#pentagram@irc.freenode.net logs for 3 Nov 2003 (GMT)

Archive Today Yesterday Tomorrow
Pentagram homepage


[00:28:45] <servus> I made up a test of collinearity that has 3 divides and 3 subtracts per test
[00:30:15] <servus> Try this http://4.60.123.159/files/Collinearity.txt
[00:31:34] <Coren_> Requires ordered points, though.
[00:32:09] <servus> A macro will do that
[00:32:14] <servus> #define MIN(a,b,c) ...
[00:32:28] * Coren_ will try it.
[00:38:23] --> Kirben has joined #pentagram
[00:38:23] --- ChanServ gives channel operator status to Kirben
[00:38:29] <servus> It'd seem that it actually doesn't need ordered points.
[00:40:39] <servus> You've just got to watch out for a degenerate case on a1,b1,c1...
[00:47:15] <servus> Reload the .txt; If one of the equations tested has a degenerate case (a1, b1, or c1 == 0), then don't test that case. If you don't test any of the cases, then return a point test (p1 == p2 == p3)
[00:48:07] <Coren_> My points are already know not coincident. At any rate, move this to #low. :-)
[01:08:43] <-- Dark-Star has left IRC ()
[01:31:45] <-- Coren_ has left IRC (Read error: 54 (Connection reset by peer))
[01:36:46] <-- Fingolfin has left IRC ("42")
[03:29:50] --> sdogi has joined #pentagram
[03:29:55] <sdogi> hi
[03:30:32] <SB-X> hello
[03:32:44] <sdogi> what do u do when everything you do seems meaningless
[03:33:41] <SB-X> stop doing it
[03:34:09] <sdogi> i have slept 3 days now.. done nothing.. i have to shorten two plays and also i have to give out a newspaper
[03:34:15] <sdogi> though that
[03:34:29] <sdogi> i guess i'll drop the newspaper part to some other geeks
[03:34:36] <sdogi> i have had enough
[03:35:21] <sdogi> huh
[03:35:54] <sdogi> stressful days... not that i have done something, but i have had lot of time to think
[03:36:15] <sdogi> hehe
[03:37:15] <sdogi> this channel made me come back to programming.. i wonder how far i get... maby so far that i can have one of the thousands jobs that needs c/c++ knowledge
[05:19:49] <-- sdogi has left IRC ("Client exiting")
[05:36:38] <-- Kirben has left IRC (Read error: 104 (Connection reset by peer))
[05:47:01] --> Kirben has joined #pentagram
[05:47:01] --- ChanServ gives channel operator status to Kirben
[06:26:59] <-- watt has left IRC ("using sirc version 2.211+KSIRC/1.2.4")
[07:38:15] <-- Kirben has left IRC ("brb")
[07:40:04] --> Kirben has joined #pentagram
[07:40:04] --- ChanServ gives channel operator status to Kirben
[08:05:46] <-- Kirben has left IRC ("System Meltdown")
[08:21:36] --> Kirben has joined #pentagram
[08:21:36] --- ChanServ gives channel operator status to Kirben
[08:44:35] <-- DarkeZzz has left IRC (Read error: 110 (Connection timed out))
[08:44:45] --> DarkeZzz has joined #pentagram
[10:27:15] --> wjp has joined #pentagram
[10:27:15] --- ChanServ gives channel operator status to wjp
[10:27:23] <wjp> servus: collinearity can be done much easier than that
[10:27:41] <wjp> five subtractions, two multiplications, one test
[10:28:15] <wjp> um, although that's the 2d case; the 3d case would be a bit more complex, but not too much, probably
[10:29:06] <wjp> you calculate (twice) the area of the triangle created by three points (two on the line, one you want to test)
[10:31:06] <wjp> but I see Coren_ already mentioned that in #low :-)
[11:00:41] <servus> Ah, so the area of the triangle should be 0? OK, well my post was the first thing I thought of..:)
[11:14:45] * wjp nods
[11:15:12] <wjp> the "subtractions and cross-product" Coren_ mentioned is probably calculating that area
[11:31:07] <servus> Well, I've been writing vertex programs for the last few days, I wouldn't actually trust anything I say. ;-)
[11:31:12] <servus> (My brain has melted!)
[12:07:35] * wjp hands servus some icecubes to re-solidify his brain
[12:07:46] --> Colourless has joined #Pentagram
[12:07:46] --- ChanServ gives channel operator status to Colourless
[12:08:12] <Colourless> hi
[12:08:29] * DarkeZzz drops a frozen anvil on servus, just to see if that helps any.
[12:08:33] <DarkeZzz> Hi!
[12:08:58] <Colourless> damn it darke, that's how we lost the other one
[12:09:25] <DarkeZzz> Hrm... actually this might be the lost one. I just found it in the back of the freezer.
[12:10:03] <wjp> hi Darke, Colourless
[12:10:05] <Colourless> cool.... in more than one way
[12:10:57] * DarkeZzz digs around in the back of the freezer and pulls out a frozen neanderthal, ponders it, then pushes it back in. Hrm... freezer hasn't been cleaned in a while it seems.
[12:11:59] <Colourless> wjp: cross products are the correct and usual way to do things
[12:12:40] <Colourless> primitive non humans in your freezer? Sounds like you have the freezer of a dragon
[12:14:23] <DarkeZzz> Yeah. Got it at a garage sale. He was selling his hoard to invest it in property.
[12:14:48] <Colourless> bought it for ember i gather
[12:17:46] <DarkeZzz> More or less. She prefers vegetables to meat, though occasionally tosses live prey into it.
[12:18:08] <Colourless> i'm sure a lovely female dragon would much prefer jewels and gold, to some old dry frozen corpse
[12:19:38] <DarkeZzz> No doubt. You can have the frozen monkeyboy if you want, I don't think she even knows it's in there. Must have come with the freezer. Still wondering how that anvil got in there though. *eartwitch*
[12:20:25] <Colourless> i'll pass on the corpse
[12:21:15] <Colourless> frozen meat is not for me. there is such a wide selection out there, why bother with second best
[12:21:49] <DarkeZzz> Besides, it's long past it's best used before date.
[12:22:14] <Colourless> you should donate it to some museum
[12:23:26] <wjp> Colourless: yeah, I expected as much
[12:24:06] <Colourless> just make sure it is what you think it is. I wouldn't put it past a dragon to say the corpse was neanderthal, when it might have just been a most unattractive human
[12:24:18] <Colourless> you know, rare sells for more
[12:24:49] <Colourless> and there are many unattractive humans out there, but few frozen neaderthals
[12:26:22] <DarkeZzz> Probably just ugly monkeyboy. Far, far too many of *them* around.
[12:26:48] <DarkeZzz> (Present company politely excluded, of course.)
[12:29:28] <DarkeZzz> Anyway, must hop off to nap. Night!
[12:30:28] <Colourless> sleep bunny, just make sure you haven't left that corpse out.... your friend the dragon might have frozen the corpse for a reason... i wouldn't want you to wake up with a zombie wondering around
[12:31:22] <DarkeZzz> Provided it's not wandering around me, it's fine!
[12:31:37] * DarkeZzz hides the frozen critter in the back of Colourless' lair. There we go.
[12:31:50] <Colourless> ooh. zombie.. i like zombies
[12:32:09] <DarkeZzz> Braiiiiiiins!
[12:33:05] <DarkeZzz> Anyway, naptime! Zzzzz...
[12:33:05] <Colourless> time to lure some hairless monkeys for some fun
[12:38:18] <-- Kirben has left IRC ("System Meltdown")
[12:53:01] * servus resents the comparison to a frozen monkeyboy:)
[12:53:20] * servus resents Microsoft's implementation of OpenGL...
[12:54:22] <Colourless> what exactly do you resent about OpenGL? That it's a pain in the ass to set up?
[12:54:54] <servus> As I understand it, I don't have to manually set up pointers to extension functions and query their address in other environments.
[12:55:00] <Colourless> IMO, which means nothing since it will never happen anyway, MS should create a new WGL
[12:55:07] <servus> WGL! *shiver*
[12:55:35] <Colourless> actually, using function pointers is useful, and GLX even has an extension that lets you do it
[12:55:44] <servus> I wish I could just find a library that had symbols for extension functions! I'm using a GL 1.1 library!
[12:55:47] <Colourless> WLG just sucks
[12:56:01] <Colourless> being integrate with GDI just makes WGL difficult to use
[12:56:13] <Colourless> servus, you would be in luck then
[12:56:38] <servus> Well, I use SDL_GL_GetProcAddress, but that is beside the point!
[12:56:50] <servus> I just want to use the function that has already been prototyped in gl.h!
[12:56:52] <Colourless> SGI itself has header files that contain all the extensions
[12:57:03] <Colourless> ah, you want to do that
[12:57:10] <servus> Yes, I have the header file, though it is not anywhere on the the WWW, as far as I can tell. b-(
[12:57:19] <servus> It took countless hours of searching to find it...
[12:57:28] <Colourless> http://oss.sgi.com/projects/ogl-sample/registry/
[12:57:32] <Colourless> http://oss.sgi.com/projects/ogl-sample/ABI/glext.h
[12:57:35] <Colourless> http://oss.sgi.com/projects/ogl-sample/ABI/glxext.h
[12:57:38] <Colourless> http://oss.sgi.com/projects/ogl-sample/ABI/wglext.h
[12:57:42] <servus> Erm, no. That won't help.
[12:57:42] <Colourless> not on the web?
[12:57:43] <Colourless> :-)
[12:57:59] <servus> You apparently haven't tried this in my environment. :)
[12:58:00] <Colourless> so you want function pointers for all the normal gl functions?
[12:58:11] <servus> My problem is in linking, not in compiling.
[12:58:18] <Colourless> there is some reason you can statically link to the dll?
[12:58:50] <Colourless> opengl32.lib or whatever you need for your compiler?
[12:59:07] <servus> Interestingly enough, not even my opengl32.dll has these entry points! I'm not sure where they are...
[12:59:15] <Colourless> which ones?
[12:59:20] <servus> If that were the case, I'd convert my opengl32.dll to a lib and use that
[12:59:33] <servus> glActiveTextureARB, for instance... my DLL does not contain a single extended function.
[12:59:48] <Colourless> ah, you need to get that as an extension
[13:00:41] <servus> Of course, I've "solved" this problem ages ago, but by "solved" I of course meant "made a cruddy and annoying hack which requires me to make a pointer to each function explicitely"
[13:00:56] <Colourless> Multitexture is an extension
[13:01:28] <Colourless> only the standard OpenGL 1.1 functions can be statically linked in Windows
[13:02:17] <Colourless> any extensions must be dynamically linked using wglGetProcAddress in windows
[13:02:20] <servus> Which was my original qualm. :P
[13:02:23] <Colourless> or SDL_GL_GetProcAddress if you are using sdl
[13:03:15] <Colourless> what exactly is your problem just doing
[13:03:24] <servus> Erm, I have no problem, per se.
[13:03:26] <Colourless> PFNGLACTIVETEXTUREPROC glActiveTextureARB = 0;
[13:03:30] <Colourless> then
[13:03:38] <servus> I am just complaining about the cruddy implementation! </rant>
[13:03:44] <Colourless> glActiveTextureARB = (PFNGLACTIVETEXTUREPROC) SDL_GL_GetProcAddress("glActiveTextureARB");
[13:03:44] <Colourless> ?
[13:04:53] <servus> Well, since I'm using the Opengl1.4 header, I have to use a doctored name, like glActiveTextureARB_x, as to not mess with the already prototyped [and completely useless!] functions in gl.h
[13:05:29] <servus> I'll just doctor my gl.h and make things a nightmare for anyone else, Mwauahhaha.
[13:05:37] <Colourless> you should be able to include gl.h without defining all the extension
[13:05:55] <servus> My gl.h prototypes every single extension function up to 1.4
[13:06:06] <Colourless> where did you get it from?
[13:07:11] <servus> Coren. It works fine, except for the fact that Windows is fighting with me.:)
[13:10:45] <Colourless> should use the headers from say this: http://oss.sgi.com/projects/ogl-sample/GLsdk.zip
[13:11:40] <servus> Nah, too old.
[13:11:59] <servus> I already went through all of this. I appreciate it, but don't go through the trouble. :)
[13:12:03] <servus> I'm fine. :P
[13:13:26] <Colourless> you you 'want' to have trouble compiling? you absoultely must use GL 1.4 headers, despite the fact that including a gl1.1 or gl1.2 gl/gl.h + most recent glext.h will give you all the functions
[13:14:27] <servus> I have no trouble compiling. :P
[13:14:34] <servus> I'm back to using a gl1.1.h though
[13:37:59] <servus> Ah, I got it all to work without too much hackiness, by using an ATI-provided .c to declare and fill in all the function prototypes
[13:38:29] <servus> It seems to be complete to GL 1.4, with a suspicious absence of NV extensions, which I can live with. heh.
[13:52:26] <wjp> are there any (modern) ATI cards without fans?
[13:53:52] <Colourless> not sure.
[13:54:04] <wjp> it's surprisingly hard to find that info online :-)
[13:54:06] <Colourless> well, actually I know 'some' don't
[13:54:14] <Colourless> but they are sort of expensive
[13:54:30] <Colourless> they use a heatpipe and a rather large heatsink
[13:56:42] * wjp hmms
[13:56:45] <Colourless> as far as I know, by standard all the cards use a fan. the 9600 cards use a really small fan
[14:01:11] <servus> Yepyep, they even have GF FX5900 Ultras without a fan(!!)
[14:01:36] <servus> (They don't use water cooling etc, either. It's a passing cooling system)
[14:01:54] <Colourless> Sapphire Atlantis Radeon 9800 and 9600 Ultimate Editions use passive cooling
[14:02:18] <Colourless> but they are anything but standard
[14:04:57] <servus> Just put a lavalamp globe over your chip and it'll cool off via convection ;-)
[14:08:08] <Colourless> It would seem that Radeon 9200 cards don't use fans, but they aren't entirely what i would call modern. Technology behind them is about 2 years old
[14:29:35] <wjp> how new are the FX5200's that Coren mentioned?
[14:30:00] <Colourless> they are new, but i wouldn't touch one.
[14:31:21] <Colourless> they are very much low end cards. Their speed using shaders is quite bad
[14:32:23] <wjp> well, my requirements for a 3d card are a bit lower than yours, probably :-)
[14:32:34] <wjp> (given how I'm perfectly happy with a tnt2ultra currently :-) )
[14:32:45] <Colourless> well, for you it would be a step up
[14:32:59] <Colourless> and on older games their speed is reasonable.
[14:33:16] <wjp> any idea if NWN would run ok?
[14:33:30] <Colourless> nwn should be fine as far as I know
[14:33:51] <Colourless> probably run better than for me (NWN was/is heavily Nvidia optimized)
[14:34:41] <Colourless> of course NWN will run on your TNT2 :-)
[14:35:00] <wjp> a tnt2 is the minimum system req...
[14:35:12] <Colourless> yeah
[14:36:05] <Colourless> thing is, you are running linux, so getting a nvidia card would probably be a good idea.
[14:37:13] <Colourless> just got to be completely sure about the card you are getting. There are all sorts of configurations for the 5200's
[14:37:35] <Colourless> things like 64 bit busses should really be avoided
[14:38:16] <Colourless> they as you much imagine are ultra cheap, and are have really terrible speed
[14:40:00] <wjp> so I'd want a 128 bit bus?
[14:40:26] <Colourless> yep
[14:46:34] <wjp> any way to tell which one I want? :-)
[14:47:02] <Colourless> um. no :-)
[14:47:20] <Colourless> you see, even if it came in a box, it's usually impossible to tell.
[14:47:57] <Colourless> information is vague. companies websites aren't always a help either. oddly enough, many have 'incorrect' information on them
[14:48:17] <Colourless> but, if you know the name and brand of the cardd, usually it will be possible to find out
[14:50:04] <wjp> let's see
[14:50:12] * wjp checks local store's website
[14:50:36] <wjp> Asus V9520 Magic GF FX5200, 128 MB, DDR, 8 x AGP, TV-out, Retail
[14:50:38] <wjp> Asus V9520 TDGF, FX5200, 128 MB, DDR, 8 x AGP, TV-DVI, Retail
[14:50:41] <wjp> Asus V9520 VS , FX5200, 128 MB, 2 X DVI/VIVOl
[14:50:48] <wjp> Gainward PRO/660 FX5200 128 MB/TV+DVI
[14:50:53] <wjp> PINE XFX FX5200 256Mb 8x AGP TVI/DVI
[14:51:13] * wjp checks asus.com
[14:51:46] <wjp> heh, 9520 is apparently too old for their homepage :-)
[14:51:53] <Colourless> unless you have a flat panel display with dvi, then the Asus V9520 VS probably isn't going to be useful for you
[14:52:07] <wjp> asus.com only lists the 9560/9950 :-)
[14:52:59] <wjp> well, this page is pretty much useless
[14:53:16] <Colourless> http://www.asus.com/products/vga/vgaindex.htm
[14:53:26] <wjp> yeah, that's where I'm looking
[14:53:43] <wjp> oh, those headers next to the cards are links it seems :-)
[14:55:59] <wjp> is that "128-bit" in the "superior innovations" list the bus width?
[14:56:11] <Colourless> nope :-)
[14:57:28] <wjp> it's amazing how little those info pages tell me :-)
[14:57:29] <Colourless> but, looking at the card itself, it only seems to have 2 ram chips, which isn't enough got a 128 bit bit
[14:57:34] <Colourless> s/bit bit/bus/
[14:58:11] <wjp> the 'magic' one?
[14:58:20] <Colourless> yeah
[14:58:21] <wjp> the 'td' one looks like it might have 4
[14:58:33] <wjp> but it also looks like it has a weird fan on it :-)
[14:59:34] <wjp> some of the cards on that vgaindex page have really interesting-looking cooling :-)
[15:01:33] <Colourless> the PINE XFX FX5200 will be 128 bit bus
[15:01:52] <Colourless> can't put 256 mb of memory on a 64 bit bus on a gfx card :-)
[15:02:01] <wjp> I expected as much, yes :-)
[15:04:56] <Colourless> gainward looks like 64 bit
[15:10:59] <Colourless> the pine XFX http://www.xfxforce.co.uk/product_view.php?sku=PVT34KUA
[15:12:28] <Colourless> but it has really slow memory on it instead of a 64 bit bus :-)
[15:13:32] <wjp> heh :-)
[15:14:02] <wjp> "AGP 8X - Provides double the bandwidth of AGP 4X" <-- no, really? :-)
[15:14:16] <wjp> pointless marketing statement of the day :-)
[15:14:47] <wjp> it does have DDR, doesn't it?
[15:14:51] <Colourless> yeah
[15:15:14] <Colourless> 333 MHz effective means it's actually running at 166
[15:16:09] <Colourless> my card runs at 310 MHz actual and it's got a 256 bit bus :-)
[15:16:31] <wjp> ouch :-)
[15:16:54] <wjp> hehe, I clicked on one of the pictures on that asus site and nothing happened
[15:17:19] <wjp> now I just looked back at my other browser window (containing a math-related site), and one frame was replaced by a picture of that card :-)
[15:17:40] <Colourless> a problem you might find is you want a decent performing card with no fan, you'll have to go expensive
[15:18:10] <Colourless> or go old
[15:18:46] <wjp> I think there are a few fanless gf4's
[15:24:03] * wjp sighs
[15:24:17] <wjp> too much choice :-)
[15:25:05] <wjp> I should get back to doing some work now :-)
[15:25:17] <Colourless> :-)
[15:25:20] <wjp> thanks for the advice, btw
[15:38:32] --> Cless has joined #Pentagram
[15:38:32] <-- Colourless has left IRC (Read error: 104 (Connection reset by peer))
[15:38:51] --- Cless is now known as Colourless
[15:38:53] --- ChanServ gives channel operator status to Colourless
[15:38:58] <-- Colourless has left IRC (Nick collision from services.)
[15:39:05] --> Colourless has joined #Pentagram
[15:39:16] --- ChanServ gives channel operator status to Colourless
[16:52:11] <-- wjp has left IRC ("bbl")
[17:09:47] --> wjp has joined #pentagram
[17:09:47] --- ChanServ gives channel operator status to wjp
[17:45:13] --> Fingolfin has joined #pentagram
[17:45:13] --- ChanServ gives channel operator status to Fingolfin
[17:45:57] <Fingolfin> hi
[17:46:21] <Colourless> hi
[18:52:50] <-- Colourless has left IRC ("casts invisibility")
[23:02:35] --> Kirben has joined #pentagram
[23:02:35] --- ChanServ gives channel operator status to Kirben
[23:08:48] <-- wjp has left IRC ("Zzzz...")
[23:27:19] --> Coren_ has joined #pentagram
[23:43:20] <-- Fingolfin has left IRC ("C++ is *not* Java")