248-165-55.dsl.teksavvy.com) |
| 07:06.09 | *** join/#brlcad d_rossberg (n=rossberg@bz.bzflag.bz) |
| 07:10.30 | *** join/#brlcad clock_ (n=clock@zux221-122-143.adsl.green.ch) |
| 09:26.25 | *** join/#brlcad cad71 (n=53060ec3@bz.bzflag.bz) |
| 11:00.03 | *** join/#brlcad Axman6 (n=Axman6@210-9-143-173.netspeed.com.au) |
| 11:04.52 | *** join/#brlcad mafm (n=mafm@elnet-111.lip.pt) |
| 11:06.45 | mafm | hallo |
| 11:12.43 | *** join/#brlcad cad81 (n=29d14bd3@bz.bzflag.bz) |
| 11:34.18 | *** join/#brlcad thing0 (n=ric@123.208.172.250) |
| 11:34.29 | thing0 | hey |
| 11:52.20 | *** part/#brlcad thing0 (n=ric@123.208.172.250) |
| 11:59.39 | *** join/#brlcad CIA-20 (n=CIA@208.69.182.149) |
| 12:00.23 | Axman6 | brlcad: yt? |
| 12:01.09 | Axman6 | or any devs really... |
| 12:03.04 | pacman87 | ~ask |
| 12:03.04 | ibot | ask is probably Questions in the channel should be specific, informative, complete, concise, and on-topic. Don't ask if you can ask a question first. Don't ask if a person is there; just ask what you intended to ask them. Better questions more frequently yield better answers. We are all here voluntarily or against our will. |
| 12:03.58 | alex_joni | pacman87: I assume the "against our will" applies more frequently.. :) |
| 12:07.55 | Axman6 | well i was just looking for a discussion on some stuff. i was wondering if veclib (specifically the vForce stuff) was being used in brl-cad on OS X (or whether it would be at all useful) |
| 12:10.05 | Axman6 | has functions that do things like "Set y[i] to the exponential function of x[i], for i=0,..,n-1" that uses the vector engine on 32, 64 bit, PowerPC and Intel procs, and uses them fully. |
| 12:30.15 | *** join/#brlcad CIA-21 (n=CIA@208.69.182.149) |
| 12:33.17 | *** join/#brlcad thing0 (n=ric@58.171.117.63) |
| 12:33.35 | Axman6 | i think i killed CIA-20... |
| 12:40.45 | alex_joni | Axman6: nope, it's beeing restarted |
| 13:00.20 | brlcad | Axman6: nope, it's not being used for anything |
| 13:00.51 | Axman6 | would it be at all helpful? |
| 13:01.14 | brlcad | where would you hook it in? :) |
| 13:02.06 | Axman6 | i dunno, you're the dev :P |
| 13:02.12 | Axman6 | i'm just the ideas man! |
| 13:02.13 | brlcad | vector libs works best when they are fully integrated vertically |
| 13:02.40 | Axman6 | what do you mean? |
| 13:03.21 | brlcad | aside from being platform specific, which historically we've avoided for longevity/portability/maintenance as they tend to come and go into popularity over the years |
| 13:03.44 | Axman6 | ah, fair enough. thought you might say that |
| 13:03.49 | Axman6 | still... it looks cool :P |
| 13:03.54 | clock_ | brlcad: platforms are like fads ;-) |
| 13:03.55 | brlcad | what I mean is that of course it's been thought about ;) but it's not just a matter of "just using it" in the least |
| 13:04.07 | Axman6 | yeah |
| 13:04.55 | *** join/#brlcad docelic (n=docelic@78.134.192.78) |
| 13:05.03 | Axman6 | oh well, just a thought. seems like it's the sort of thing brl-cad would do a lot of, and the routines used here are designed to be _fast_. (not that brl-cad needs much help there) |
| 13:05.11 | brlcad | even if we hooked it into the few places in our math library as a backend, you wouldn't likely see any performance gains because it wouldn't be pervasive |
| 13:05.25 | brlcad | you'd still end up stalling the vector unit every few intructions |
| 13:05.36 | Axman6 | how come? |
| 13:05.52 | clock_ | brlcad: I still didn't do the last bit to finish the video into releasability :( |
| 13:05.59 | Axman6 | i'm not all that familiar (at all) with the workings of brl-cad |
| 13:06.55 | brlcad | Axman6: because of the way vector math works, they get their speed by keeping a vector pipeline fed and crossing the boundary from system mem to the vector units infrequently |
| 13:07.31 | brlcad | without vertical integration where you restructure all computations specifically around that, you don't get those speed benefits because the pipeline isn't fed |
| 13:07.37 | brlcad | it stalls |
| 13:07.46 | brlcad | you can actually end up running slower |
| 13:07.47 | Axman6 | i thought that the point of this library was to do that as fast as it could be done. i could be wrong though. |
| 13:08.01 | Axman6 | interesting |
| 13:08.10 | brlcad | you're just reading the marketing materials, there's a lot more to it |
| 13:08.49 | Axman6 | i'm sure this'll all make more sense once once i've finished my 5 years at uni |
| 13:09.07 | brlcad | they're intentionally written as OMG!! *FAST* ponies!?!! |
| 13:09.49 | Axman6 | well i'm reading the header file, which just tells you what it does |
| 13:10.08 | brlcad | there have been efforts to take advantage of vectorization, but that requires a fair bit of data restructuring |
| 13:10.53 | brlcad | if someone(tm) implemented all of the vshot() routines in librt, then a vector lib would be more beneficial |
| 13:10.56 | Axman6 | fair enough. i wasn't sure how data was represented in the source |
| 13:11.02 | brlcad | but just hooking it in to the math lib wouldn't help |
| 13:12.05 | Axman6 | what does vshot() do? |
| 13:15.55 | brlcad | shoots a bundle of rays at a primitive |
| 13:16.00 | brlcad | vectorized |
| 13:16.14 | CIA-21 | BRL-CAD: 03bob1961 * r30978 10/brlcad/trunk/misc/win32-msvc8/brlcad/brlcad.sln: Added libged and support for x64. |
| 13:16.30 | Axman6 | anyway, i had better get to sleep. i've got google tech talk to watch tomorrow ay ANU :) |
| 13:16.37 | Axman6 | at* |
| 13:17.01 | Axman6 | night all, and don't have too much fun or you'll hurt yourselves! |
| 13:18.00 | brlcad | cya |
| 13:18.11 | CIA-21 | BRL-CAD: 03bob1961 * r30979 10/brlcad/trunk/misc/win32-msvc8/ (143 files in 143 dirs): Added builds for x64. |
| 13:33.38 | *** join/#brlcad prasad_ (n=psilva@h-67-103-183-185.mclnva23.covad.net) |
| 13:44.20 | *** join/#brlcad pacman_87 (n=timothy@nat-204-179.arlut.utexas.edu) |
| 14:06.29 | *** join/#brlcad docelic_ (n=docelic@78.134.199.54) |
| 14:10.32 | *** join/#brlcad whymarkwh (i=dsfsdfsd@196.211.34.3) |
| 14:14.09 | *** part/#brlcad whymarkwh (i=dsfsdfsd@196.211.34.3) |
| 14:34.39 | CIA-21 | BRL-CAD: 03bob1961 * r30980 10/brlcad/trunk/src/libbu/magic.c: Include bu.h |
| 14:37.21 | CIA-21 | BRL-CAD: 03bob1961 * r30981 10/brlcad/trunk/src/libbn/globals.c: Include bn.h |
| 14:40.52 | *** join/#brlcad prasad_ (n=psilva@h-67-103-183-185.mclnva23.covad.net) |
| 14:49.18 | *** part/#brlcad thing0 (n=ric@58.171.117.63) |
| 15:03.27 | yukonbob | waves in, packs out -- later cadheads |
| 15:04.28 | brlcad | howdy cya yukonbob |
| 15:04.29 | CIA-21 | BRL-CAD: 03brlcad * r30982 10/brlcad/trunk/ (include/magic.h src/libbu/magic.c): sort entries so they are alphabetical by group and consistent with the header (though noteworthy that not all in the header are listed here) |
| 15:27.20 | CIA-21 | BRL-CAD: 03brlcad * r30983 10/brlcad/trunk/ (4 files in 2 dirs): rename BUHOOK to BU_HOOK for consistency |
| 15:42.18 | *** join/#brlcad andrecastelo (n=chatzill@189.71.26.193) |
| 15:43.22 | andrecastelo | good morning everyone :D |
| 15:43.48 | prasad_ | gmorning |
| 16:32.27 | *** join/#brlcad Axman6 (n=Axman6@pdpc/supporter/student/Axman6) [NETSPLIT VICTIM] |
| 16:32.27 | *** join/#brlcad vedge (i=vedge@vedge.org) [NETSPLIT VICTIM] |
| 16:32.30 | *** join/#brlcad mafm (n=mafm@elnet-111.lip.pt) [NETSPLIT VICTIM] |
| 17:00.18 | *** join/#brlcad quentusrex (n=quentusr@c-71-197-244-228.hsd1.or.comcast.net) |
| 18:01.25 | *** join/#brlcad Elperion (n=Bary@p54875C42.dip.t-dialin.net) |
| 18:07.29 | brlcad | hello quentusrex |
| 18:11.17 | *** join/#brlcad docelic (n=docelic@78.134.199.54) |
| 18:14.47 | mafm | :) |
| 18:14.57 | mafm | that looked like a tsunami |
| 18:15.07 | mafm | see you tomorrow! |
| 18:15.11 | brlcad | see ya! |
| 18:29.28 | CIA-21 | BRL-CAD: 03brlcad * r30984 10/brlcad/trunk/src/libbu/globals.c: missed commit for BUHOOK |
| 18:52.06 | CIA-21 | BRL-CAD: 03brlcad * r30985 10/brlcad/trunk/ (5 files in 2 dirs): consolidate all of the magic functions, defines, and other logic for magic number checking over into magic.[ch] |
| 19:02.54 | CIA-21 | BRL-CAD: 03brlcad * r30986 10/brlcad/trunk/src/mged/clone.c: clean up comments |
| 19:33.18 | CIA-21 | BRL-CAD: 03brlcad * r30987 10/brlcad/trunk/src/mged/attach.c: need ged.h for vo funcs |
| 19:39.54 | CIA-21 | BRL-CAD: 03brlcad * r30988 10/brlcad/trunk/src/mged/mged.h: ws |
| 19:53.26 | CIA-21 | BRL-CAD: 03brlcad * r30989 10/brlcad/trunk/src/mged/clone.c: fix some places in the v4 support where NAMESIZE needs to be used instead of CLONE_BUFSIZE, make tracker pair up bu_free with the bu_calloc calls |
| 20:14.55 | CIA-21 | BRL-CAD: 03bob1961 * r30990 10/brlcad/trunk/misc/win32-msvc8/ (6 files in 6 dirs): Mods to get things working with libged. |
| 20:17.28 | CIA-21 | BRL-CAD: 03bob1961 * r30991 10/brlcad/trunk/include/ (dg.h ged.h raytrace.h): Mods related to moving things from librt to libged. |
| 20:20.20 | CIA-21 | BRL-CAD: 03bob1961 * r30992 10/brlcad/trunk/ (4 files in 4 dirs): Update to 7.12.3 |
| 20:21.37 | CIA-21 | BRL-CAD: 03bob1961 * r30993 10/brlcad/trunk/src/librt/track.c: This moved to libged. |
| 20:22.24 | CIA-21 | BRL-CAD: 03bob1961 * r30994 10/brlcad/trunk/src/librt/ (importFg4Section.c wdb_comb_std.c): This moved to libged. |
| 20:26.08 | CIA-21 | BRL-CAD: 03bob1961 * r30995 10/brlcad/trunk/src/ (57 files in 5 dirs): Mods related to moving things from librt to libged. |
| 20:38.30 | CIA-21 | BRL-CAD: 03bob1961 * r30996 10/brlcad/trunk/include/raytrace.h: Add ifdef __RTGEOM_H__ around rt_bot_sort_faces and rt_bot_decimate declarations. |
| 22:08.37 | CIA-21 | BRL-CAD: 03brlcad * r30997 10/brlcad/trunk/src/libged/Makefile.am: need importFg4Section.c track.c wdb_comb_std.c |
| 22:27.15 | CIA-21 | BRL-CAD: 03brlcad * r30998 10/brlcad/trunk/src/other/tcl/generic/regex.h: |
| 22:27.15 | CIA-21 | BRL-CAD: revert 30995 as regfree does need to be defined to TclReFree else unresolved |
| 22:27.15 | CIA-21 | BRL-CAD: symbols on other platforms. the problem (presumably on windows) is probably the |
| 22:27.15 | CIA-21 | BRL-CAD: build finding the src/other/libregex/regex.h header before tcl's from |
| 22:27.16 | CIA-21 | BRL-CAD: src/other/tcl/generic/regex.h; the include paths should include tcl first in |
| 22:27.18 | CIA-21 | BRL-CAD: order for tcl to compile its regex interface. |
| 22:32.13 | *** part/#brlcad pacman_87 (n=timothy@nat-204-179.arlut.utexas.edu) |