2012-04-23T00:01:35 *** alehorst has quit IRC (Ping timeout: 246 seconds) 2012-04-23T00:17:30 *** dvladim has quit IRC (Ping timeout: 260 seconds) 2012-04-23T00:26:24 *** alehorst has joined #aichallenge 2012-04-23T00:31:20 *** amstan has quit IRC (Ping timeout: 246 seconds) 2012-04-23T00:50:59 *** pgpaskar1 has joined #aichallenge 2012-04-23T00:52:45 *** Alexer has joined #aichallenge 2012-04-23T00:58:07 *** welterde has quit IRC (*.net *.split) 2012-04-23T00:58:08 *** Alexer- has quit IRC (*.net *.split) 2012-04-23T00:58:08 *** pgpaskar_ has quit IRC (*.net *.split) 2012-04-23T00:58:08 *** ronchaine has quit IRC (*.net *.split) 2012-04-23T00:58:08 *** Masxmasx has quit IRC (*.net *.split) 2012-04-23T00:58:09 *** loglog has quit IRC (*.net *.split) 2012-04-23T00:58:26 *** replore_ has quit IRC (Remote host closed the connection) 2012-04-23T01:07:30 *** replore_ has joined #aichallenge 2012-04-23T01:10:53 *** Cyndre_ has quit IRC (Quit: Leaving) 2012-04-23T01:13:02 *** ronchamon has joined #aichallenge 2012-04-23T01:14:50 *** loglog has joined #aichallenge 2012-04-23T01:15:20 *** welterde has joined #aichallenge 2012-04-23T01:17:31 *** Masxmasx has joined #aichallenge 2012-04-23T02:40:12 *** replore_ has quit IRC (Remote host closed the connection) 2012-04-23T02:57:20 *** NoxiaZ^ has joined #aichallenge 2012-04-23T02:58:51 *** _dr <_dr!~johannes@faui32o.informatik.uni-erlangen.de> has joined #aichallenge 2012-04-23T03:29:01 *** Elverion has joined #aichallenge 2012-04-23T03:29:40 *** Elverion has left #aichallenge 2012-04-23T03:41:23 *** mviel_ has joined #aichallenge 2012-04-23T04:06:32 *** replore_ has joined #aichallenge 2012-04-23T04:27:00 *** pairofdice has joined #aichallenge 2012-04-23T05:02:31 *** moises_ has joined #aichallenge 2012-04-23T05:04:05 *** moises_ has quit IRC (Remote host closed the connection) 2012-04-23T05:24:21 antimatroid: You'll just have to invent your own langauge or look around the internet some more. I guess the list of languages that offer pattern matching and C-like performance is rather short at the moment. 2012-04-23T05:24:56 *** replore_ has quit IRC (Remote host closed the connection) 2012-04-23T05:25:37 Some functions could be written more cleanly structured with pattern matching, like working with angles and special casing 0°, 90° and so on 2012-04-23T05:56:41 Removing simple pattern matching languages and oldies, I see: Scala, Objective CAML, ML, Haskell 2012-04-23T05:58:52 and I'll just add Rust, since I heard of it and it has switch-cases that do _some_ pattern matching 2012-04-23T06:10:19 Now let's remove from these 5 languages the ones that don't have "anything as an operator": there goes Rust, Haskell, ML dialects 2012-04-23T06:12:32 Wikipedia also has a list of languages that allow new definable operators: ALGOL 68, Eiffel, Fortran, F#, Haskell, Io, Perl 6, R, Scala, Smalltalk. Maybe you want to learn ALGOL instead or Perl? :D 2012-04-23T06:12:39 *** mleise1 has quit IRC (Quit: Leaving.) 2012-04-23T06:22:09 *** mleise has joined #aichallenge 2012-04-23T06:22:26 *** Elverion has joined #aichallenge 2012-04-23T06:25:50 *** Elverion has quit IRC (Client Quit) 2012-04-23T06:48:27 Also F# offers pretty much the same overloading scheme as ALGOL. You get a set of symbolic characters that you can use to define new operators. ALGOL uses proper mathematical symbols, while F# can combine multiple ASCII characters. But neither allows you to just put down anything you like as an operator. Honestly I don't know if it wouldn't create more problems than it solves in a language implementation. 2012-04-23T06:48:56 *** delt0r_ has quit IRC (Ping timeout: 245 seconds) 2012-04-23T07:02:13 *** delt0r_ has joined #aichallenge 2012-04-23T07:08:42 mleise: i suspect my largest problem is insufficient knowledge of basic programming concepts 2012-04-23T07:08:48 it seems to trip me up a lot 2012-04-23T07:09:04 what do you mean by basic concepts? 2012-04-23T07:09:20 you know what inheritance is, what a recursion is etc. 2012-04-23T07:42:32 *** foRei has joined #aichallenge 2012-04-23T07:45:23 *** ecidforiap has joined #aichallenge 2012-04-23T07:46:39 *** pairofdice has quit IRC (Ping timeout: 265 seconds) 2012-04-23T07:46:39 *** ecidforiap is now known as pairofdice 2012-04-23T08:01:19 *** cyphase has quit IRC (Ping timeout: 276 seconds) 2012-04-23T08:08:54 *** alehorst has quit IRC (Remote host closed the connection) 2012-04-23T08:09:58 *** alehorst has joined #aichallenge 2012-04-23T08:23:40 *** cyphase has joined #aichallenge 2012-04-23T08:57:06 *** mceier has joined #aichallenge 2012-04-23T08:57:47 mleise: you gotta nic me or I forget :P 2012-04-23T08:58:01 i don't know, but haskell just confuses the hell out of me every time i play with it 2012-04-23T08:59:17 *** mcstar has joined #aichallenge 2012-04-23T09:01:32 ahoy 2012-04-23T09:03:44 antimatroid: haskell lets you add new infix operators, define a precedence for them, you can define associativity to the left or to the right, or none, you can use any 2 argument function as an infix operator, you can add new types to numerical typeclasses(or others), what else do you need? 2012-04-23T09:04:08 it sounds so awesome 2012-04-23T09:04:09 but i don't get it 2012-04-23T09:04:17 :( 2012-04-23T09:04:24 im failed as a teacher 2012-04-23T09:04:29 i get what you just said 2012-04-23T09:04:39 but i don't get programming in haskell 2012-04-23T09:04:44 and there's so much syntax to remember just to do stuff like input 2012-04-23T09:04:49 i like >> and << in c++ 2012-04-23T09:04:56 it's simple and composable 2012-04-23T09:05:19 >> has a similarity to haskell's >> 2012-04-23T09:05:27 (monadic composition) 2012-04-23T09:06:11 anyway, haskell is very much syntax less 2012-04-23T09:06:25 it is quite logical 2012-04-23T09:06:48 antimatroid: how much did you read from the gentle introduction? 2012-04-23T09:06:57 first few things 2012-04-23T09:07:00 then tried to play around :P 2012-04-23T09:07:03 chapters? 2012-04-23T09:07:05 i don't have time atm anyway 2012-04-23T09:07:08 k 2012-04-23T09:07:18 swamped atm :( 2012-04-23T09:07:54 im waiting for new firmware to arrive... 2012-04-23T09:08:28 antimatroid: machine learning starts soon, got an email from them 2012-04-23T09:08:36 yeah i know 2012-04-23T09:08:40 i don't have time though 2012-04-23T09:08:46 yeah... 2012-04-23T09:10:45 mleise: clean lets you define operators like haskell, the two language is very similar, and there is pure, which is based on term rewriting, it is very configurable 2012-04-23T09:10:55 languages are* 2012-04-23T09:14:18 mcstar: about that failing as a teacher... I don't know. You can chose: either its Haskell's fault, our exposition to C++ or yours. 2012-04-23T09:15:57 mleise: how do you solve en-route dihydration? 2012-04-23T09:16:36 I take a PET bottle of water and apple juice with me 2012-04-23T09:16:43 mleise: btw, you can judge the above a bit, since i tried to explain some things to you 2012-04-23T09:16:47 or I jog in the morning hours 2012-04-23T09:17:00 i sleep in the morning hours 2012-04-23T09:17:24 well, anyway, i woke at 7, but i dont think im gonna start running that early 2012-04-23T09:17:45 i just hate holding a bottle all the time 2012-04-23T09:18:18 i should strap it on my back, and have a straw to drink from it 2012-04-23T09:29:15 mcstar: ride a bike :P 2012-04-23T09:29:37 antimatroid: good idea, but it takes too long 2012-04-23T09:29:46 pedal harder then 2012-04-23T09:29:51 XD 2012-04-23T09:34:05 mleise: have you heard of something like compressing memory to increse throughput? 2012-04-23T09:35:07 mcstar: compressing data in RAM with a sophisticated algorithm (like zip) for local processing? no 2012-04-23T09:35:36 I've heard of compressing files on low storage or for internet transfers (obvious) 2012-04-23T09:36:06 i mean something like always writing compressed data to dram 2012-04-23T09:36:11 And it is a nice optimization to store a collection bit flags instead of larger data types 2012-04-23T09:36:15 and uncompress it into cache, and process it 2012-04-23T09:36:41 mcstar: no, do some memory controllers do that? 2012-04-23T09:36:42 im just thinking about the memory bandwidth/latency problem 2012-04-23T09:36:49 i dont know 2012-04-23T09:38:57 it's probably too complicated to implement on hardware (you need to be able to address the memory byte-by-byte) 2012-04-23T09:39:13 mleise: you know what us annoying? video cards have the same kind of memory as cpu's, and they can reach 100GB/s+ , why cant these ppl, integrate dram into the motherboard? 2012-04-23T09:39:16 unless you restrict yourself somehow 2012-04-23T09:39:19 right next to the cpu? or something 2012-04-23T09:39:42 us->is 2012-04-23T09:39:49 fast ram is expensive and ... yes you need it close to the cpu 2012-04-23T09:41:22 After all CPUs *have* L1 + L2 caches, what more do you want :) 2012-04-23T09:41:45 that is just patching the problem, it isnt a solution 2012-04-23T09:41:48 the latency is actually hidden *perfectly* when you access data in sequence 2012-04-23T09:42:05 what happens when you have 4 or more cores? 2012-04-23T09:42:16 the modern x86 cpus only 'stall' when you use a random pattern 2012-04-23T09:42:41 mleise: im reading an article, i will be much smarter in 30 minutes 2012-04-23T09:42:47 they can follow up to 16(?) data streams either forward or backwards 2012-04-23T09:43:02 but it definitely says, today's cpus starve more than ever 2012-04-23T09:43:28 They just say that to support their claims :-| 2012-04-23T09:43:49 mleise: what does follow data stream mean? 2012-04-23T09:44:17 they have 16 pointers to fetched memory regions? 2012-04-23T09:44:27 when the CPU detects a sequential access pattern it will preload the data into the cache 2012-04-23T09:44:39 yeah 2012-04-23T09:44:42 arr[0], arr[1], arr[2] and so on 2012-04-23T09:45:06 and yes, they can follow multiple such streams, but I don't know the exact number 2012-04-23T09:45:25 So I tried it out with a linked list 2012-04-23T09:45:44 Once with the linked items actually in the correct order in RAM and once with a random linking 2012-04-23T09:46:06 what was the test? copying? 2012-04-23T09:46:15 The random order was slow as molasses, whereas the ordered list was as fast as using L1 or L2 caches 2012-04-23T09:46:26 no I just accessed the pointers *g* 2012-04-23T09:46:40 but still, it means reading data at that memory location 2012-04-23T09:47:01 how did you make a random list? 2012-04-23T09:47:19 i guess memory allocation happens pretty sequantially 2012-04-23T09:47:20 using a uniform RNG 2012-04-23T09:47:38 i dont get it 2012-04-23T09:47:39 i just created some random indexes and then linked the items accordingly 2012-04-23T09:47:45 ah 2012-04-23T09:47:50 ok 2012-04-23T09:47:52 i had a preallocated array 2012-04-23T09:47:56 pfff 2012-04-23T09:47:57 stupid me 2012-04-23T09:48:09 i think for the very first time i understood it like that 2012-04-23T09:48:17 but then started thinking(which i shouldnt) 2012-04-23T09:48:39 hey, it took me a while to understand a simple function call in haskell 2012-04-23T09:49:20 the important thing is that you had fun 2012-04-23T09:49:24 You can actually write the test yourself. It's quickly done in any language 2012-04-23T09:49:45 i believe you 2012-04-23T09:50:01 no, seeing is believing :p 2012-04-23T09:50:19 how long was the list? 2012-04-23T09:51:06 long enough to exceed both cache levels 2012-04-23T09:51:25 was it larger than operational memory? 2012-04-23T09:51:28 XD 2012-04-23T09:51:31 a multiple of the L2 cache actually. i tried different lengths and compared them 2012-04-23T09:51:53 what is the operational memory? all RAM? ... well no. 2012-04-23T09:52:04 more like maybe 512 MB 2012-04-23T09:52:34 yeah, ram is called operational memory too 2012-04-23T09:53:01 in contrast with swapped memory i guess 2012-04-23T09:54:01 I once implemented this: http://en.wikipedia.org/wiki/Fibonacci_heap 2012-04-23T09:54:06 It was a total fail 2012-04-23T09:54:12 :) 2012-04-23T09:54:24 why? 2012-04-23T09:54:30 It has nice O(...) bounds, but it just isn't practical on the usual hardware 2012-04-23T09:54:38 too many cache misses, pointers etc... 2012-04-23T09:55:13 A normal array for storing the data (like Java uses) was 2 times faster for A* in Ants 2012-04-23T09:55:18 when i needed a priority queue, i tried like 4 heap implementations in CL 2012-04-23T09:55:49 i think our problem size was really small 2012-04-23T09:58:06 Fibonacci heaps are not the easiest to implement, and I had to debug mine quite a bit, but it showed me that you cannot trust big-O notations for everything. (I hoped that the A* application was a long enough list to hide the constant overhead of F.heaps) 2012-04-23T09:58:57 in fact you may often need n>=1000 to see any benefit over naive approaches 2012-04-23T10:06:09 *** NoxiaZ^ has quit IRC () 2012-04-23T10:07:24 mleise: why do you say you cannot trust big-oh? 2012-04-23T10:08:27 big-oh(well, theta), just says, there is an n, when this will be better than that 2012-04-23T10:09:00 you really cant blame a theoretically sound technique, for poor performance 2012-04-23T10:09:02 yes, that's what I liked to ignore. 2012-04-23T10:21:23 *** amstan has joined #aichallenge 2012-04-23T10:21:23 *** ChanServ sets mode: +o amstan 2012-04-23T10:32:51 i can clearly see my lcd flicker 2012-04-23T10:34:18 On flat bakcground colors? 2012-04-23T10:34:31 Mine flickers on some gray tones 2012-04-23T10:35:06 yeah, on gray stripes 2012-04-23T10:35:45 pairofdice: http://fc20630405.wordpress.com/ 2012-04-23T10:35:56 that background is very flickery 2012-04-23T10:36:09 some flicker 2012-04-23T10:36:51 Here it's more noticeable on flat 2012-04-23T10:37:36 i think nouveau is running my screen at 60hz 2012-04-23T10:37:56 yeah 2012-04-23T11:09:33 *** Cyndre has joined #aichallenge 2012-04-23T11:25:38 *** X-Scale has joined #aichallenge 2012-04-23T11:56:31 *** choas has joined #aichallenge 2012-04-23T12:05:00 *** Palmik has joined #aichallenge 2012-04-23T12:22:57 *** mleise has quit IRC (Ping timeout: 250 seconds) 2012-04-23T12:43:05 *** dvladim has joined #aichallenge 2012-04-23T13:03:11 *** kilae has joined #aichallenge 2012-04-23T13:24:21 *** thestinger has joined #aichallenge 2012-04-23T13:26:46 *** cyphase has quit IRC (Ping timeout: 252 seconds) 2012-04-23T13:29:59 *** cyphase_ has joined #aichallenge 2012-04-23T13:34:49 *** cyphase_ is now known as cyphase 2012-04-23T13:35:06 *** cyphase has joined #aichallenge 2012-04-23T13:40:14 *** mleise has joined #aichallenge 2012-04-23T14:07:22 *** dvladim has quit IRC (Ping timeout: 260 seconds) 2012-04-23T14:24:41 *** mcstar has quit IRC (Quit: mcstar) 2012-04-23T14:34:53 *** kilae_ has joined #aichallenge 2012-04-23T14:35:30 *** alehorst has quit IRC (Quit: Leaving.) 2012-04-23T14:37:07 *** kilae has quit IRC (Ping timeout: 246 seconds) 2012-04-23T14:43:01 *** bbruh has joined #aichallenge 2012-04-23T14:44:46 *** dici has joined #aichallenge 2012-04-23T14:47:09 *** bbruh has quit IRC (Ping timeout: 245 seconds) 2012-04-23T14:49:02 *** delt0r_ has quit IRC (Ping timeout: 246 seconds) 2012-04-23T14:56:57 *** kurnevsky has joined #aichallenge 2012-04-23T15:02:48 *** delt0r_ has joined #aichallenge 2012-04-23T15:06:04 *** Accoun has quit IRC () 2012-04-23T15:18:38 *** Accoun has joined #aichallenge 2012-04-23T15:32:48 *** dvladim has joined #aichallenge 2012-04-23T16:12:01 *** dvladim has quit IRC (Ping timeout: 248 seconds) 2012-04-23T16:37:00 *** kilae_ has quit IRC (Quit: ChatZilla 0.9.88.2 [Firefox 11.0/20120312181643]) 2012-04-23T16:51:49 *** pairofdice has quit IRC (Quit: It is time for you to leave) 2012-04-23T16:52:38 *** Palmik has quit IRC () 2012-04-23T16:57:22 *** kurnevsky has quit IRC (Read error: Connection reset by peer) 2012-04-23T17:03:44 *** dici has quit IRC (Read error: Connection reset by peer) 2012-04-23T17:40:05 *** choas has quit IRC (Ping timeout: 245 seconds) 2012-04-23T17:59:33 *** GeorgeH has quit IRC (Ping timeout: 260 seconds) 2012-04-23T18:00:09 *** alehorst has joined #aichallenge 2012-04-23T18:02:42 *** GeorgeH has joined #aichallenge 2012-04-23T18:17:15 *** loglog has quit IRC (Remote host closed the connection) 2012-04-23T18:17:42 *** loglog has joined #aichallenge 2012-04-23T18:34:26 *** Redgis_ is now known as Redgis 2012-04-23T18:43:54 *** alehorst has quit IRC (Read error: Connection reset by peer) 2012-04-23T18:51:21 *** mceier has quit IRC (Quit: leaving) 2012-04-23T19:02:02 *** michael_alex has joined #aichallenge 2012-04-23T19:23:30 *** robertorocky has joined #aichallenge 2012-04-23T19:23:49 *** robertorocky has quit IRC (Client Quit) 2012-04-23T19:29:24 hey kids 2012-04-23T19:30:28 You don't get any pudding if you don't come to hear me! 2012-04-23T19:31:12 I think that's actually a nice filter. Young people surely don't know Pink Floyd 2012-04-23T19:32:00 *** cyphase has quit IRC (Ping timeout: 246 seconds) 2012-04-23T19:33:40 *that's* what he says? 2012-04-23T19:34:56 Now where did that "that's what she said" phrase come from anyway? 2012-04-23T19:35:09 I had a hard time tracking down "all your base are belong to us" 2012-04-23T19:35:53 I'm just a german guy. and every day there is some new inside joke or phrase on the internet that i don't get 2012-04-23T19:37:36 I just came back from the pub where a guy tried to couple me with a friend of his. She was a business woman. Can you imagine that? She wore a needle-stripe suit. That's so not my cup of tea 2012-04-23T19:40:42 man, this photo i shot and use for my background looks so noisy. 2012-04-23T19:42:01 why is it that your perception is so different under alcohol? it's not all getting blurry. actually it makes you take any time you need to look at things the way they are ... sometimes 2012-04-23T19:43:25 damn those small digital cameras. they only make good photos when your look at them sober 2012-04-23T19:44:14 now all that's left of this marveleous sky is noise. random noise all over the screen 2012-04-23T19:47:36 *** cyphase has joined #aichallenge 2012-04-23T19:49:24 http://imageshack.us/f/542/img0075it.jpg/ 2012-04-23T19:51:49 and what's more... why am i the only one writing something? 2012-04-23T19:52:13 I don't think that's how a chat is supposed to work 2012-04-23T19:54:24 ok good night you hollow emptiness of a chat window 2012-04-23T19:58:22 good night 2012-04-23T20:17:48 *** replore has joined #aichallenge 2012-04-23T20:40:08 *** X-Scale has quit IRC (Ping timeout: 272 seconds) 2012-04-23T20:43:24 *** X-Scale has joined #aichallenge 2012-04-23T20:43:41 *** X-Scale is now known as Guest33334 2012-04-23T20:53:26 *** Guest33334 has quit IRC (Ping timeout: 272 seconds) 2012-04-23T20:57:22 *** Guest33334 has joined #aichallenge 2012-04-23T21:41:30 *** Guest33334 is now known as X-Scale 2012-04-23T21:52:58 *** X-Scale has quit IRC (Ping timeout: 272 seconds) 2012-04-23T22:02:49 janzert1: http://forums.aichallenge.org/viewtopic.php?f=3&t=2423 2012-04-23T22:03:18 janzert1: tell me if you agree or not, or if you even want to be part of the next iteration 2012-04-23T22:06:03 Zannick: you might be interested too ^ 2012-04-23T22:46:23 *** delt0r_ has quit IRC (Read error: Operation timed out) 2012-04-23T22:57:20 *** foRei has quit IRC (Quit: Bye) 2012-04-23T22:57:40 *** mleise has quit IRC (Ping timeout: 252 seconds) 2012-04-23T22:58:33 *** mleise has joined #aichallenge 2012-04-23T23:01:51 *** delt0r_ has joined #aichallenge 2012-04-23T23:15:11 @later tell amstan I'm in complete agreement with the first two paragraphs. For the complete rewrite I'm still rather dubious but if someone(s) can get it done then more power to 'em. 2 weeks from game implementation to full launch is also fairly worrisome to me that we will end up with a quality game. But for the next several months at least I don't have much time to put into the contest so those that do should do what they wa 2012-04-23T23:15:11 janzert1: I think that worked... 2012-04-23T23:25:29 *** dvladim has joined #aichallenge 2012-04-23T23:52:15 "should do what they wa" 2012-04-23T23:52:32 amstan: i might be game 2012-04-23T23:52:38 for the rewrite