#pentagram@irc.freenode.net logs for 29 Nov 2006 (GMT)

Archive Today Yesterday Tomorrow
Pentagram homepage


[02:30:05] --> SB-X has joined #pentagram
[02:38:23] <-- Lord_Nightmare has left IRC (Read error: 104 (Connection reset by peer))
[02:38:51] --> Lord_Nightmare has joined #pentagram
[05:33:21] --> watt_ has joined #pentagram
[05:43:47] <-- SB-X has left IRC ("*casts gate travel*")
[05:47:37] --> SB-X has joined #pentagram
[09:53:45] <-- SB-X has left IRC (Remote closed the connection)
[10:36:36] --> servus_ has joined #pentagram
[10:37:19] <-- servus has left IRC (Read error: 110 (Connection timed out))
[11:27:40] <-- Colourless has left IRC ("casts improved invisibility")
[13:25:13] <-- Kirben has left IRC (Read error: 110 (Connection timed out))
[13:52:29] <-- watt_ has left IRC ()
[14:16:17] --- Lord_Nightmare is now known as LordNAway
[14:26:21] --> watt has joined #pentagram
[14:26:21] --- ChanServ gives channel operator status to watt
[17:29:15] --> Darklock has joined #pentagram
[17:29:38] <Darklock> http://www.math.leidenuniv.nl/~wpalenst/bg2/paperdolling.png woo, look what wjp did!
[17:29:49] <Darklock> the characters show their equipment now
[17:29:50] <Darklock> :P
[17:29:52] <-- Darklock has left #pentagram ("Verlassend")
[19:31:00] --> Fingolfin has joined #pentagram
[19:31:19] --- ChanServ gives channel operator status to Fingolfin
[20:20:47] <-- Fingolfin has left IRC ()
[21:34:03] --> Kirben has joined #pentagram
[21:34:03] --- ChanServ gives channel operator status to Kirben
[21:36:57] --> SB-X has joined #pentagram
[21:37:18] <SB-X> wjp: what game are those paperdoll screenshots of?
[21:38:59] <wjp> baldur's gate 2
[21:39:45] <wjp> (running in gemrb)
[21:45:56] --> Fingolfin has joined #pentagram
[21:46:15] --- ChanServ gives channel operator status to Fingolfin
[22:33:18] <SB-X> I see. Good work, wjp. :)
[22:33:59] <SB-X> Someone on U6O uses one of those character portraits for his avatar and I was curious what it was from. :)
[22:35:28] --> Colourless has joined #Pentagram
[22:35:28] --- ChanServ gives channel operator status to Colourless
[22:35:43] <wjp> SB-X: yeah, they show up from time to time
[22:36:03] <wjp> hi Colourless
[22:39:14] <Colourless> hi
[22:40:00] * Colourless looks at bg2 screenshot. do infinity engine games really need to run at 100+ fps ? ;-)
[22:40:30] <wjp> there's a bit of a bug in the fps counter, actually :-)
[22:42:10] <wjp> oh, that reminds me, I need to see how many fps pentagram gets on my new PC :-)
[22:42:54] * Colourless remembers when pentagram div by zero'd on 'fast' computers when showing fps at 320x200
[22:42:58] <wjp> rendering time 9ms
[22:43:18] <wjp> (320x200 scaled with scale2x)
[22:43:51] <wjp> heh, yeah, I think Dominus ran into that at some point
[22:44:19] <Colourless> i think that once you hit sub 10 ms in pentagram, it's something else bottle necking your system, not the cpu
[22:44:31] <-- Fingolfin has left #pentagram ()
[22:44:53] <Colourless> a hardware buffer over PCI express should drop the ms right down
[22:44:58] <wjp> cpu usage is at 40% or so
[22:45:14] <Colourless> and shouldn't get a huge speed hit when reading back to do transparency
[22:45:45] * Colourless thinks for a moment
[22:46:09] <Colourless> doesn't pentagram play nice in its main loop and 'sleep' for a little bit?
[22:46:17] <wjp> it should, yes
[22:46:32] <Colourless> that will cut down the max fps
[22:50:08] <Colourless> of course back in the day when we were hitting 2-3ms rendering time, we weren't running any usecode or doing much of anything actually
[22:50:58] <wjp> ok, up to 300fps now
[22:51:06] <Colourless> heh
[22:52:58] <wjp> I guess this is one way of 'solving' the pathfinding issues :-)
[22:53:25] <wjp> Speaking of which, a Brian Tietz emailed me recently. He's working on improving pathfinding
[22:53:48] <wjp> He has some interesting ideas, and already implemented them as well
[22:54:16] <Colourless> cool
[22:55:10] <Colourless> i still wonder how the original managed to path find
[22:55:18] <Colourless> and quickly too
[22:55:23] <wjp> yes...
[22:55:57] <Colourless> while it may not have been perfect (there is no perfect solution to the problem), it was way more than adequate and ran significantly better than ours does
[22:56:33] <wjp> I'm beginning to wonder if running full animations in the pathfinder is the right thing to do
[22:57:24] <wjp> I tried approximating the results of animations by simple sweeps from start to end, but that didn't really seem to work out well
[22:57:41] <wjp> (but I can't say I tried very hard yet)
[22:58:11] <Colourless> allowing partial steps would probably help alot
[22:59:03] <wjp> that's already in there
[22:59:35] <Colourless> hm
[23:00:00] <wjp> at least, if I committed it
[23:00:57] <wjp> ah, I didn't
[23:01:06] <Colourless> :-)
[23:01:12] <wjp> I worked on this a couple of months ago
[23:01:22] <Colourless> well other than that.. i'm not entirely sure what to do
[23:01:36] <wjp> that did improve the paths, but didn't help the speed
[23:01:41] <Colourless> if someone else wants to play around with it... then yay for everyone
[23:01:55] <wjp> indeed :-)
[23:09:09] <wjp> I pointed Brian here by the way, so he may appear some time
[23:09:15] <wjp> I'll be away for two days myself
[23:09:44] <Colourless> i'm here most days
[23:09:59] <Colourless> though our time overlap isn't as big as it used to be :-)
[23:22:21] <-- watt has left IRC ()