| 00:32.30 | silvap | hrm |
| 00:34.01 | silvap | srcimg is still running |
| 00:38.21 | silvap | oh snap |
| 00:38.27 | silvap | it's writing out the img :o |
| 00:42.54 | *** part/#brlcad Semhirage (n=Semhirag@unaffiliated/semhirage) |
| 00:43.50 | *** join/#brlcad Semhirage (i=semhirag@unaffiliated/semhirage) |
| 00:44.02 | silvap | well this is slow |
| 00:44.07 | silvap | 200K/s |
| 00:44.14 | silvap | on SATA |
| 01:18.02 | silvap | iowait at 93%.. @_@ |
| 01:18.17 | brlcad | :) |
| 01:18.24 | brlcad | you've hosed the machine |
| 01:26.32 | Twingy | o.O |
| 01:29.59 | brlcad | rub it in |
| 01:49.03 | Twingy | because I like to snort |
| 01:49.08 | Twingy | ...on my pillow |
| 01:49.10 | Twingy | o.O |
| 01:50.30 | silvap | well its getting there |
| 01:51.53 | silvap | 644MB/2GB |
| 01:51.56 | silvap | slowly.. |
| 01:52.08 | Twingy | :) |
| 01:52.16 | Twingy | make your code faster |
| 01:52.25 | Twingy | and less memory intensive |
| 01:52.28 | silvap | hdd must be thrashing every mem access |
| 01:52.38 | Twingy | ick |
| 01:52.38 | silvap | 2GB cache used |
| 01:52.44 | silvap | 7GB mem full |
| 01:52.56 | Twingy | :P |
| 01:53.21 | silvap | only if the 36K x 22K image has 1 color |
| 01:56.45 | Twingy | hrm |
| 01:56.52 | Twingy | I've speant the last 2 weeks working on nurbs |
| 01:56.59 | Twingy | feels very nostalgic |
| 01:59.43 | Twingy | high performance nurbs tracing |
| 02:00.00 | silvap | why do u have a 15% cutoff instead of a 4tris cutoff? |
| 02:00.21 | Twingy | hrm? |
| 02:00.33 | silvap | 'threshold' |
| 02:00.42 | Twingy | that's for the span |
| 02:00.49 | Twingy | if it doesn't find a span |
| 02:00.55 | Twingy | then it reverts to weighted graph |
| 02:01.08 | Twingy | termination is still 4 tri cutoff |
| 02:01.13 | Twingy | and depth |
| 02:01.24 | Twingy | and I'm about to add the last amount of logic |
| 02:01.37 | Twingy | as soon as I finish helping emanuel with integrating my nurbs code into blender |
| 02:02.54 | silvap | why dont u add memory management at the same time |
| 02:03.04 | Twingy | cause I'm not thinking about low level stuff right now |
| 02:03.09 | Twingy | optimizations are pass #2 |
| 02:03.10 | silvap | heresy! |
| 02:03.37 | silvap | so m3 quazi-meeting on monday |
| 02:03.37 | silvap | re: performance |
| 02:03.42 | Twingy | :) |
| 02:03.45 | Twingy | tell them I'm over a million rays/sec now :) |
| 02:03.52 | Twingy | one a lease pc! |
| 02:03.54 | silvap | will do |
| 02:04.25 | silvap | 'justin is getting 10 million rays/sec on stryker' |
| 02:04.59 | silvap | man i cant wait for this to render |
| 02:05.10 | silvap | stupid pngwriter lib is ass slow |
| 02:05.55 | silvap | i think the encoding is chewing most of the time |
| 02:06.23 | learner | heh, when did you start it? :) |
| 02:06.35 | silvap | i thought u were asleep |
| 02:07.13 | silvap | it's writing out at 200K/s |
| 02:07.58 | silvap | it'd be great if my structs were the cause of this bottleneck |
| 02:08.03 | silvap | then id be able to fix it |
| 02:08.45 | silvap | 778MB |
| 02:08.47 | learner | dude, it is a big image, and it's being compressed massively |
| 02:09.03 | silvap | heh |
| 02:09.15 | silvap | whats the compression btw |
| 02:09.20 | silvap | 2:1 on avg? |
| 02:09.45 | learner | noo |
| 02:10.16 | silvap | pixel space is 2GB |
| 02:10.17 | Twingy | heh |
| 02:10.37 | learner | it's more like 10:1 |
| 02:10.47 | Twingy | me and my silly poster idea |
| 02:10.55 | silvap | scene dependant, yes? |
| 02:11.18 | silvap | well the png file is at 800MB now |
| 02:11.27 | learner | it's all the black background, longs streams of it |
| 02:11.43 | silvap | well not exactly |
| 02:11.56 | silvap | there is no spacing between my glyphs |
| 02:12.03 | silvap | as in no straight black lines |
| 02:12.11 | learner | well, yeah, that's true -- yours is more tight |
| 02:12.20 | learner | but there is still lots of black |
| 02:12.26 | Twingy | tighter than a ... |
| 02:12.26 | learner | maybe 4:1 or so |
| 02:12.31 | silvap | png uses an index? |
| 02:12.42 | Twingy | you should use jpeg at 1% quality |
| 02:12.44 | silvap | well if it quit now, it'd be 2:1 so... |
| 02:12.46 | learner | it can, but doesn't by default iirc |
| 02:13.16 | learner | lessless wavelet compression if I had to guess |
| 02:13.29 | silvap | top is reporting 186:22 time |
| 02:13.46 | silvap | im guessing cpu only |
| 02:16.23 | Twingy | hrm, 340 days till siggraph? |
| 02:16.56 | silvap | the siggraph grifter |
| 02:17.05 | silvap | siggrifter |
| 02:17.08 | Twingy | hehe |
| 02:19.42 | silvap | hmm |
| 02:19.53 | silvap | seems i didnt specify compression |
| 02:19.54 | silvap | oops |
| 02:19.59 | Twingy | haha |
| 02:20.11 | Twingy | small oops :) |
| 02:20.50 | silvap | ah well |
| 02:21.02 | silvap | at least it'll finish in an hr |
| 02:21.38 | silvap | yay! |
| 02:21.40 | silvap | no more thrashing |
| 02:22.23 | silvap | 1.1GB |
| 02:22.47 | silvap | holy crap it's done |
| 02:22.58 | silvap | final size 1.1GB |
| 02:23.19 | learner | heh, okay, that's horrible compression |
| 02:23.50 | silvap | im gonna make a smaller one |
| 02:24.02 | learner | stupid 16bit pngs -- i didn't see a simple way to make em 8bit by default in pngwriter either without hacking it up quite a bit -- lots of assumptions on the bitsize |
| 02:32.11 | Twingy | neat |
| 02:32.21 | silvap | that's only src/anim tho |
| 02:32.31 | silvap | the GB file should be the whole src/ |
| 02:34.17 | Twingy | afk'ish |
| 02:37.05 | silvap | off to dc |
| 02:53.02 | Twingy | back |
| 03:10.37 | PrezKennedy | partay at Twingy's!! |
| 03:10.42 | Twingy | no |
| 03:10.45 | Twingy | headache at twingy's |
| 03:12.01 | PrezKennedy | i had one last night |
| 03:13.10 | Twingy | I'm having one helping emanuel try to grok ton's old nurbs code in blender and convincing him to scrap the backwards compat cause of his shitty code and just use my nurbs library I wrote that is atleast consistent |
| 03:13.24 | Twingy | and now he's agrees |
| 05:29.48 | PrezKennedy | this just in: hollywood has finally realized that remakes of old movies and television series suck |
| 05:34.13 | PrezKennedy | ...and 10 minutes of commercials before the movie is stupid |
| 06:00.03 | *** join/#brlcad Maloeran (n=alexis@modemcable233.164-81-70.mc.videotron.ca) [NETSPLIT VICTIM] |
| 06:02.30 | *** join/#brlcad ibot (i=ibot@rikers.org) |
| 06:02.32 | *** join/#brlcad TheLastSpartan (i=guu@208.48.199.217) |
| 06:04.47 | *** join/#brlcad Semhirage (i=semhirag@unaffiliated/semhirage) |
| 09:37.40 | *** join/#brlcad docelic (n=docelic@ri02-006.dialin.iskon.hr) |
| 12:33.04 | *** join/#brlcad brlcad (n=brlcad@bz.bzflag.bz) [NETSPLIT VICTIM] |
| 12:33.05 | *** join/#brlcad archivist_ (n=archivis@host217-35-76-52.in-addr.btopenworld.com) [NETSPLIT VICTIM] |
| 12:34.07 | *** join/#brlcad DTRemenak (n=Daniel_R@DHCP-170-143.caltech.edu) |
| 12:34.10 | *** join/#brlcad docelic (n=docelic@ri01-042.dialin.iskon.hr) |
| 12:34.32 | *** join/#brlcad learner (n=brlcad@pdpc/supporter/silver/brlcad) [NETSPLIT VICTIM] |
| 12:34.32 | *** mode/#brlcad [+o learner] by ChanServ |
| 12:35.02 | docelic | 'day folks |
| 14:57.36 | *** join/#brlcad Twingy_ (n=justin@pcp0011647505pcs.aberdn01.md.comcast.net) |
| 14:58.18 | CIA-5 | BRL-CAD: 03twingy * 10brlcad/src/adrt/libtie/ (define.h tie.c): |
| 14:58.18 | CIA-5 | BRL-CAD: New kd-tree building algorithm (yet to be named) is now suitable for general purpose |
| 14:58.19 | CIA-5 | BRL-CAD: use as an improvement (avg 25% - 100% faster) over the old algorithm and uses about |
| 14:58.19 | CIA-5 | BRL-CAD: half the memory consumption as before. One more criteria is yet to be added, more |
| 14:58.19 | CIA-5 | BRL-CAD: improvements to come. Algorithm is 10x slower than mid-split so kd-tree caching |
| 14:58.19 | CIA-5 | BRL-CAD: is on the way. |
| 16:35.24 | *** join/#brlcad docelic (n=docelic@ri02-128.dialin.iskon.hr) |
| 17:56.40 | CIA-5 | BRL-CAD: 03twingy * 10brlcad/src/adrt/libtie/tie.c: kd-tree code clean-up. |
| 20:54.54 | *** join/#brlcad docelic (n=docelic@ri01-151.dialin.iskon.hr) |
| 21:03.15 | learner | docelic, g'day |
| 21:04.14 | docelic | gentlemen' |
| 21:19.32 | Twingy_ | o.O |
| 22:16.56 | CIA-5 | BRL-CAD: 03twingy * 10brlcad/src/adrt/libtie/ (define.h tie.c): working on final changes |
| 22:22.24 | CIA-5 | BRL-CAD: 03brlcad * 10brlcad/src/libfb/Makefile.am: damn two character omission that ends up disabling the X and ogl interfaces.. they're CFLAGS dammit |
| 22:35.49 | *** join/#brlcad ibot (i=ibot@pdpc/supporter/active/TimRiker/bot/apt) |