00:14.15 | trepan | might also be nice to allow both compressed binary and compressed text formats, as the compressed binary can still be half the size |
00:16.43 | trepan | if bzfs is using a URL'ed world, it would good for it to actually load from that file, and maybe even force clients to use the URL and not provide the built-in transfer mechanism (would make libcurl mandatory, or require an alternate http d/l routine) |
00:17.05 | freakazoid | downloading via http makes sense |
00:17.20 | trepan | that why it's already in place |
00:17.24 | trepan | ~tupone++ |
00:17.55 | freakazoid | Do you really think that the size of the world file is an issue though? |
00:18.14 | freakazoid | how big is the biggest world file you've seen? |
00:18.20 | trepan | freakazoid: i've got 23Mbyte text files |
00:18.39 | freakazoid | and a compressed binary is 1/2 the size of a compressed text file? |
00:18.44 | trepan | 2.5 Mbyte compressed text, 1.3Mbyte compressed binary |
00:18.54 | *** part/#bzflag amathis (~Aaron@amathis.linuxfordummies) |
00:18.56 | freakazoid | hardly worth the obfuscation |
00:19.04 | freakazoid | considering it's only loaded once on joining |
00:19.06 | trepan | what's your connection? |
00:19.14 | freakazoid | internet connection? |
00:19.26 | *** join/#bzflag Aribeth (~aribeth@dsl081-067-057.sfo1.dsl.speakeasy.net) |
00:20.19 | trepan | no, your genetic relationship to Elvis |
00:20.26 | lan56 | oh, well in that case :) |
00:21.21 | freakazoid | He and I are both from zeta2 reticuli IV |
00:21.29 | freakazoid | but I never met the guy |
00:27.32 | *** join/#bzflag GenRommel (~GenRommel@pcp09355007pcs.jersyc01.nj.comcast.net) |
00:28.17 | lan56 | did that feature freeze ever take place? I guess yes since there have been very few new features lately, but I can't be sure |
00:29.13 | trepan | it did not take place, and a few new features have been added |
00:29.35 | lan56 | ahh, good |
00:31.40 | trepan | althought, after the /date command was implemented, we were all convinced that that was the closing linch pin, and almost released that day =) |
00:31.43 | *** join/#bzflag Havoc_ (~Havoc@MTL-HSE-ppp188348.qc.sympatico.ca) |
00:32.23 | blast007_ | hello Havoc_ |
00:32.25 | lan56 | well, what can I say, my features are moving ;) |
00:32.29 | Havoc_ | hi blast |
00:42.49 | *** join/#bzflag orchid (orchid@orchid.user) |
00:50.16 | freakazoid | Patlabor221: The ogre samples from cvs don't seem to look right on my machine.. could that be due to radeon bugs? |
00:50.49 | freakazoid | for example, I get a bunch of dark triangles in the farther away LODs in the terrain sample |
00:50.59 | Patlabor221 | posibly |
00:51.13 | Patlabor221 | I haven't run the tuts in a while |
00:51.35 | freakazoid | hmmm, the skeletal animation sample doesn't show anything at all |
00:51.38 | freakazoid | and the water is mostly white |
00:51.59 | freakazoid | not a single one so far looks right |
00:52.08 | Patlabor221 | water generaly uses shaders |
00:52.21 | freakazoid | i need a new laptop |
00:53.38 | Patlabor221 | best of luck |
00:53.46 | freakazoid | :( |
00:53.53 | Patlabor221 | as far as I know for the 7000 series, it's DRI, or the ones you pay for |
00:54.02 | Patlabor221 | newer versions of DRI may help |
00:54.35 | Patlabor221 | as they are allways atempting to hack in more support for those cheap chipsets |
00:54.54 | freakazoid | Crap, *nothing* for Radeon 7500 |
00:54.56 | freakazoid | bastards |
00:55.02 | Patlabor221 | yeah 8000+ |
00:55.13 | Patlabor221 | they had to pick somewhere to start |
00:55.18 | Patlabor221 | and the 8000's have shaders |
00:55.21 | freakazoid | yeah, I suppose |
00:55.45 | Patlabor221 | nvidia supports every chipset down to the TNT |
00:57.00 | *** part/#bzflag freakazoid (~seanl@h-67-100-214-38.snvacaid.covad.net) |
00:57.36 | *** join/#bzflag freakazoid (~seanl@h-67-100-214-38.snvacaid.covad.net) |
00:57.54 | freakazoid | Patlabor221: I don't compile my kernel |
00:58.13 | freakazoid | Patlabor221: Unless you mean grabbing different X drivers |
00:58.34 | Patlabor221 | DRI should be a module |
00:58.40 | Patlabor221 | it may have updates |
00:58.48 | Patlabor221 | it's jsut an X driver |
00:59.28 | freakazoid | ok, so you don't mean the kernel part of DRI |
00:59.39 | Patlabor221 | no I mean the X driver part |
00:59.52 | Patlabor221 | tho dono if yours is built in, or a module |
00:59.59 | Patlabor221 | I think it can be done ether way |
01:00.05 | Patlabor221 | mine's allways been a module |
01:00.18 | Patlabor221 | http://dri.sourceforge.net/cgi-bin/moin.cgi/ |
01:00.39 | freakazoid | thanks! |
01:01.03 | freakazoid | btw, do ATI's linux drivers have any less functionality than the Windows drivers? |
01:01.13 | Patlabor221 | dono |
01:01.15 | Patlabor221 | never used them |
01:01.21 | Patlabor221 | I would asume not |
01:01.26 | Patlabor221 | tho dono how X handles dual head |
01:01.36 | freakazoid | I don't care about dual head since it's for a laptop |
01:01.43 | Patlabor221 | you may |
01:01.46 | Patlabor221 | you can dual head on lapys |
01:01.58 | Patlabor221 | external CRT and internal LCD |
01:02.04 | Patlabor221 | it's kinda nice |
01:02.15 | Patlabor221 | gives you more screen if your at a desk |
01:02.23 | freakazoid | aah |
01:02.26 | freakazoid | hmmm, good idea |
01:02.30 | Patlabor221 | lets see you have an rv200 right? |
01:02.31 | freakazoid | I have a monitor on my desk at work that does nothing |
01:02.33 | bryjen | X does divides a display into screens. :0.1 :0.2 |
01:02.40 | freakazoid | radeon 7500 mobility |
01:02.47 | Patlabor221 | ok yeah so a rv200 |
01:03.26 | freakazoid | what is rv200? |
01:03.28 | bryjen | laptop hardware is often oddball |
01:03.29 | freakazoid | is that the GPU or something? |
01:03.36 | freakazoid | is it different on the mobility than on the desktop version? |
01:03.46 | Patlabor221 | that is the chip name yes |
01:03.58 | Patlabor221 | it has the same basic specs |
01:03.59 | freakazoid | do I still use the radeon driver? |
01:04.04 | Patlabor221 | yeah |
01:04.07 | freakazoid | I see one called r200 also |
01:04.09 | Patlabor221 | it's slower bus speed |
01:04.10 | freakazoid | is that different? |
01:04.14 | Patlabor221 | no |
01:04.18 | freakazoid | yeah, I noticed it's a lot slower than my desktop |
01:04.18 | Patlabor221 | I don't think so |
01:04.28 | Patlabor221 | yeah it's so you get that kick ass batery life |
01:04.41 | Patlabor221 | basicly it's a lower power version |
01:04.52 | Patlabor221 | difrent die, slower bus to the vram, slower ram |
01:04.59 | bryjen | i've got a radeon 7500 agp card in my desktop. the r200 driver doesn't work with it |
01:05.02 | Patlabor221 | all designed to reduce power and heat |
01:05.07 | Patlabor221 | hmm them maybe it's not |
01:05.22 | bryjen | its the radeon driver |
01:05.52 | freakazoid | well, I'm using the radeon driver now and except for the bugginess it works |
01:05.55 | Patlabor221 | ahh the r200 is an 8500+ |
01:05.57 | *** join/#bzflag GenRommel (~GenRommel@pcp09355007pcs.jersyc01.nj.comcast.net) |
01:06.06 | freakazoid | bzflag works ok with the depth buffer turned off |
01:06.23 | Patlabor221 | hah, I would not call that "OK" |
01:06.40 | freakazoid | Patlabor221: is that a lot slower? |
01:06.56 | Patlabor221 | no, it looks like ass |
01:06.59 | freakazoid | it doesn't work at all with it on :) |
01:07.05 | freakazoid | or maybe it is on and I'm smoking crack |
01:07.10 | bryjen | perhaps a purple something |
01:07.14 | Patlabor221 | then your driver must suck bad |
01:07.37 | freakazoid | oh, never mind, it's on |
01:07.46 | trepan | BSP will be much slower in 1.11.x, no octree culling for regular nodes or shadoes |
01:08.09 | Patlabor221 | does it still jack some of the spliting planes? or did you fix that? |
01:08.23 | trepan | didn't use splitting for dynamic nodes |
01:08.25 | freakazoid | Hrrm |
01:08.41 | Patlabor221 | no the zbuffer splits, in the old code, it wouldn't allways work right |
01:08.44 | freakazoid | Patlabor221: Do you think I'll need to compile the kernel module to use the snapshot? |
01:08.56 | Patlabor221 | I dono, I use windows mostly :) |
01:08.59 | trepan | Patlabor221: i did fix the nodes that are split, but tanks are never split |
01:09.07 | Patlabor221 | ahh |
01:09.12 | trepan | so they look like junk with BSP, so i disabled animated treads there |
01:09.27 | trepan | as well, once mesh blasting is in place, BSP users will not get that benefit either |
01:09.28 | Patlabor221 | anything concave will look like ass |
01:09.44 | Patlabor221 | I say remove the option |
01:09.56 | Patlabor221 | software people would probably STILL be faster with a software z buffer |
01:10.30 | *** join/#bzflag lan56 (lan56@dialup-4.233.44.186.Dial1.LosAngeles1.Level3.net) |
01:10.37 | trepan | prolly, i don't really care how those 4 users find the game, but Tim does |
01:10.49 | trepan | it's a needless nuissance in my opinion |
01:11.12 | Patlabor221 | if it's shown to be faster then we tell tim that software people are still fine, and the option isn't needed anymore |
01:11.25 | trepan | we should track down those users, and have 'em test the relative speeds |
01:11.41 | Patlabor221 | that would be chestal |
01:11.54 | Patlabor221 | on his work system |
01:12.01 | trepan | how about we make a pool, raise $15, and buy him a new card |
01:12.12 | Patlabor221 | I offerd to send em cards |
01:12.14 | Patlabor221 | I have some here |
01:12.21 | Patlabor221 | but he can't change the card in his work system |
01:12.31 | Patlabor221 | I sent tim a card even |
01:12.32 | trepan | it would be especially good to do before the OpenGLGState rip |
01:12.40 | trepan | he shouldn't be playing games at work |
01:12.46 | trepan | ;) |
01:12.55 | Patlabor221 | IIRC thats where he hosts ducati |
01:13.04 | trepan | that's just bzfs |
01:13.17 | Patlabor221 | yeah but I think he plays from there every so often too |
01:13.27 | Patlabor221 | he works for a uni IIRC, so it's not real work :) |
01:13.57 | Patlabor221 | it's like a matrix g200 or something horible like that |
01:14.08 | trepan | as a side note, if anyone is thinking about better shadows, that'll probably require 8-bit stencil buffers + zbuffer, so it'll be yet another excluded item for BSP folks |
01:14.30 | Patlabor221 | yeah stencel shadows will kill em |
01:14.40 | Patlabor221 | the curent shadows are very much crap |
01:14.50 | trepan | if I see Chestal, i'll ask hiim to test it |
01:15.17 | trepan | (in an appropriate biased world ;) |
01:15.42 | freakazoid | sound used to require a completely separate processor to handle it. Now we use the CPU. Sprites used to be done in hardware. Now we use blitting. |
01:15.56 | freakazoid | date -R |
01:16.02 | Patlabor221 | in code |
01:16.12 | Patlabor221 | freakazoid, actualy sound is done on card hardware now |
01:16.12 | freakazoid | strftime('something') |
01:16.20 | Patlabor221 | and sprites are processed in the GPU |
01:16.23 | lan56 | asctime() |
01:16.29 | Patlabor221 | for cards that support them |
01:16.36 | freakazoid | hrrm |
01:16.47 | freakazoid | ok, I guess my theory about doing more on the CPU as the CPU gets faster is bunk then |
01:16.49 | Patlabor221 | hell OSX even uses openGL for all it's drawing |
01:16.56 | freakazoid | I was hoping eventually there wouldn't be separate 3d hardware |
01:16.59 | freakazoid | guess not |
01:17.12 | Patlabor221 | GPUs may outpace CPUs |
01:17.17 | lan56 | patlabor221: http://www.acm.uiuc.edu/webmonkeys/book/c_guide/2.15.html <- just as future reference about time/date |
01:17.26 | Patlabor221 | they have the advantage of being on the other side of the bus |
01:17.32 | freakazoid | I know John Carmack is a big opponent of putting general purpose functions on GPUs |
01:17.35 | Patlabor221 | since the bus is the lag |
01:17.48 | Patlabor221 | yeah cus then you can do more on that side of the bus |
01:17.49 | freakazoid | Patlabor221: I like the PS2 strategy of just making the bus really huge |
01:17.59 | freakazoid | no, OPPONENT |
01:18.01 | Patlabor221 | have you ever tried to program for that thing? |
01:18.11 | freakazoid | Patlabor221: no, but it can't be that hard |
01:18.14 | Patlabor221 | it is |
01:18.24 | Patlabor221 | in your code you have to know about hardware timing |
01:18.39 | Patlabor221 | the bus is fast because you have direct controll over it. |
01:18.42 | freakazoid | that's how it was on the commodore 64 too |
01:18.48 | Patlabor221 | so you do your own stuff like memory management |
01:19.00 | freakazoid | Patlabor221: that's what compilers and development kits are for :) |
01:19.22 | Patlabor221 | sure, it'll compile, but the basic lib dosn't do mem management |
01:19.27 | Patlabor221 | you jsut get a big ass address space |
01:19.31 | Patlabor221 | break it up how you want |
01:19.35 | Patlabor221 | there is no "malloc" |
01:19.55 | freakazoid | well, it's pretty much impossible to make a general purpose memory manager anyway |
01:19.59 | freakazoid | actually, it's been proven to be impossible |
01:20.08 | freakazoid | to letting the user do their own memory management makes sense |
01:20.22 | freakazoid | besides, that has nothing to do with having a really fat bus |
01:20.34 | Patlabor221 | now the xbox way is nice, where the cpu uses GPU ram as system ram |
01:20.41 | Patlabor221 | so there is no bus |
01:22.29 | freakazoid | That's old |
01:22.37 | freakazoid | UMA |
01:22.42 | freakazoid | SGI pioneered that |
01:22.48 | Patlabor221 | didn't say it was |
01:23.05 | Patlabor221 | it works well for dedicated graphics applications |
01:23.08 | freakazoid | the PS2 has no graphics RAM :) |
01:23.13 | freakazoid | just renders straight to the screen |
01:23.14 | Patlabor221 | nope |
01:23.20 | Patlabor221 | yeah |
01:23.29 | freakazoid | I find that cooler |
01:23.31 | Patlabor221 | so you have to make minicode for rendering everything |
01:23.39 | Patlabor221 | makes a lot of multipass stuff harder |
01:23.49 | Patlabor221 | shaders are a much better way to go |
01:24.16 | Patlabor221 | advantage of gp coding, and all the ram and buffers |
01:24.30 | Patlabor221 | and they run at the same time as CPU |
01:24.53 | Patlabor221 | on the current FX chips you can get 16 executions at the same time |
01:25.07 | Patlabor221 | no CPU can do that, if it could it would flood the bus |
01:25.25 | freakazoid | we just need to get rid of the bus |
01:25.49 | Patlabor221 | and put tons of execution units on the CPU? |
01:26.04 | freakazoid | sure |
01:26.13 | Patlabor221 | you'd end up just puting a GPU in the CPU |
01:26.22 | Patlabor221 | then you cant change it when something better comes out next month |
01:26.31 | freakazoid | a better CPU comes out every month too |
01:27.04 | Patlabor221 | yeah but who wants to replace both |
01:27.08 | freakazoid | while we're at it, let's put all the RAM on the die too |
01:27.15 | Patlabor221 | this way you get a bit more flex when you upgrade |
01:27.28 | Patlabor221 | not everyone needs the same amount of ram |
01:27.28 | trepan | freakazoid: you have no concept of die yields, do you? |
01:27.30 | freakazoid | Patlabor221: most mac users replace both at the same time |
01:27.40 | freakazoid | trepan: Dies get bigger and bigger all the time |
01:27.46 | Patlabor221 | freakazoid, because that's the only choice for disposable computers |
01:27.50 | Patlabor221 | same as laptops |
01:28.10 | freakazoid | I generally don't replace my graphics card on my desktop any faster than my CPU either |
01:28.23 | freakazoid | I usually replace everything at the same time, in fact |
01:28.56 | Patlabor221 | yeah but you are odd ;) |
01:29.03 | freakazoid | I don't deny that |
01:29.06 | Patlabor221 | a number of people piecemeal up |
01:29.28 | Patlabor221 | and I need a bit more ram then say somone's mom checking mail on AOL :) |
01:29.45 | Patlabor221 | when debuging this app takes 700-1000 megs before I even open a part |
01:30.31 | freakazoid | My desktop only has 512M I think |
01:30.47 | Patlabor221 | I consider that minimal, the starting point |
01:31.00 | Patlabor221 | I'd fill that by hiting F7 |
01:31.20 | Patlabor221 | parasolid is not a lib for the weak :) |
01:31.32 | trepan | ouch, what would F12 fill? ;) |
01:31.47 | Patlabor221 | none, I think that stops the project |
01:32.02 | trepan | ah, i was hoping for an exponential factor |
01:32.06 | Patlabor221 | are parts are often in the gig ranes |
01:32.23 | freakazoid | Patlabor221: What are you modelling? |
01:32.25 | Patlabor221 | so keeping them around, and an undo copy can take a bit :) |
01:32.35 | Patlabor221 | what ever our customers want to make |
01:32.48 | Patlabor221 | space shuttle parts |
01:32.51 | Patlabor221 | cat chow |
01:32.53 | freakazoid | Patlabor221: http://www.masten-space.com |
01:33.07 | freakazoid | Patlabor221: startup I do some work for |
01:33.11 | lan56 | mmm, cat chow |
01:33.19 | freakazoid | Patlabor221: They use pro-e and solidworks |
01:33.21 | Patlabor221 | do they use CNC machines? |
01:33.35 | Patlabor221 | yeah what do the do with the solidworks files to make them out of metal? |
01:33.47 | freakazoid | they have a CNC mill and one of their partners has some CNC EDM machines |
01:34.18 | Patlabor221 | there is a peice of software that comes inbetween the soldworks design phase and that CNC |
01:34.21 | Patlabor221 | that's what we do |
01:34.32 | Patlabor221 | take that model and compute all the toolpaths to cut the suker |
01:34.36 | freakazoid | Patlabor221: The SolidWorks stuff has only gone to vendors so far |
01:34.48 | freakazoid | Patlabor221: I think they've only programmed the CNC mill by hand |
01:34.50 | Patlabor221 | those vendors use a software like ours |
01:34.52 | Patlabor221 | hah |
01:34.59 | Patlabor221 | good luck making space parts then :) |
01:35.14 | freakazoid | Patlabor221: Well, they don't fabricate most of their own parts |
01:35.16 | Patlabor221 | really hard to program a trubine blade when it's got 7 billion moves :) |
01:35.18 | Patlabor221 | yeah |
01:35.27 | Patlabor221 | your vendors don't program stuff by hand |
01:35.32 | freakazoid | I'm sure they don't |
01:35.34 | Patlabor221 | they bring it into a CAM packate |
01:35.37 | Patlabor221 | that's us |
01:35.39 | freakazoid | cool |
01:36.33 | bryjen | "make the cpu really fast so it can do everything" is intel's mantra. of course that makes sense for them, as they want to sell lots of cpus |
01:36.39 | Patlabor221 | so we have to take that nice big solid you got and slice it up into little tiny chunks :) |
01:36.41 | freakazoid | Ok, I'm gonna reboot and see if X still wors |
01:36.47 | freakazoid | works, too |
01:36.52 | Patlabor221 | victory or death |
01:37.23 | Patlabor221 | I should take the 3ds to bz converter and make it into a plugin for our cam system :) |
01:37.33 | *** join/#bzflag amathis (~amathis@amathis.linuxfordummies) |
01:37.45 | lan56 | bryjen: sounds like a good seller to me, could be backed by some examples or evidence, but the idea is good business |
01:37.59 | bryjen | tanks fighting inside a turbine? |
01:38.05 | trepan | Patlabor221: you smoking crack over there? =) |
01:38.10 | Patlabor221 | no |
01:38.19 | Patlabor221 | it'd be easy to dump out the facets as whatever |
01:38.24 | Patlabor221 | I've made obj exporters |
01:38.30 | Patlabor221 | and an importer for marathon maps :) |
01:38.42 | Patlabor221 | no textures tho |
01:38.47 | lan56 | or bobs |
01:38.49 | lan56 | :) |
01:38.50 | Patlabor221 | you'd have full boolean support when modeling |
01:39.02 | trepan | as with blender (iirc) |
01:39.05 | Patlabor221 | yeah |
01:39.14 | Patlabor221 | tho they don't require manifolds |
01:39.26 | trepan | nope, but neither does bz at the moment |
01:39.27 | Patlabor221 | this would keep it all as surfaces untill export time |
01:39.36 | Patlabor221 | so you could pick the teselation |
01:39.53 | Patlabor221 | even more then surfaces b-rep solids |
01:40.08 | Patlabor221 | nice slider to pick your face count :) |
01:40.33 | trepan | i think blender has that sort of feature as well |
01:40.41 | Patlabor221 | depends on the input |
01:40.45 | Patlabor221 | they have surface support |
01:40.51 | Patlabor221 | so it can easaly teselate that |
01:40.57 | Patlabor221 | if it's a mesh it has to subdivide |
01:41.06 | Patlabor221 | and that sometimes gets "funky" |
01:41.34 | Patlabor221 | here every edge is defines by a geometric primitive |
01:41.42 | trepan | depending on how it handles atrributes, blender may make for a fine default world file format |
01:41.52 | Patlabor221 | probably |
01:41.55 | trepan | bases, teles, and links could simple be done with attribs |
01:42.01 | Patlabor221 | yeah |
01:42.14 | Patlabor221 | it would need some global atributes too |
01:42.35 | trepan | yup, worldSize, flagSize ... |
01:42.53 | trepan | but it would missed the "options" section, and probably a few other niceties |
01:42.58 | *** join/#bzflag PrezKennedy (~mkennedy@resnet-253-150.resnet.umbc.edu) |
01:46.06 | *** join/#bzflag GenRommel (~GenRommel@pcp09355007pcs.jersyc01.nj.comcast.net) |
01:46.29 | *** join/#bzflag freakazoi1 (~seanl@h-67-100-214-38.snvacaid.covad.net) |
01:46.36 | freakazoi1 | Hmm, glxgears locked up X |
01:48.24 | freakazoid | Terrain looks correct now! |
01:48.25 | freakazoid | Yay! |
01:49.33 | Patlabor221 | welcome to the 21st centry |
01:50.48 | lan56 | now, welcome to the 21st cent-u-ry :) |
01:51.19 | amathis | hrm.. |
01:51.31 | amathis | where are the flying cars? |
01:52.32 | trepan | with is the euthanasia for idiots that i was promised? |
01:53.00 | lan56 | and the suction tubes to get me miles away in less than a second? |
01:53.01 | trepan | s/with/where |
01:53.09 | Patlabor221 | I demand flying cars! |
01:53.23 | lan56 | I believe those are called airplanes :) |
01:55.07 | trepan | there's another 95.x years to go, patience... |
01:57.07 | lan56 | feature ideas these days for bzflag are scarce |
01:58.17 | trepan | why don't you speed up the radar, that should keep ya busy for a couple of hours |
01:58.37 | Patlabor221 | wow a slot A athalon processor |
01:58.41 | *** join/#bzflag freakazoid (~seanl@h-67-100-214-38.snvacaid.covad.net) |
01:58.41 | lan56 | isn't it already updating fast enough? |
01:58.42 | freakazoid | oh well |
01:58.50 | freakazoid | 20040924 is out of the question |
01:58.53 | freakazoid | need something more recent |
01:58.57 | trepan | lan56: no, it draws slowly |
01:59.12 | freakazoid | maybe I should be really brave and upgrade to xorg |
01:59.34 | lan56 | the only way I coudl think of to speed any rendering up is to eliminate calls to functions, other than that, I don't know of any way to make it faster |
01:59.46 | trepan | lan56: then you aren't the right person for the job |
01:59.56 | lan56 | I'd say so |
02:00.00 | Patlabor221 | I don't think he knows about textures :) |
02:00.12 | trepan | aren't those the pretty picture thingies? |
02:00.19 | Patlabor221 | totaly |
02:00.23 | lan56 | well, aren't they? :) |
02:01.20 | Patlabor221 | theya re |
02:01.23 | trepan | Patlabor221: any idea how many texture units the ATI 9x00 card and NV 5x00/6x00 cards have? |
02:01.39 | Patlabor221 | at least 4 |
02:01.59 | Patlabor221 | the full blown FX has 16 IIRC |
02:02.07 | Patlabor221 | 6800 may have 32 |
02:02.07 | trepan | my GF3 has four, i was just curious as to how many they're jamming in there these days |
02:02.16 | Patlabor221 | 16+ |
02:02.30 | Patlabor221 | tho a 5200 has like 4 or 8 I think |
02:02.37 | Patlabor221 | it's on the low end |
02:03.26 | Patlabor221 | 5700+ has 16 |
02:04.37 | CIA-2 | BZFlag: 03lan56 * 10bzflag/README.DEVC++: note about building with everything |
02:05.12 | Patlabor221 | I'm asuming a texture unit is what ATI calls it's " Up to sixteen extreme parallel pixel pipelines" |
02:05.18 | Patlabor221 | so 16 in it's best chip too |
02:05.28 | Patlabor221 | tho only 5 vertex shaders |
02:05.40 | Patlabor221 | sorry 6 |
02:05.45 | trepan | that could also be frag shaders |
02:06.10 | trepan | but who knows with marketoid talk |
02:06.20 | Patlabor221 | no it says vertex |
02:06.39 | trepan | na, the 16 parallel pipelines |
02:06.43 | Patlabor221 | ahh |
02:06.51 | Patlabor221 | next chip goes down to 8 of em |
02:08.52 | CIA-2 | BZFlag: 03lan56 * 10bzflag/Dev-C++/ (bzflag.dev bzfs.dev config.h): remove meshmaterial.cxx |
02:10.09 | Patlabor221 | hmm what to do with a k7 750 |
02:12.02 | lan56 | one way I could think of speeding up radar rendering would be to have a few vectors store all the info for each obstacle, therefore every time it is redrawn instead of having to recall every function to get info for each obstacle, it can just access a vector cell and boom it has it |
02:13.01 | lan56 | that might take up an enourmous amount of memory though |
02:13.38 | Patlabor221 | lan56 it's because it's drawing all the stuff, not looking it up |
02:13.53 | Patlabor221 | each object should have it's radar profile drawn to a texture |
02:14.00 | Patlabor221 | then just blast the texture on a quad |
02:14.07 | Patlabor221 | a lot less geo, hardware friendly |
02:14.09 | Patlabor221 | etc..... |
02:14.13 | Patlabor221 | the geo never changes |
02:14.16 | lan56 | ahh, I see. |
02:14.37 | Patlabor221 | you can even store the stuff that is at the same Z in one texture layer |
02:14.40 | Patlabor221 | and just rotate it |
02:14.46 | Patlabor221 | or at least put em in a list |
02:15.10 | lan56 | I see. |
02:15.12 | Patlabor221 | cards have fillrate up the wazzoo |
02:15.14 | Patlabor221 | so use that |
02:15.24 | Patlabor221 | geo transforms are the lillers |
02:15.35 | freakazoid | ok, brb |
02:16.32 | lan56 | let me guess the next line |
02:16.38 | lan56 | trepan is now known as trepan_away |
02:16.40 | lan56 | :) |
02:26.14 | *** part/#bzflag doughecka (~Doug@doughecka.user) |
02:54.16 | *** join/#bzflag Patlabor221 (~JeffM@Patlabor221.active.supporter.pdpc) |
02:54.16 | *** mode/#bzflag [+o Patlabor221] by ChanServ |
03:04.14 | *** join/#bzflag DTRemenak (~DTRemenak@131.215.170.165) |
03:05.53 | *** join/#bzflag sgk284 (~sgk284@AC8DF51A.ipt.aol.com) |
03:06.12 | sgk284 | nidhoggr: ping |
03:06.19 | Nidhoggr | ? |
03:06.29 | sgk284 | hey, you read the list recently? |
03:06.35 | Nidhoggr | which bits? |
03:06.46 | Nidhoggr | <--- has been on vacation. just got back yesterday. |
03:06.49 | sgk284 | about people randomly having secretplace.us in their host and ips as their names |
03:07.19 | Nidhoggr | you mean in /playerlist? |
03:07.24 | sgk284 | yea |
03:07.30 | sgk284 | sorry |
03:07.49 | amathis | wb Nidhoggr |
03:07.53 | amathis | Nidhoggr, how was trip? |
03:07.55 | Nidhoggr | because there is a bug if their hostname can't be looked up in dns, the slot isn't cleared and the last person's stuff shows up instead. |
03:08.06 | sgk284 | oh... okay |
03:08.08 | Nidhoggr | people in bzadmin were probably in there previously. |
03:08.16 | Nidhoggr | it was pretty good, amathis. |
03:08.22 | Nidhoggr | mexico can be very relaxing. |
03:08.30 | Nidhoggr | spent time on the beach and in fresh air. |
03:31.32 | *** join/#bzflag sgk284 (~sgk284@AC8DF51A.ipt.aol.com) |
03:32.12 | sgk284 | :/ |
03:32.19 | sgk284 | ww |