2011-12-04T00:11:35 *** delt0r_ has joined #aichallenge 2011-12-04T00:14:24 *** amstan_ has quit IRC (Read error: No route to host) 2011-12-04T00:14:25 *** amstan__ has joined #aichallenge 2011-12-04T00:14:25 *** ChanServ sets mode: +o amstan__ 2011-12-04T00:15:21 *** keith_ has joined #aichallenge 2011-12-04T00:28:41 *** rwest__ has joined #aichallenge 2011-12-04T00:28:41 *** rwest_ has quit IRC (Read error: Connection reset by peer) 2011-12-04T00:31:13 *** Surya has quit IRC (Quit: Page closed) 2011-12-04T00:35:08 *** rwest_ has joined #aichallenge 2011-12-04T00:38:20 *** rwest__ has quit IRC (Ping timeout: 244 seconds) 2011-12-04T00:53:47 *** antimatroid has quit IRC (Quit: Leaving.) 2011-12-04T00:53:55 *** antimatroid has joined #aichallenge 2011-12-04T00:54:23 antimatroid: how's your bot coming along? 2011-12-04T00:54:42 i mostly finished all my info stuff but now i have some weird bug 2011-12-04T00:54:48 it was taking up to like 1second to output "go" wtf? 2011-12-04T00:54:58 but i'm busy doing other stuff i've been neglecting today 2011-12-04T00:55:00 what about you? 2011-12-04T00:55:46 i settled on just recalculating hill information every time i see a previously unseen square or a hill has been discovered/razed 2011-12-04T00:56:03 I had to fix up my battle partition stuff too 2011-12-04T00:56:16 just found a bug, + sign should have been -, it was making my ants go away from enemy hills in some cases. so now my newest bot is starting to beat my current bot consistently, good sign 2011-12-04T00:56:30 that's good 2011-12-04T00:57:00 i have no idea what's causing this, i flush everytime i output something and yet the timer is at like 25ms before the end of the turn and then 1300ms after i output go 2011-12-04T00:57:17 and it only happens like randomly once every 100 or so turns 2011-12-04T00:57:26 could it be an OS scheduling issue? 2011-12-04T00:57:34 maybe 2011-12-04T00:57:44 i don't know anything about that kind of stuff :P 2011-12-04T00:58:02 I've seen that in my testing, especially if the disk decides to do something in the middle of a game. 2011-12-04T00:58:20 how do you test? 2011-12-04T01:00:35 usually with very specific examples i've set up and outputting info to a file 2011-12-04T01:00:45 for this it's popping up when I run a game 2011-12-04T01:00:49 once every hundred turns seems pretty frequent for a 1+ second pause caused by the OS, it seems like you'd notice something like that during general use 2011-12-04T01:00:59 i just feed my bot input data from a file and output stuff to a file 2011-12-04T01:01:09 do you use the --serial option when running games? 2011-12-04T01:01:23 janzert: yeah i don't think it's the os's fault 2011-12-04T01:01:41 i've never had that problem previously as far as i know, and it's happening at a specific point of the code every time 2011-12-04T01:01:45 there's no way it's not me somewhere 2011-12-04T01:01:49 I recommend it, because otherwise all the bots in the game compete for CPU at the same time and I've seen very inconsistent timings in that case 2011-12-04T01:01:58 i haven't used that 2011-12-04T01:02:04 although it's playing against lefty and hunter 2011-12-04T01:02:20 yeah, not likely the problem then 2011-12-04T01:02:48 I still recommend using --serial though, in provides much more consistent time behavior when you're running the game and all bots locally 2011-12-04T01:02:55 yep 2011-12-04T01:02:58 i will add that 2011-12-04T01:03:28 *** jstrong has quit IRC (Quit: Leaving.) 2011-12-04T01:04:28 antimatroid: whatever the problem is, it sounds very frustrating, but if it's always a specific spot in the code that helps narrow it down. 2011-12-04T01:04:51 not when the piece of code between them is "cout << "go" << endl;" :P 2011-12-04T01:05:20 that's as helpful as not knowing where the problem is at all 2011-12-04T01:05:41 it seems to me like i'm outputting a heap of stuff and not flushing it or something 2011-12-04T01:05:50 but i am adament that i'm not 2011-12-04T01:06:13 is there an easy way to get my output for a game? 2011-12-04T01:06:14 that could be, but I'm pretty sure < yeah it would then 2011-12-04T01:06:21 use the -O option 2011-12-04T01:06:38 it will write the bot output to the game_logs directory 2011-12-04T01:06:42 but if it outputted heaps before and didn't endl it it could potentially take ages to flush 2011-12-04T01:06:50 that's the only reasonable explanation i can think of 2011-12-04T01:07:06 antimatroid: maybe just try a cout << endl; before the first timing call and see if the problem disappears? 2011-12-04T01:07:19 janzert: that's a good idea 2011-12-04T01:07:26 it'll at least tell me it's happening cause of that 2011-12-04T01:07:36 * janzert nods 2011-12-04T01:08:39 janzert: what's the syntax for doing those * emotes? 2011-12-04T01:08:50 start with /me 2011-12-04T01:08:57 thx 2011-12-04T01:09:30 *** u_ has quit IRC (Ping timeout: 244 seconds) 2011-12-04T01:09:31 it's definitely the cout << endl; 2011-12-04T01:11:01 it's only losing like 10ms now 2011-12-04T01:11:08 but that's still not right 2011-12-04T01:11:58 got frustrated with my ants moving to the target in queue instead of encircling it when there is a friendly ant in front. 2011-12-04T01:12:22 antimatroid: can you try C style printf and see what would happen? 2011-12-04T01:14:19 remind me again? printf("go \n"); doesn't seem to flush? 2011-12-04T01:15:26 call fflush() after 2011-12-04T01:16:16 claims too few parameters? 2011-12-04T01:16:50 fflush(stdout); 2011-12-04T01:17:22 it's been a while since I did C++, but I seem to remember that C++ streams and C io calls don't necessarily play nice together when it comes to buffering and flushing, etc. 2011-12-04T01:17:44 i thought they had separate buffers 2011-12-04T01:17:58 yeah they interact strangely, you might not want to intermix them 2011-12-04T01:18:30 problem's still there 2011-12-04T01:18:59 fflush is just a hack for the printf() i've suggested to try if the cout << endl is really the problem 2011-12-04T01:19:36 i'm going to check the output 2011-12-04T01:19:42 ChrisH_: they sync by default (makes it really slow though) 2011-12-04T01:19:47 you can turn off sync though 2011-12-04T01:20:06 thestinger: ahhh, thanks for the correction 2011-12-04T01:20:08 std::ios::sync_with_stdio(false); 2011-12-04T01:20:17 yeah, that rings a bell now 2011-12-04T01:20:19 but then you can't mix them 2011-12-04T01:20:26 antimatroid: that means cout < yes, but the problem is occuring when i cout endl 2011-12-04T01:20:53 which would suggest there's something not getting flushed no? 2011-12-04T01:21:07 wow, it's been over six years since I did C++ regularly, hard to believe, that was my bread and butter for a long time 2011-12-04T01:21:08 i literally check the time before and after doing this 2011-12-04T01:21:12 and it's using the time there 2011-12-04T01:21:12 antimatroid: do you have sync between stdio and iostream off for stdout? 2011-12-04T01:21:27 probably 2011-12-04T01:21:37 i have "std::cout.sync_with_stdio(0);" 2011-12-04T01:21:55 i'd always found that made it faster 2011-12-04T01:21:59 yeah 2011-12-04T01:22:01 it should 2011-12-04T01:22:10 *** WW_ has joined #aichallenge 2011-12-04T01:22:11 std::ios::sync_with_stdio(false); 2011-12-04T01:22:17 try that, so it's off for stderr/stdin too 2011-12-04T01:22:36 i always found 0/1 faster than false/true 2011-12-04T01:22:41 although that's probably compiler specific 2011-12-04T01:23:31 they're the same 2011-12-04T01:23:41 take the md5sum of your binary and recompile with false instead of 0 2011-12-04T01:24:44 when i was on windows i tested them and there was a difference 2011-12-04T01:25:11 with debugging symbols, etc. off? 2011-12-04T01:25:19 no idea :P 2011-12-04T01:25:59 antimatroid: it wasn't changing the 0 to false that he was suggesting above though 2011-12-04T01:26:03 in case you missed that 2011-12-04T01:26:08 no i got that :) 2011-12-04T01:26:10 i was just noting 2011-12-04T01:26:23 didn't fix the problem either :( 2011-12-04T01:26:37 * antimatroid starts combing his source code 2011-12-04T01:26:50 i've searched every file for cout's 2011-12-04T01:30:35 grep "cout\|stdout" * :/ 2011-12-04T01:31:14 antimatroid: regarding your thought that a big output without a flush could be the problem: most implementations of std::cout that I've seen will only buffer a few K-bytes before doing an automatic flush anyway 2011-12-04T01:31:36 i'm looking for anything wrong at this point 2011-12-04T01:32:51 antimatroid: since it only happens every ~100 turns, is there anything in your code that is only performed that often? 2011-12-04T01:32:52 you mentioned testing with direct file input/output, are you seeing it when actually running under the engine too? 2011-12-04T01:33:05 i'm not sure it's that infrequent 2011-12-04T01:33:12 try cstdio :P 2011-12-04T01:33:19 i'm only seeing it with the engine atm 2011-12-04T01:34:04 are you writing a log of logging to stderr? 2011-12-04T01:34:14 (a lot of logging) 2011-12-04T01:35:02 all my logging is straight to an output file 2011-12-04T01:35:04 i can see that 2011-12-04T01:38:23 does it happen during the whole game or only in the later stages? 2011-12-04T01:39:56 and do you see python using a lot of CPU during the game if you run top or your OS equivalent? 2011-12-04T01:43:19 *** WW_ has quit IRC (Quit: Page closed) 2011-12-04T01:44:07 i just look near the start of the game 2011-12-04T01:44:09 like first 40 turns 2011-12-04T01:44:19 but it isn't consistent on the turn which is also annoying 2011-12-04T01:44:40 *** sigh has joined #aichallenge 2011-12-04T01:46:12 if you set the turntime for the game below the typical delay, does the engine cause your bot to timeout? 2011-12-04T01:46:49 *** praveen has joined #aichallenge 2011-12-04T01:47:55 it sounds to me like the game engine is not ready to read your moves, and when the flush happens your code is blocking waiting for the write to complete, but the game engine isn't reading from the pipe for some reason, and when it finally does then your code starts running again and measures the delay 2011-12-04T01:50:12 if this theory is correct then the problem would be some sort of systematic but irregular delay in the python runtime, not sure what that could be though 2011-12-04T01:51:16 i've seen an output in the commandline that says: INFO: invalid o x y D # duplicate order, where x and y seems to be row and col and D is direction. what does that mean? that i sent that specific line twice? 2011-12-04T01:51:28 another test would be to run the game with all bots being your code (remove the sample bots) and see if the delay happens to all instances of your bot 2011-12-04T01:51:30 AntDroid: yes 2011-12-04T01:51:49 AntDroid: you sent to orders for the same ant 2011-12-04T01:52:01 AntDroid: GreedyBot does that a lot 2011-12-04T01:53:10 antimatroid: also a game between three of your bot could ferret out if the delay happens on the same turn for all instances of your bot or on different turns in the same game 2011-12-04T01:53:56 it might be a bug then. i have code to take care of that but i may have missed something. 2011-12-04T01:56:01 *** janzert has quit IRC (Read error: Connection reset by peer) 2011-12-04T01:56:50 okay that's just weird 2011-12-04T01:57:02 not happening on the same turn, and one of them just did it after i just output endl 2011-12-04T01:58:12 AntDroid, just be sure that it's not one of the sample bots first 2011-12-04T01:59:10 *** janzert has joined #aichallenge 2011-12-04T02:00:09 *** yoden has quit IRC (Quit: Leaving.) 2011-12-04T02:00:21 * antimatroid cries 2011-12-04T02:01:40 antimatroid: what about the timeout test, is the delay always around 1sec? if so, set the turn time to 500ms and see if your bots timeout because of the delay or not 2011-12-04T02:02:06 ChrisH_: the delay is currently down at around 10ms 2011-12-04T02:02:23 oh 2011-12-04T02:02:27 antimatroid: the engine does that iirc 2011-12-04T02:02:35 are you sure whatever you're using for timing has a resolution higher than that? 2011-12-04T02:02:37 it has 10ms pauses between every 50 lines or something like that 2011-12-04T02:03:15 http://pastebin.com/AD2ZhaiQ 2011-12-04T02:03:17 i remember someone saying the engine checks for timeout every 10ms, not sure if that is related 2011-12-04T02:03:17 that's my timer 2011-12-04T02:03:22 but it had larger time things before 2011-12-04T02:03:36 thestinger: this is the time it is taking him to output the "go" line 2011-12-04T02:04:03 it took the 10 seconds after janzerts extra cout << endl line just before 2011-12-04T02:04:05 10ms 2011-12-04T02:05:47 well i'm still curious if the engine is receiving the "go\n" before or after the delay, which is why I suggest the timeout test, so maybe removing the extra cout< antimatroid: check clock_getres() 2011-12-04T02:06:02 that'll give you the resolution of gettimeofday 2011-12-04T02:06:33 doesn't that get you the res of clock_gettime()? 2011-12-04T02:06:40 I guess they're probably the same 2011-12-04T02:07:12 yeah, I suppose they may not be 2011-12-04T02:07:30 gettimeofday() is actually deprecated in POSIX:2008 :P 2011-12-04T02:07:41 but I think you need -lrt to use clock_gettime() 2011-12-04T02:08:01 the engine does a 10ms pause before reading anything 2011-12-04T02:08:03 and then reads 100 lines 2011-12-04T02:08:04 also, if the other 50 lines of output show smaller time steps than 10ms, then that would indicate a higher resolution. 2011-12-04T02:08:08 then another 10ms pause 2011-12-04T02:08:10 etc. 2011-12-04T02:08:16 ChrisH_: most of the time now the delay is happening with a "cout << endl;" line 2011-12-04T02:08:18 so the minimum time to output your orders is 10ms 2011-12-04T02:08:44 *** treeform has joined #aichallenge 2011-12-04T02:09:02 thestinger: yes, but unless he is filling the stdout buffer before the engine can read it, that shouldn't be an effect here 2011-12-04T02:09:23 what am I calling? clock_getres(gettimeofday, ?); 2011-12-04T02:09:26 no 2011-12-04T02:09:41 thestinger: want to pastebin your timer for me? :P 2011-12-04T02:09:43 and in the first 40 turns I doubt he is outputing enough to overflow the buffer 2011-12-04T02:09:48 that way i can see if it's that 2011-12-04T02:09:50 antimatroid: http://pubs.opengroup.org/onlinepubs/9699919799/ 2011-12-04T02:10:03 if thestinger's explaination of the engine's logic is correct, then I think he has a good point 2011-12-04T02:10:03 I just use gettimeofday() because the new replacement stuff needs -lrt 2011-12-04T02:10:15 http://pastie.org/2963242 2011-12-04T02:10:17 look 2011-12-04T02:10:22 it pauses for 10ms before it starts reading 2011-12-04T02:10:26 so you fill the buffer, and output go 2011-12-04T02:10:35 it will take 10ms at minimum to read it all 2011-12-04T02:11:12 it probably takes under 10ms for you to do the output 2011-12-04T02:11:19 it also means that if you have 200 ants to give orders for, you can't do it in less than 30ms because that loop in the engine will have to go around 3 times to read 200 orders plus a "go" line 2011-12-04T02:11:20 so the engine is still sleeping when you get to "go" 2011-12-04T02:11:53 i don't get what i'm calling? 2011-12-04T02:12:03 what do i pass to clock_getred? 2011-12-04T02:12:05 getres? 2011-12-04T02:12:31 antimatroid: int res = clock_getres(CLOCK_MONOTONIC, NULL); 2011-12-04T02:12:50 and try using clock_gettime() to measure time 2011-12-04T02:13:17 CLOCK_MONOTONIC isn't wall clock time, it's an arbitrary number which increases at the same rate as wall clock time 2011-12-04T02:13:38 CLOCK_REALTIME is the actual time, but the res is probably lower 2011-12-04T02:13:58 oh 2011-12-04T02:15:05 and you have to define _C_POSIX_SOURCE to a year/month and use -lrt to compile 2011-12-04T02:15:21 #define _POSIX_C_SOURCE 200809L 2011-12-04T02:15:35 before your includes 2011-12-04T02:15:53 anyway, I think it's just the time.sleep(0.01) the engine does 2011-12-04T02:16:01 *** treeform has quit IRC (Remote host closed the connection) 2011-12-04T02:16:11 is >8kb of orders data for a single turn in the first 100 turns actually likely? 2011-12-04T02:16:27 http://pastebin.com/FgY7pAZK 2011-12-04T02:16:35 i get that when i try to do int res = clock_getres(CLOCK_MONOTONIC, NULL); 2011-12-04T02:17:02 err 2011-12-04T02:17:05 since you have to completely fill the OS buffers before that sleep will have any effect on this timing that antimatroid is doing 2011-12-04T02:17:06 but with gettimeofday 2011-12-04T02:17:17 so what is a ClOCK_MONOTONIC? 2011-12-04T02:17:26 antimatroid: are you using -lrt and defining _POSIX_C_SOURCE? 2011-12-04T02:17:32 i have no idea what you guys are going on about with timer stuff :P 2011-12-04T02:17:47 thestinger: I couldn't find -lrt in code::blocks 2011-12-04T02:17:50 i'll look agian 2011-12-04T02:17:50 janzert: I think you are wrong. what if the engine has just issued the sleep when his code submits it's orders? 2011-12-04T02:17:58 antimatroid: add it to the end of the compiler command 2011-12-04T02:18:07 it'll all go into the OS buffers until they fill 2011-12-04T02:18:13 ChrisH_: but i submit the orders when i find them, it should happen in other places if that's the case 2011-12-04T02:18:38 but doesn't the flush block until the pipe is read from? 2011-12-04T02:18:55 if not then ignore me 2011-12-04T02:19:02 I don't think so 2011-12-04T02:19:09 then my bot would take forever to run :P 2011-12-04T02:19:10 ok, nevermind then 2011-12-04T02:19:44 timespec start_time; 2011-12-04T02:19:45 thestinger: i can't find the compiler command, only flags to tic box 2011-12-04T02:19:46 clock_gettime(CLOCK_MONOTONIC, &start_time); 2011-12-04T02:19:57 thestinger: why, do you flush after every line or something 2011-12-04T02:20:05 ChrisH_: yes, atm I do 2011-12-04T02:20:18 ok, that's curious 2011-12-04T02:20:43 you're half right actually 2011-12-04T02:20:55 it'll flush the bots buffer, but still keep the engine buffer 2011-12-04T02:21:10 so 4kb of order data is probably the actual cutoff 2011-12-04T02:21:52 *** janzert has quit IRC (Read error: Connection reset by peer) 2011-12-04T02:21:54 yeah, but that's a LOT of orders 2011-12-04T02:22:21 interesting 2011-12-04T02:22:26 i think the timer might be borked 2011-12-04T02:22:36 i it just skipped from 1 to 4 ms 2011-12-04T02:22:38 it's not unheard of 2011-12-04T02:22:40 when just couting get time 2011-12-04T02:23:04 *** janzert has joined #aichallenge 2011-12-04T02:23:08 thestinger: are you willing to share your timer code? 2011-12-04T02:23:15 I don't have any atm 2011-12-04T02:23:17 but 2011-12-04T02:23:26 use uintmax_t to do any conversions/calculations 2011-12-04T02:23:27 not double 2011-12-04T02:24:10 antimatroid: are you just using what you put in the C++ starter? 2011-12-04T02:24:17 yeah 2011-12-04T02:24:24 except stripping that junk for windows 2011-12-04T02:25:08 I think it might be wrong 2011-12-04T02:25:15 startTime = (double)timer.tv_sec+((double)timer.tv_usec/1000000.0); 2011-12-04T02:25:30 you probably want usec accuract, so you want to turn tv_sec into usec 2011-12-04T02:25:33 accuracy* 2011-12-04T02:27:48 sec 2011-12-04T02:29:39 *** mleise has joined #aichallenge 2011-12-04T02:31:28 antimatroid: http://pastie.org/2963294 try this 2011-12-04T02:31:51 gets microsecond accuracy rather than millisecond though 2011-12-04T02:32:42 hopefully it doesn't ever overflow... I didn't really think it over :P 2011-12-04T02:33:16 oh, comment is wrong now 2011-12-04T02:33:25 it seems to work better 2011-12-04T02:34:47 I think it might overflow though 2011-12-04T02:35:35 yeah i changed those to doubles and converted to milliseconds 2011-12-04T02:35:50 i probably don't need the double part 2011-12-04T02:35:58 oh, I know what it should do 2011-12-04T02:36:20 you don't need to convert timeval to something else 2011-12-04T02:36:27 that's the proper way, I forgot 2011-12-04T02:37:04 *** Jak_o_Shadows has joined #aichallenge 2011-12-04T02:38:24 what? 2011-12-04T02:38:32 like you just store timeval start_time 2011-12-04T02:38:34 sec 2011-12-04T02:38:38 http://pastebin.com/fgeBVLht 2011-12-04T02:38:40 that seems to work 2011-12-04T02:38:51 oh yeah 2011-12-04T02:38:52 okayt 2011-12-04T02:39:01 I think that will overflow int though 2011-12-04T02:39:44 if it worked with double? 2011-12-04T02:39:47 long in? 2011-12-04T02:39:52 sec 2011-12-04T02:40:34 http://pastie.org/private/75fvn0s53b2ow830ffzcvq 2011-12-04T02:40:35 *** treeform has joined #aichallenge 2011-12-04T02:40:36 that's better 2011-12-04T02:40:56 really long line but whatever 2011-12-04T02:41:20 that won't ever overflow because you leave it as internal types until you want the delta 2011-12-04T02:41:59 tv_sec is a long int (seconds since 1970 or whatever) 2011-12-04T02:42:07 so if you multiply it by a number, that's huge 2011-12-04T02:42:32 http://pastebin.com/vjDgUPbs 2011-12-04T02:42:34 how about that? 2011-12-04T02:42:54 well, you want ms right? 2011-12-04T02:42:58 yeah 2011-12-04T02:43:15 my last one gives it 2011-12-04T02:43:35 http://pastie.org/2963324 2011-12-04T02:43:48 in ms now 2011-12-04T02:43:50 *** SUITOR has joined #aichallenge 2011-12-04T02:44:13 done 2011-12-04T02:44:15 i like it 2011-12-04T02:44:20 cheers 2011-12-04T02:44:44 now to see if that was my problem :P 2011-12-04T02:46:18 *** SUITOR has quit IRC () 2011-12-04T02:47:39 *** Palmik has joined #aichallenge 2011-12-04T02:49:25 anyway the return type doesn't have to be uintmax_t now since it's just ms 2011-12-04T02:50:55 is uintmax_t == unsigned long ? 2011-12-04T02:51:00 no 2011-12-04T02:51:11 it's probably 128 bit 2011-12-04T02:51:13 xD 2011-12-04T02:51:36 UINTMAX_MAX 2011-12-04T02:51:41 is the macro for it, sec 2011-12-04T02:52:17 oh, it's just 64-bit 2011-12-04T02:52:25 on x86_64 gcc 2011-12-04T02:52:38 the idea is that it's the largest int the implementation could use for something 2011-12-04T02:52:55 *** CowTipperVirus has joined #aichallenge 2011-12-04T02:53:12 so yeah, it's uint64_t which is the same as unsigned long long or unsigned long 2011-12-04T02:53:20 "for something" may exclude sse numbers when they can't be handled the same way as ints 2011-12-04T02:53:25 Hey guys, has anybody else had the ai viewer mess up (not displaying most buttons) w/ the latest firefox? 2011-12-04T02:53:59 CowTipperVirus: lol, stuff like this happens with every major release starting from version 4 2011-12-04T02:54:09 *** praveen has quit IRC (Ping timeout: 265 seconds) 2011-12-04T02:54:13 CowTipperVirus: no, it works fine 2011-12-04T02:54:29 but I may just have to extend the filter for the shadow effect 2011-12-04T02:54:30 thestinger: should i use int or uintmax_T? 2011-12-04T02:54:39 just leave it as uintmax_t 2011-12-04T02:54:43 k 2011-12-04T02:54:45 mleise: mean you have the bug too? 2011-12-04T02:54:53 i mean disable shadows in that version of FF. 2011-12-04T02:55:14 no i don't have the latest FireFox, but as the developer of the visualizer I know the issue 2011-12-04T02:55:28 mleise: cool, thanks, I'll give it a shot 2011-12-04T02:55:40 which version of FF is it exactly? 2011-12-04T02:55:58 oh wait, i need double 2011-12-04T02:56:02 8.0.1 is the latest stable :P 2011-12-04T02:56:04 thestinger: double will work fine yeah? 2011-12-04T02:56:10 It is a Windows only problem btw. since they introduced a new renderer for Windows a while back, so thestinger wont see it on Linux 2011-12-04T02:56:18 mleise: latest stable windows sys, 8.0.1 2011-12-04T02:56:18 i pass functions a maxTimeProportion 2011-12-04T02:56:25 antimatroid: yeah, that will be fine 2011-12-04T02:56:28 since it's just ms 2011-12-04T02:56:33 just making sure :P 2011-12-04T02:56:38 i don't trust myself at this point 2011-12-04T02:56:45 mleise: btw, works when it's not fullscreen 2011-12-04T02:56:46 just leave the internal thing the same 2011-12-04T02:57:05 like, start() stores a timeval and then getTime() gives you the delta based on a new timeval 2011-12-04T02:57:28 you can make the return type anything that can store the max turn time in ms 2011-12-04T02:57:29 antimatroid: the times come from a hardware clock. it is impossible that you get the numbers as 134352355235-bit integers from there ^^ 2011-12-04T02:58:36 and my problem is still there :( 2011-12-04T02:58:37 CowTipperVirus: Not really. It is a matter of size 2011-12-04T02:58:54 You may notice some pixels are missing on the buttons on the left 2011-12-04T02:59:01 antimatroid: is it just a 10ms delay? 2011-12-04T02:59:08 *** CowTipperVirus_ has joined #aichallenge 2011-12-04T02:59:11 it was 7 that time 2011-12-04T02:59:17 try only flushing at the end 2011-12-04T02:59:26 like replace endl; with '\n'; except at the end 2011-12-04T02:59:38 mleise: disabling hardware accel fixed it 2011-12-04T02:59:43 I couldn't figure out which was better 2011-12-04T02:59:46 the 7 second delay is happening with a "cout << endl;" line 2011-12-04T02:59:55 i don't like flushing just at the end either 2011-12-04T02:59:58 CowTipperVirus_: nice workaround ;) 2011-12-04T03:00:07 7ms* 2011-12-04T03:00:17 mleise: what was your suggestion about shadows? prefer to use the < hammer approach if I can :) 2011-12-04T03:01:24 *** CowTipperVirus has quit IRC (Ping timeout: 265 seconds) 2011-12-04T03:02:11 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has quit IRC (Quit: Lost terminal) 2011-12-04T03:03:21 antimatroid: you get 7 ms delay when flushing? 2011-12-04T03:03:46 mleise: there's something wrong with my bot 2011-12-04T03:03:48 do you send more than 64kb over the pipe? 2011-12-04T03:03:51 and i can't work out what 2011-12-04T03:04:04 i have no idea what the problem is:P 2011-12-04T03:04:15 i had it doing 1000ms last night 2011-12-04T03:04:20 same problem 2011-12-04T03:04:33 so.. why don't you run it in a profiler? 2011-12-04T03:04:51 use no inlining and keep the line number information 2011-12-04T03:05:33 that would require me to go and learn how to use a profiler :P 2011-12-04T03:05:38 If you are on Linux, then OProfile is a good choice, although it is a bit overwhelming at first 2011-12-04T03:05:39 and take out all my inline functions etc. 2011-12-04T03:05:58 i'm going to just comb my source code :P 2011-12-04T03:06:05 i'm sure i'm doing something retarded somewhere to cause it 2011-12-04T03:06:13 antimatroid: ok, then keep the inline functions. It's just that they wont show up seperately in the output then 2011-12-04T03:06:49 ok :) 2011-12-04T03:07:32 mleise: just poking again, what was your shadow suggestion? 2011-12-04T03:07:54 you can disable inlining with a compiler argument 2011-12-04T03:08:39 gcc rarely cares about your views on inlining anyway :) 2011-12-04T03:08:45 I thought he was using C#, don't think it has many compiler args 2011-12-04T03:09:16 MyBot.cpp:133:8: warning: inlining failed in call to ‘Steps::~Steps()’: --param inline-unit-growth limit reached [-Winline] 2011-12-04T03:09:44 that's the default destructor for a tiny struct :) 2011-12-04T03:10:04 yeah... as soon as this competition is over I'm going to put in a push request for their compiler flags... the C/C++ -o3 -funroll-loops is just retarded 2011-12-04T03:10:18 seriously... -o2 -march=native 2011-12-04T03:10:32 maybe... MAYbe some other flags 2011-12-04T03:10:37 -O3 is better for "small" programs like this 2011-12-04T03:10:38 *** analyst74 has joined #aichallenge 2011-12-04T03:10:58 CowTipperVirus_: I've fixed the problem in the source code repository 2011-12-04T03:11:01 got a sample of code I was playing around w/ where the funroll loops slows it down on my pc by a factor of 10 or more 2011-12-04T03:11:01 -funroll-loops is a really bad thing most of the time without PGO 2011-12-04T03:11:08 CowTipperVirus_: yeah, but -O3 is fine 2011-12-04T03:11:24 thestinger: meh... doubt it 2011-12-04T03:11:27 it's just not fine to compile a massive program like firefox completely with -O3 and no PGO 2011-12-04T03:11:40 gcc manual doesn't recommend it unless you can prove your app 2011-12-04T03:11:52 it makes optimizations that may not really help 2011-12-04T03:11:57 http://gcc.gnu.org/onlinedocs/gcc/Optimize-Options.html 2011-12-04T03:12:40 it doesn't do anything bad at -O3 unless the program is too big for gcc to properly analyze it 2011-12-04T03:12:51 -funroll-loops is also losing weight with the processors becoming smarter about loops 2011-12-04T03:13:00 yeah, -funroll-loops is often bad 2011-12-04T03:13:17 if you compile with PGO, gcc will unroll loops where it's a good thing 2011-12-04T03:13:24 you shouldn't ever explicitly enable it without testing 2011-12-04T03:13:27 yeah, I agree w/ the link, but some of the optimizations enabled by -o3 should have warnings 2011-12-04T03:13:40 PGO? 2011-12-04T03:13:49 profile-guided optimization 2011-12-04T03:14:13 build with -fprofile-generate, run the program through a benchmarking profile, then rebuild with -fprofile-use 2011-12-04T03:14:31 firefox is built like that by default 2011-12-04T03:14:32 interesting, never heard of that... will have to look into it :) 2011-12-04T03:14:45 it lets the compiler actually see which branches are taken most 2011-12-04T03:14:48 which code is hot/cold 2011-12-04T03:14:55 I recently see more and more software use -O3 in the make file. Sounds like it is stable now and doing more good than bad on the average. Or people really just want to write -O9999 and not care about the consequences. 2011-12-04T03:14:55 yeah, makes tons of sense 2011-12-04T03:14:58 which functions should be inlined, which loops to unroll 2011-12-04T03:15:03 like a JIT, but compiled ahead of time 2011-12-04T03:15:24 mleise: anything over -O3 is meaningless with gcc 2011-12-04T03:15:39 lol, yeah... :) has the - - -omg_just_go_fast look to it 2011-12-04T03:15:43 I tried PGO once and it *failed*. The program became slower than without PGO. 2011-12-04T03:15:43 with clang -O4 is basically an alias for '-O3 -flto' 2011-12-04T03:16:02 mleise: which compiler? 2011-12-04T03:16:07 I think gcc just got it with 4.6 2011-12-04T03:16:33 but yeah... the ai challenge has a bunch of retarded flags set, big bummer 2011-12-04T03:16:43 and of all the flags, seriously... not setting march 2011-12-04T03:16:45 thestinger: some older version of gcc 2011-12-04T03:16:57 oh 2011-12-04T03:17:14 if I had noticed earlier in the competition I would have requested a change, but so far in I think it would be too big of a change for them to do anything 2011-12-04T03:17:16 sorry, link-time optimization first appeared in 4.6 2011-12-04T03:17:18 I got confused xD 2011-12-04T03:17:55 CowTipperVirus_: -O3 is definitely sane for a small program, but the way they compile isn't very good 2011-12-04T03:17:58 *** Palmik has quit IRC (Read error: Connection reset by peer) 2011-12-04T03:18:01 are link time optimizations part of ld then? 2011-12-04T03:18:05 well 2011-12-04T03:18:08 or part of g++ 2011-12-04T03:18:21 when gcc compiles a .o, it loses all the internal representation 2011-12-04T03:18:29 so you don't get optimization for your whole program 2011-12-04T03:18:41 a function in one .o cannot be inlined into another, or changed based on what happens in other files 2011-12-04T03:18:50 thestinger: k, I'll let you have that one 2011-12-04T03:19:08 with -flto, it basically keeps all the information in the .o so it can do optimization across the whole program 2011-12-04T03:19:43 CowTipperVirus_: -funroll-loops is a bad idea though 2011-12-04T03:19:44 You know what is funny? That byte-code languages now have features that make them faster than native compiled languages. Like devirtualizing method calls 2011-12-04T03:19:59 that's what PGO is for :) 2011-12-04T03:20:36 *** Palmik has joined #aichallenge 2011-12-04T03:20:44 I'm sure JIT compilers will eventually be faster than AOT ones though, because you can't just distribute your code with -march=native 2011-12-04T03:20:54 that's interesting. we need that stuff in D some day ;) 2011-12-04T03:21:11 llvm actually has a JIT, so there's a C++ JIT/REPL now 2011-12-04T03:21:23 http://root.cern.ch/drupal/content/cling 2011-12-04T03:22:31 CowTipperVirus_: if people want -funroll-loops they can actually enable it with a pragma 2011-12-04T03:22:46 http://gcc.gnu.org/onlinedocs/gcc/Function-Specific-Option-Pragmas.html 2011-12-04T03:23:04 thestinger: I agree... which is why I'm going to try and get them to change it in the next competition 2011-12-04T03:23:18 thestinger: because I can't "un-enable" it very easily 2011-12-04T03:23:23 CowTipperVirus_: are you using C++? 2011-12-04T03:23:30 thestinger: yeah 2011-12-04T03:23:33 it's not enabled for the C++11 build 2011-12-04T03:23:38 just switch your extension to .cpp 2011-12-04T03:23:47 *** epicmonkey has joined #aichallenge 2011-12-04T03:23:54 and use -std=c++0x locally 2011-12-04T03:24:12 or -std=c++11 if you're using gcc 4.7 like I am :) 2011-12-04T03:24:34 and clang 3.0 2011-12-04T03:24:43 thestinger: didn't even think to try that... good workaround. for now I just changed my code so the compiler wouldn't figure out the size of the loop at compile time 2011-12-04T03:24:45 maybe he uses MSVC++? 2011-12-04T03:25:01 CowTipperVirus_: YOU DID WHAT? 2011-12-04T03:25:05 well, C++03 will still compile as C++11 2011-12-04T03:25:15 unless you actually use 'auto' with the old meaning 2011-12-04T03:25:25 mleise: well... the funroll-loops was owning it 2011-12-04T03:25:34 can't unroll something if you don't know what it is :) 2011-12-04T03:25:35 hmm. ok. 2011-12-04T03:25:50 CowTipperVirus_: you an also use a pragma to disable it 2011-12-04T03:26:03 -fno-unroll-loops ? 2011-12-04T03:26:09 yeah 2011-12-04T03:26:12 you use the optimize pragma 2011-12-04T03:26:18 thestinger: looked into that, was pretty sure the cl flag would override it 2011-12-04T03:26:19 ...you can just upload your own makefile 2011-12-04T03:26:32 pairofdice: no, the server doesn't use your Makefile 2011-12-04T03:26:33 Can't you? 2011-12-04T03:26:34 pairofdice: doesn't matter... look at their code, won't use your makefile 2011-12-04T03:26:58 they compile each file into a .o with -O3, then link it together with -O2 2011-12-04T03:26:58 Weird, it's said on the website that it should use one if provided 2011-12-04T03:27:14 nice! can you provide a link? 2011-12-04T03:27:21 would love to get them to do that 2011-12-04T03:27:26 https://github.com/aichallenge/aichallenge/blob/epsilon/worker/compiler.py#L320 2011-12-04T03:27:29 lol, ofc...then people would do the thread 2011-12-04T03:27:53 what's command -parallelize-loops or something 2011-12-04T03:28:05 just use a .cpp extension and it won't use -funroll-loops :P 2011-12-04T03:28:53 or push no-unroll-loops onto the optimize stack 2011-12-04T03:29:12 wow... wtf, that's the only diff between the two languages 2011-12-04T03:29:19 seriously... why would they be different??? 2011-12-04T03:29:38 well, using -funroll-loops was a bad idea 2011-12-04T03:29:50 but maybe they don't want to change it before the next contest 2011-12-04T03:29:56 but they didn't use it for the new C++11 option 2011-12-04T03:30:13 yeah, and I'm not going to petition for that... because of that exact reason 2011-12-04T03:30:15 yeah 2011-12-04T03:30:40 iirc, amstan said they are going to make C++11 and C99 the default for the next contest anyway 2011-12-04T03:30:45 how can your code push something on the optimize stack or override cl options anyway? 2011-12-04T03:31:10 http://gcc.gnu.org/onlinedocs/gcc/Function-Specific-Option-Pragmas.html <- like that 2011-12-04T03:32:25 #pragma GCC optimize ("no-unroll-loops") 2011-12-04T03:32:26 I think 2011-12-04T03:32:29 thestinger: cool... thanks 2011-12-04T03:32:55 test it though, you might also have to push it 2011-12-04T03:32:59 you can also set a global optimization level 2011-12-04T03:33:22 thestinger: will def look into that to see if it helps. doubt it'll get much saving from letting the compiler know the loop time, but worth a shot 2011-12-04T03:33:31 #pragma GCC optimize ("O2") 2011-12-04T03:33:39 etc. 2011-12-04T03:33:45 thestinger: now I'm almost positive that'll get overriden 2011-12-04T03:33:49 ridden* 2011-12-04T03:33:55 #pragma GCC push_options 2011-12-04T03:33:57 #pragma GCC optimize ("O0") 2011-12-04T03:33:59 like that 2011-12-04T03:34:02 I think you need both lines 2011-12-04T03:34:11 then you can #pragma GCC pop_options to undo them 2011-12-04T03:34:23 so you can set per-function, per-block, etc. optimization 2011-12-04T03:35:00 wow, have to be pretty hardcore to profile that much of your code 2011-12-04T03:35:57 http://stackoverflow.com/a/2188780 there's also that 2011-12-04T03:36:13 CowTipperVirus_: well, in the linux kernel they use the gcc attributes _everywhere_ 2011-12-04T03:36:16 and profile everything 2011-12-04T03:36:53 #define likely(x) __builtin_expect((x),1) 2011-12-04T03:36:55 #define unlikely(x) __builtin_expect((x),0) 2011-12-04T03:37:04 and then you tell the compiler if a condition will be likely or unlikely 2011-12-04T03:37:10 if (likely(some_condition)) { } 2011-12-04T03:37:21 obviously need to profile 2011-12-04T03:37:41 yeah... way too much work 2011-12-04T03:38:07 I could see it mattering there though 2011-12-04T03:38:47 http://gcc.gnu.org/onlinedocs/gcc-4.1.2/gcc/X86-Built_002din-Functions.html#X86-Built_002din-Functions 2011-12-04T03:39:02 that lets you use SSE vector stuff, etc. 2011-12-04T03:40:16 http://gcc.gnu.org/onlinedocs/gcc-4.1.2/gcc/Other-Builtins.html 2011-12-04T03:40:30 prefetching stuff to cache, etc. 2011-12-04T03:40:55 yeah, my understanding is those should come automatically if you allow march compiler options 2011-12-04T03:40:59 Looks like xathis is about to lose #1 2011-12-04T03:41:21 CowTipperVirus_: compilers are still really bad at vectorization 2011-12-04T03:41:23 pairofdice: oh noes... he'll have to write v2 2011-12-04T03:41:26 lol 2011-12-04T03:41:28 heh 2011-12-04T03:41:52 like x86_64 implies a level of SSE support 2011-12-04T03:42:13 but you still have to do it manually if you actually want it to be used for stuff other than memset, etc. 2011-12-04T03:42:22 maybe clang will be better 2011-12-04T03:42:41 clang/llvm still make slower code than gcc most of the time though :) 2011-12-04T03:42:57 yeah... I'm an electrical engineer by trade, power electronics, mostly the hardware and controls side 2011-12-04T03:43:14 the latter forces me to dabble in code occasionally, not my main job though, and others are way better at it 2011-12-04T03:43:26 anyway, writing code for a dsp I have to do stupid shit like that all the time 2011-12-04T03:43:48 the compilers are insanely bad at optimizing on those things 2011-12-04T03:44:06 pairofdice: yeah, I think pguillory will pass him soon. 2011-12-04T03:45:24 I haven't really bothered trying to optimize at all yet beyond changing my algorithms 2011-12-04T03:46:03 yeah, realistically I'm not taking much of a penalty, so I don't care 2011-12-04T03:46:12 I'm lazy, so I'm just using std::unordered_set, std::unordered_map, 2D std::vector, etc. everywhere 2011-12-04T03:46:16 but the loops thing... omg... like I said, 10x - 20x is noticable 2011-12-04T03:46:39 yeah, it slows down my bot a lot too 2011-12-04T03:46:58 the std::vector is actually really fast if you compile w/ the -no-exceptions flag 2011-12-04T03:47:02 err -fno-? 2011-12-04T03:47:07 yeah 2011-12-04T03:47:12 well 2011-12-04T03:47:33 the thing is that I could have a 2D array as a big block of memory 2011-12-04T03:47:46 and instead I have pointers to arrays, which probably slows it down a lot 2011-12-04T03:48:05 because I'm not resizing any of the arrays 2011-12-04T03:48:17 if it's known at compile time, that should go away in optimizations I'd assume 2011-12-04T03:48:26 it doesn't 2011-12-04T03:48:40 gcc won't change container types/semantics 2011-12-04T03:48:49 yeah... I'm also allocating a bunch of arrays on the heap, might be able to drop some on the stack for performance 2011-12-04T03:49:02 dunno... may actually hurt me 2011-12-04T03:49:05 well I would still put them on the heap 2011-12-04T03:49:17 but as a 1D array with row * cols + col indexing 2011-12-04T03:49:21 instead of pointers to arrays 2011-12-04T03:49:28 yeah 2011-12-04T03:49:28 pairofdice: pguillory should get another game in about 45 minutes, and I'm expecting one in about 30, so it should be interesting to see what happens 2011-12-04T03:49:42 Yeah 2011-12-04T03:50:02 You both have higher mu than xathis 2011-12-04T03:50:06 xathis's last game was the same as mine, so he's due in 30 mins also 2011-12-04T03:50:09 thestinger: thanks for the convo, out for the night 2011-12-04T03:50:13 cya 2011-12-04T03:50:27 sweet, g'luck ChrisH 2011-12-04T03:50:43 i think my bot is about equal or slightly worse than xathis right now, but pguillory is clearly better, it's just a matter of time 2011-12-04T03:50:55 *** CowTipperVirus_ has quit IRC (Quit: Page closed) 2011-12-04T03:51:37 but who knows what xathis has been working on for the last two months. he'll probably put a upload a super bot just before finals and whip us all 2011-12-04T03:51:49 he's had his bot on tcp sometimes 2011-12-04T03:51:58 I think he definitely worked on it a lot more 2011-12-04T03:52:05 yeah, i've seen that but haven't looked at the games 2011-12-04T03:52:41 i've got a few bug fixes and improvements ready for my next version as well, so we'll see 2011-12-04T04:00:10 heh, pguillory is using Go too now - I didn't realize that 2011-12-04T04:00:45 I'm liking C++ a lot now that I've learned the C++11 stuff :) 2011-12-04T04:01:34 yeah, he switched mid stream, I saw him comment in IRC that it wasn't as fast, but he liked it was easier to work with so he was switching anyway 2011-12-04T04:02:07 C++11 has local type inference like Go (auto v = something() instead of v := something()) 2011-12-04T04:02:10 and lambdas/closures 2011-12-04T04:02:18 so I'm not really missing anything from python/Go 2011-12-04T04:02:24 we have to do more hand optimization in Go than you C++ guys, but luckily there is an easy to use profiler built into Go that makes it easy to find the hot spots 2011-12-04T04:02:24 it's just really verbose/ugly :) 2011-12-04T04:02:39 yeah, I learned C++ for this contest 2011-12-04T04:02:42 I already knew Go 2011-12-04T04:02:54 cool 2011-12-04T04:03:27 I did PW in Java and it was fast enough, but syntactially a poor fit for this kind of coding, so it was some ugly looking stuff 2011-12-04T04:03:35 yeah 2011-12-04T04:03:48 but I'm liking Go for the most part 2011-12-04T04:03:53 *** Euphony has joined #aichallenge 2011-12-04T04:03:55 I wouldn't be having much fun if I was using the pre-C++11 language 2011-12-04T04:04:19 I've ended up using lambdas a _lot_ because I don't like using classes for code-reuse most of the time 2011-12-04T04:04:36 well, I also don't really understand how you would do reuse via classes for lots of stuff 2011-12-04T04:06:25 the syntax is definitely pretty bad though 2011-12-04T04:07:32 yeah, this sort of code strikes me as more functional/procedural in nature 2011-12-04T04:08:00 thestinger: that's because object orientation is an anti-pattern :) 2011-12-04T04:08:28 bmh: well, OOP is fine but not the Java/C++ style hurrr make a singleton style OOP :P 2011-12-04T04:08:29 bmh: that depends how you use it 2011-12-04T04:08:51 ChrisH_: I am a functional programming warrior ;) 2011-12-04T04:09:05 * bmh gets paid to write Java. 2011-12-04T04:09:06 i'm a multi-paradigm guy myself 2011-12-04T04:09:06 bmh: I think you would be pretty amused by how I'm coding C++ 2011-12-04T04:09:29 lambdas/higher-order functions everywhere, everything that's not a container/loop-control var is basically const 2011-12-04T04:09:33 etc. 2011-12-04T04:09:48 but one thing I agree on, deep inheritance structures are a red flag 2011-12-04T04:09:58 meh, inheritance 2011-12-04T04:10:24 composition all the way :) 2011-12-04T04:10:33 but i like encapsulating behaviors and data for lots of reasons, but not so much for these ai contests 2011-12-04T04:11:11 and i also like functional stuff when appopriate 2011-12-04T04:11:26 thestinger: do you know about closures in GNU C? 2011-12-04T04:11:49 I thought they were just nested functions? 2011-12-04T04:12:13 C++ actually has proper lambdas that can capture state, and I can abuse them with references 2011-12-04T04:12:46 [this](Location loc) { grid[loc.row][loc.col].last_seen = turn; } 2011-12-04T04:12:48 yeah... they are just nested function. It's neat to avoid polluting the namespace. 2011-12-04T04:12:59 That's an abomination. 2011-12-04T04:13:06 yeah. 2011-12-04T04:13:16 it lets me have a single function for painting on radius info 2011-12-04T04:13:21 if I can make my bot do something *other* than crash, I'm still hoping for "best Haskell bot" 2011-12-04T04:14:07 I haven't thought of a combat strategy better than "sample until I run out of time" 2011-12-04T04:14:18 also it only infers the return type with a single statement... 2011-12-04T04:14:34 [](int i)->int { return i * 2; }; 2011-12-04T04:14:37 that's the full syntax 2011-12-04T04:14:43 thestinger: do you ever sleep? 2011-12-04T04:14:49 antimatroid: what's sleep? 2011-12-04T04:14:56 it's glorius 2011-12-04T04:15:00 I'm sick atm actually, so I'm not really sleeping 2011-12-04T04:15:00 glorious 2011-12-04T04:15:09 What timezone are you in? 2011-12-04T04:15:22 EST 2011-12-04T04:15:31 jeeze. (PST represent) 2011-12-04T04:15:36 EST? 2011-12-04T04:15:40 EEST 2011-12-04T04:15:43 +? 2011-12-04T04:16:06 UTC-5 atm 2011-12-04T04:16:12 unless they move DST around again :) 2011-12-04T04:16:14 EST = +11GMT == eastern standard time to me 2011-12-04T04:16:25 well +10, but atm we have daylight savings 2011-12-04T04:16:46 I actually slept from like 5pm to 8pm to I'm pretty much screwed now 2011-12-04T04:16:48 'eh? EST is like GMT-5 or so 2011-12-04T04:16:58 so I'm pretty much screwed now* 2011-12-04T04:17:03 bmh: yeah -5 2011-12-04T04:17:10 -4 when it's EDT 2011-12-04T04:17:45 I once worked on a financial business app. There was a lot of mucking with timezones and market openings and closings. 2011-12-04T04:17:52 Not fun. (Praise be to Joda, though) 2011-12-04T04:17:57 eww, time 2011-12-04T04:18:11 sleep time 2011-12-04T04:18:16 *** bmh has quit IRC (Quit: Page closed) 2011-12-04T04:19:47 *** Jak_o_Shadows1 has joined #aichallenge 2011-12-04T04:21:06 well, I just lost to FlagCapper and Murashka. 2011-12-04T04:22:33 the version I uploaded on nov 30th seems to be doing well on the official site :) 2011-12-04T04:22:48 http://aichallenge.org/profile.php?user=2255 sigma just needs to drop a bit 2011-12-04T04:22:56 *** Jak_o_Shadows has quit IRC (Ping timeout: 248 seconds) 2011-12-04T04:22:59 definitely way better than my pre-combat bot xD 2011-12-04T04:23:22 yeah, combat makes a huge difference 2011-12-04T04:23:43 my last bot actually had .1 more skill than this one does 2011-12-04T04:23:51 but that's only because sigma dropped so low 2011-12-04T04:23:56 because I was keeping it activated 2011-12-04T04:24:11 the competition is a little stronger now too 2011-12-04T04:24:28 yeah 2011-12-04T04:24:31 tcp is a lot more intense 2011-12-04T04:24:44 it seems like everyone has combat now on tcp 2011-12-04T04:26:25 maybe it's only because they're using 4s per turn... 2011-12-04T04:26:42 maybe 2011-12-04T04:26:56 gotta go, bye 2011-12-04T04:27:05 *** ChrisH_ has quit IRC (Quit: Page closed) 2011-12-04T04:28:59 *** Jak_o_Shadows1 has quit IRC (Remote host closed the connection) 2011-12-04T04:29:28 *** dzukpp has joined #aichallenge 2011-12-04T04:31:01 *** choas has joined #aichallenge 2011-12-04T04:31:51 *** Jak_o_Shadows has joined #aichallenge 2011-12-04T04:31:58 *** Ashoka has quit IRC (Quit: Page closed) 2011-12-04T04:38:18 http://llvm.org/docs/CodingStandards.html#micro_anonns <- I need to do this for my classes 2011-12-04T04:44:22 *** Jak_o_Shadows has quit IRC (Read error: Connection reset by peer) 2011-12-04T04:46:44 *** artart78_ is now known as artart78 2011-12-04T04:47:27 *** antimatroid has quit IRC (Ping timeout: 240 seconds) 2011-12-04T04:52:49 In the last game, pguillory got another step closer to overtaking xathis. If pguillory overtakes him, maybe xathis will upload a new bot. 2011-12-04T04:54:35 *** antimatroid has joined #aichallenge 2011-12-04T04:58:21 *** bretep has joined #aichallenge 2011-12-04T05:01:55 *** treeform has quit IRC (Remote host closed the connection) 2011-12-04T05:05:16 *** foRei has joined #aichallenge 2011-12-04T05:10:50 hmmm 2011-12-04T05:10:53 thestinger: you still here? 2011-12-04T05:11:19 yeah 2011-12-04T05:12:02 i think the timer is still borked 2011-12-04T05:12:07 or my computer is borked 2011-12-04T05:13:03 I think it's fine now 2011-12-04T05:13:09 return (current_time.tv_sec - start_time.tv_sec) * 1000 + 2011-12-04T05:13:11 (current_time.tv_usec - start_time.tv_usec) / 1000; 2011-12-04T05:13:12 *** antimatroid has quit IRC (Quit: Leaving.) 2011-12-04T05:13:19 *** antimatroid has joined #aichallenge 2011-12-04T05:13:20 *** grwip has joined #aichallenge 2011-12-04T05:13:35 it just converts the sec delta to msec and usec to msec 2011-12-04T05:13:41 shouldn't be any overflows or anything 2011-12-04T05:14:31 http://pastebin.com/B5NR7S94 2011-12-04T05:14:42 if i compile that ant run like "./antibot > out.txt" 2011-12-04T05:14:50 it occasionally just jumps 2011-12-04T05:15:39 the loop is probably faster than the clock resolution 2011-12-04T05:17:08 what CPU do you have? 2011-12-04T05:18:35 *** antimatroid1 has joined #aichallenge 2011-12-04T05:18:41 the kernel could just be doing a context switch and letting something else run on that core for a bit 2011-12-04T05:19:20 works for me :P 2011-12-04T05:19:32 *** antimatroid has quit IRC (Ping timeout: 255 seconds) 2011-12-04T05:19:40 possibly 2011-12-04T05:20:01 but my bots problem happens at a specific place 2011-12-04T05:20:06 that can't be a cooincident 2011-12-04T05:23:15 *** Wraithan has joined #aichallenge 2011-12-04T05:27:59 *** antimatroid1 has left #aichallenge 2011-12-04T05:29:38 *** antimatroid has joined #aichallenge 2011-12-04T05:30:42 wtf 2011-12-04T05:30:46 it just jumped from 9 to 18 2011-12-04T05:30:57 i restarted too 2011-12-04T05:33:42 i guess i could compate with clock time and see if it is just my computer being a pita 2011-12-04T05:33:48 compare* 2011-12-04T05:33:58 it's multicore right? 2011-12-04T05:34:50 *** epicmonkey has quit IRC (Ping timeout: 255 seconds) 2011-12-04T05:36:02 nope 2011-12-04T05:36:08 oh 2011-12-04T05:36:15 pretty sure it's not 2011-12-04T05:36:29 you're running linux right? 2011-12-04T05:42:43 yeah, ubuntu 2011-12-04T05:43:03 *** kaemo has quit IRC (Ping timeout: 240 seconds) 2011-12-04T05:44:37 you could run your bot with higher cpu priority (nice -n 19 ./MyBot) but you'll need to be root or change limits.conf 2011-12-04T05:45:22 schedtool even lets you give something realtime priority, but then everything else will freeze while it's running if you only have 1 core :) 2011-12-04T05:45:28 *** epicmonkey has joined #aichallenge 2011-12-04T05:46:26 how do i be root? :P 2011-12-04T05:46:42 i'm not good with linux either :P 2011-12-04T05:48:53 well, iirc Ubuntu has root disabled so just do 'sudo -i' 2011-12-04T05:48:56 in a terminal 2011-12-04T05:53:11 that seems to have worked better 2011-12-04T05:53:40 oh 2011-12-04T05:53:50 I told you the backwards way of doing it xD 2011-12-04T05:53:55 nice -n 19 is lowest priority 2011-12-04T05:54:00 aha 2011-12-04T05:54:06 but it worked when i did that? 2011-12-04T05:54:11 *** mj41 has joined #aichallenge 2011-12-04T05:54:18 nice -n -20 is highest priority (lower is higher prio) 2011-12-04T05:55:15 antimatroid: well, maybe it gave the engine more time to do stuff :P 2011-12-04T05:55:15 *** amstan__ has quit IRC (Read error: Connection reset by peer) 2011-12-04T05:55:21 *** amstan__ has joined #aichallenge 2011-12-04T05:55:21 *** ChanServ sets mode: +o amstan__ 2011-12-04T05:55:41 antimatroid: if you really have 1 core I wouldn't worry about it 2011-12-04T05:55:52 your bot has to compete for cycles with the engine 2011-12-04T05:56:08 i don't mind that so much, i just want to make sure my bot isn't broken 2011-12-04T05:56:18 it's mildy annoying for debugging 2011-12-04T05:56:21 but i can live with it 2011-12-04T05:57:04 *** nick_ has joined #aichallenge 2011-12-04T05:57:08 hello 2011-12-04T05:57:30 *** nick_ is now known as Guest77607 2011-12-04T05:57:43 *** Guest77607 is now known as noctopuss 2011-12-04T05:58:06 i don't understand viewradius 2011-12-04T05:58:19 it's the radius each of your ants can see 2011-12-04T05:58:25 it should be a radius, but python visible function takes the square root of this value 2011-12-04T05:58:34 you receive information on tiles that are visible to your ants 2011-12-04T05:58:36 well 2011-12-04T05:58:39 you're passed viewradius2 2011-12-04T05:58:39 thestinger: i just dumped that value, it was 77 2011-12-04T05:58:42 not viewradius 2011-12-04T05:58:51 oh yes sorry 2011-12-04T05:58:56 viewradius2 2011-12-04T05:59:21 it's just viewradius^2, so it can be an integer 2011-12-04T05:59:22 thestinger: so what is viewradius2 ? 2011-12-04T05:59:28 oh 2011-12-04T05:59:41 why doing that ? 2011-12-04T05:59:48 why just not taking 2011-12-04T05:59:54 int(viewradius) 2011-12-04T06:00:11 because then it's not the same 2011-12-04T06:00:14 like attackradius2 is 5 2011-12-04T06:00:25 *** antimatroid has quit IRC (Read error: Operation timed out) 2011-12-04T06:00:26 so you can attack 2 tiles away forwards, but you can also attack 2 forwards and 1 to the side 2011-12-04T06:00:34 not the same as attackradius=2 2011-12-04T06:00:42 all right 2011-12-04T06:00:49 thx thestinger 2011-12-04T06:00:53 np 2011-12-04T06:03:03 *** mj41 has quit IRC (Ping timeout: 240 seconds) 2011-12-04T06:09:31 *** george_ has quit IRC (Quit: Page closed) 2011-12-04T06:17:13 *** mj41 has joined #aichallenge 2011-12-04T06:21:27 thestinger: i understand the fact that not having a round number allow to get a different range 2011-12-04T06:22:01 *** ikaros has joined #aichallenge 2011-12-04T06:22:03 thestinger: but I don't understand how I can have this exact range if if the sq of the number then the sqrt 2011-12-04T06:22:37 s/if th/i do th/ 2011-12-04T06:24:29 Err 2011-12-04T06:24:49 int(sqrt(77)) will ever do 8 .. isn't it ? 2011-12-04T06:25:29 why would you truncate it to an int? 2011-12-04T06:26:08 you can compare euclidean_distance2 and a radius2 2011-12-04T06:26:26 thestinger: i need to review some math bases ^^ 2011-12-04T06:26:29 thx :) 2011-12-04T06:26:31 or euclidean_distance and a radius, but that's more work (you need to call sqrt) 2011-12-04T06:34:43 *** kire has joined #aichallenge 2011-12-04T06:38:50 *** marijnfs has joined #aichallenge 2011-12-04T06:47:47 *** choas has quit IRC (Ping timeout: 240 seconds) 2011-12-04T06:54:27 *** epicmonkey has quit IRC (Ping timeout: 240 seconds) 2011-12-04T06:54:31 *** Redgis has joined #aichallenge 2011-12-04T06:58:20 *** treeform has joined #aichallenge 2011-12-04T07:03:12 *** ermacchi has joined #aichallenge 2011-12-04T07:03:48 quote :cookie 2011-12-04T07:03:54 *** dvladim has joined #aichallenge 2011-12-04T07:05:26 hi dvladim 2011-12-04T07:06:18 New Kids on the Block - Step By Step 2011-12-04T07:07:35 *** ermacchi has quit IRC (Quit: Leaving) 2011-12-04T07:11:50 *** dvladim has quit IRC (Read error: Operation timed out) 2011-12-04T07:21:07 *** Conorach has joined #aichallenge 2011-12-04T07:23:07 thestinger: thx a lot, until now i was wondering why my bot was freezing over 70 ants :P 2011-12-04T07:23:30 now i know that move and see require differant distance calculation ! 2011-12-04T07:24:06 see = euclidean, move = manhanttan 2011-12-04T07:24:28 *Manhattan* 2011-12-04T07:24:53 now i can win the contest :P 2011-12-04T07:25:21 well, manhattan distance doesn't get you a real distance to something 2011-12-04T07:25:43 just the minimum (assuming there was no water) 2011-12-04T07:28:04 thestinger: i'm using a* search, 2011-12-04T07:28:30 thestinger: can't I say it's a form of manhattan distance calculation ? 2011-12-04T07:28:37 as long as it calculates the real path 2011-12-04T07:28:49 hmmm, you could call it 'real distance' 2011-12-04T07:29:00 i'll call it that way then :P 2011-12-04T07:29:21 noctopuss: well, if you're using it in A* heuristic it's fine :) 2011-12-04T07:29:25 *** antimatroid has joined #aichallenge 2011-12-04T07:29:36 because it is the real amount of steps needed to get to the destination, as opposed to the minimum if no obstacles, or the euclidian, like the bird flies. 2011-12-04T07:29:37 thestinger: cool then :) 2011-12-04T07:29:54 but in the end it's all words until your bot actually moves that way ;) 2011-12-04T07:30:06 ^^ 2011-12-04T07:30:09 and nobody will really care what distance is used in the end ;) 2011-12-04T07:30:25 C++ is silly 2011-12-04T07:30:29 people who will open the source of the top 5 player 2011-12-04T07:30:30 will 2011-12-04T07:30:43 I put all the structs/classes in anonymous namespaces and gcc turns it into a binary with 50% smaller size now xD 2011-12-04T07:31:03 >< 2011-12-04T07:31:20 thestinger: a quite strange optimization 2011-12-04T07:31:53 well, before I guess the compiler assumed that inlining stuff would lead to a larger size, because other translation units might have used them 2011-12-04T07:37:27 *** g0llum has joined #aichallenge 2011-12-04T07:45:03 *** olexs has joined #aichallenge 2011-12-04T07:58:01 *** delt0r_ has quit IRC (Ping timeout: 244 seconds) 2011-12-04T08:06:43 *** epicmonkey has joined #aichallenge 2011-12-04T08:11:58 *** delt0r_ has joined #aichallenge 2011-12-04T08:15:38 *** jstemmer has joined #aichallenge 2011-12-04T08:28:46 *** mj41 has quit IRC (Ping timeout: 244 seconds) 2011-12-04T08:32:29 *** AntDroid has quit IRC (Ping timeout: 265 seconds) 2011-12-04T08:35:16 *** AntDroid has joined #aichallenge 2011-12-04T08:35:26 *** grwip has quit IRC (Quit: Leaving) 2011-12-04T08:36:32 *** xathis has joined #aichallenge 2011-12-04T08:46:55 *** antimatroid has quit IRC (*.net *.split) 2011-12-04T08:46:55 *** Conorach has quit IRC (*.net *.split) 2011-12-04T08:46:55 *** marijnfs has quit IRC (*.net *.split) 2011-12-04T08:46:55 *** ejls_ has quit IRC (*.net *.split) 2011-12-04T08:46:55 *** spacebat_ has quit IRC (*.net *.split) 2011-12-04T08:46:55 *** xathis has quit IRC (*.net *.split) 2011-12-04T08:46:55 *** g0llum has quit IRC (*.net *.split) 2011-12-04T08:46:55 *** kire has quit IRC (*.net *.split) 2011-12-04T08:46:55 *** Palmik has quit IRC (*.net *.split) 2011-12-04T08:46:55 *** artart78 has quit IRC (*.net *.split) 2011-12-04T08:46:55 *** delt0r_ has quit IRC (*.net *.split) 2011-12-04T08:46:55 *** foRei has quit IRC (*.net *.split) 2011-12-04T08:46:55 *** analyst74 has quit IRC (*.net *.split) 2011-12-04T08:46:55 *** pairofdice has quit IRC (*.net *.split) 2011-12-04T08:46:55 *** olexs has quit IRC (*.net *.split) 2011-12-04T08:46:55 *** sorki has quit IRC (*.net *.split) 2011-12-04T08:46:55 *** nplus has quit IRC (*.net *.split) 2011-12-04T08:46:55 *** heinrich5991 has quit IRC (*.net *.split) 2011-12-04T08:46:55 *** GeorgeJ has quit IRC (*.net *.split) 2011-12-04T08:46:55 *** jnwhiteh has quit IRC (*.net *.split) 2011-12-04T08:46:55 *** Minthos has quit IRC (*.net *.split) 2011-12-04T08:46:55 *** MikeyG has quit IRC (*.net *.split) 2011-12-04T08:46:55 *** AlliedEnvy has quit IRC (*.net *.split) 2011-12-04T08:46:56 *** jstemmer has quit IRC (*.net *.split) 2011-12-04T08:46:56 *** Accoun has quit IRC (*.net *.split) 2011-12-04T08:46:56 *** Lithosphere has quit IRC (*.net *.split) 2011-12-04T08:46:56 *** juuso_ has quit IRC (*.net *.split) 2011-12-04T08:46:56 *** klutometis has quit IRC (*.net *.split) 2011-12-04T08:46:56 *** Redgis has quit IRC (*.net *.split) 2011-12-04T08:46:56 *** Sir_Ragnarok has quit IRC (*.net *.split) 2011-12-04T08:46:56 *** bhasker has quit IRC (*.net *.split) 2011-12-04T08:46:56 *** dai-ra has quit IRC (*.net *.split) 2011-12-04T08:46:56 *** j3camero has quit IRC (*.net *.split) 2011-12-04T08:46:56 *** rb_ has quit IRC (*.net *.split) 2011-12-04T08:46:56 *** durarara has quit IRC (*.net *.split) 2011-12-04T08:46:56 *** Ttech has quit IRC (*.net *.split) 2011-12-04T08:46:56 *** Clex has quit IRC (*.net *.split) 2011-12-04T08:46:56 *** f4hy has quit IRC (*.net *.split) 2011-12-04T08:46:56 *** acieroid has quit IRC (*.net *.split) 2011-12-04T08:46:56 *** pgpaskar_ has quit IRC (*.net *.split) 2011-12-04T08:46:56 *** cbad has quit IRC (*.net *.split) 2011-12-04T08:46:56 *** bugnuts has quit IRC (*.net *.split) 2011-12-04T08:46:56 *** tcoppi has quit IRC (*.net *.split) 2011-12-04T08:46:57 *** antonh has quit IRC (*.net *.split) 2011-12-04T08:46:57 *** seletskiy has quit IRC (*.net *.split) 2011-12-04T08:46:57 *** hkraal has quit IRC (*.net *.split) 2011-12-04T08:46:57 *** Fluxid has quit IRC (*.net *.split) 2011-12-04T08:46:57 *** Kaji has quit IRC (*.net *.split) 2011-12-04T08:46:57 *** Thinq has quit IRC (*.net *.split) 2011-12-04T08:46:57 *** magikmw has quit IRC (*.net *.split) 2011-12-04T08:46:57 *** andar has quit IRC (*.net *.split) 2011-12-04T08:46:57 *** deafferret has quit IRC (*.net *.split) 2011-12-04T08:46:57 *** mawif has quit IRC (*.net *.split) 2011-12-04T08:46:57 *** isbric has quit IRC (*.net *.split) 2011-12-04T08:46:57 *** a1k0n has quit IRC (*.net *.split) 2011-12-04T08:46:57 *** Insti has quit IRC (*.net *.split) 2011-12-04T08:46:57 *** timos has quit IRC (*.net *.split) 2011-12-04T08:46:57 *** delt0r has quit IRC (*.net *.split) 2011-12-04T08:46:57 *** sigh has quit IRC (*.net *.split) 2011-12-04T08:46:57 *** oele2 has quit IRC (*.net *.split) 2011-12-04T08:46:57 *** raemde has quit IRC (*.net *.split) 2011-12-04T08:46:57 *** modafinil has quit IRC (*.net *.split) 2011-12-04T08:46:57 *** Zepp has quit IRC (*.net *.split) 2011-12-04T08:46:57 *** Raimondi has quit IRC (*.net *.split) 2011-12-04T08:46:57 *** cyphase has quit IRC (*.net *.split) 2011-12-04T08:46:57 *** McLeopold has quit IRC (*.net *.split) 2011-12-04T08:46:57 *** Israfel has quit IRC (*.net *.split) 2011-12-04T08:46:58 *** AndrewBC has quit IRC (*.net *.split) 2011-12-04T08:46:58 *** zostay has quit IRC (*.net *.split) 2011-12-04T08:46:58 *** ikaros has quit IRC (*.net *.split) 2011-12-04T08:46:58 *** mleise has quit IRC (*.net *.split) 2011-12-04T08:46:58 *** Extrarius has quit IRC (*.net *.split) 2011-12-04T08:46:58 *** rofer has quit IRC (*.net *.split) 2011-12-04T08:46:58 *** rmmh has quit IRC (*.net *.split) 2011-12-04T08:46:58 *** emilk has quit IRC (*.net *.split) 2011-12-04T08:46:58 *** jix has quit IRC (*.net *.split) 2011-12-04T08:46:58 *** Seeker` has quit IRC (*.net *.split) 2011-12-04T08:46:58 *** rwest has quit IRC (*.net *.split) 2011-12-04T08:46:58 *** bysin has quit IRC (*.net *.split) 2011-12-04T08:46:58 *** Wraithan has quit IRC (*.net *.split) 2011-12-04T08:46:58 *** mansfeld has quit IRC (*.net *.split) 2011-12-04T08:46:58 *** ruser has quit IRC (*.net *.split) 2011-12-04T08:46:58 *** suid has quit IRC (*.net *.split) 2011-12-04T08:46:58 *** ivan`_ has quit IRC (*.net *.split) 2011-12-04T08:46:58 *** rintaun has quit IRC (*.net *.split) 2011-12-04T08:46:58 *** sh4wn has quit IRC (*.net *.split) 2011-12-04T08:46:58 *** epicmonkey has quit IRC (*.net *.split) 2011-12-04T08:46:58 *** murr4y has quit IRC (*.net *.split) 2011-12-04T08:46:58 *** Zannick has quit IRC (*.net *.split) 2011-12-04T08:46:58 *** Kommander has quit IRC (*.net *.split) 2011-12-04T08:46:58 *** magiik has quit IRC (*.net *.split) 2011-12-04T08:46:58 *** Apophis has quit IRC (*.net *.split) 2011-12-04T08:46:58 *** rexou has quit IRC (*.net *.split) 2011-12-04T08:46:58 *** gazpachoking has quit IRC (*.net *.split) 2011-12-04T08:46:58 *** Bluedgis has quit IRC (*.net *.split) 2011-12-04T08:46:58 *** staalebk has quit IRC (*.net *.split) 2011-12-04T08:46:58 *** CIA-53 has quit IRC (*.net *.split) 2011-12-04T08:46:58 *** smly- has quit IRC (*.net *.split) 2011-12-04T08:46:58 *** treeform has quit IRC (*.net *.split) 2011-12-04T08:46:58 *** jmcarthur has quit IRC (*.net *.split) 2011-12-04T08:46:58 *** dom7b5 has quit IRC (*.net *.split) 2011-12-04T08:46:58 *** dr- has quit IRC (*.net *.split) 2011-12-04T08:46:59 *** Chris_0076 has quit IRC (*.net *.split) 2011-12-04T08:46:59 *** bugnuts2 has quit IRC (*.net *.split) 2011-12-04T08:46:59 *** welterde has quit IRC (*.net *.split) 2011-12-04T08:46:59 *** Alexer has quit IRC (*.net *.split) 2011-12-04T08:46:59 *** Katiska has quit IRC (*.net *.split) 2011-12-04T08:46:59 *** aperiodic has quit IRC (*.net *.split) 2011-12-04T08:46:59 *** lahwran has quit IRC (*.net *.split) 2011-12-04T08:46:59 *** AntDroid has quit IRC (*.net *.split) 2011-12-04T08:46:59 *** noctopuss has quit IRC (*.net *.split) 2011-12-04T08:46:59 *** bretep has quit IRC (*.net *.split) 2011-12-04T08:46:59 *** Euphony has quit IRC (*.net *.split) 2011-12-04T08:46:59 *** keith_ has quit IRC (*.net *.split) 2011-12-04T08:46:59 *** ssclift has quit IRC (*.net *.split) 2011-12-04T08:46:59 *** cwc has quit IRC (*.net *.split) 2011-12-04T08:46:59 *** amstan__ has quit IRC (*.net *.split) 2011-12-04T08:46:59 *** kurti has quit IRC (*.net *.split) 2011-12-04T08:46:59 *** bqf has quit IRC (*.net *.split) 2011-12-04T08:46:59 *** JorgeB has quit IRC (*.net *.split) 2011-12-04T08:46:59 *** BenJackson has quit IRC (*.net *.split) 2011-12-04T08:46:59 *** luizribeiro has quit IRC (*.net *.split) 2011-12-04T08:46:59 *** Elderwolf has quit IRC (*.net *.split) 2011-12-04T08:46:59 *** tobiassjosten has quit IRC (*.net *.split) 2011-12-04T08:46:59 *** ChanServ has quit IRC (*.net *.split) 2011-12-04T08:47:03 *** Fluxid has joined #aichallenge 2011-12-04T08:47:03 *** mawif has joined #aichallenge 2011-12-04T08:47:07 *** isbric has joined #aichallenge 2011-12-04T08:47:08 *** delt0r has joined #aichallenge 2011-12-04T08:47:09 *** seletskiy has joined #aichallenge 2011-12-04T08:47:24 *** andar has joined #aichallenge 2011-12-04T08:52:29 *** rb has joined #aichallenge 2011-12-04T08:52:33 *** rb is now known as 31NAACW1K 2011-12-04T08:52:53 *** tcoppi has joined #aichallenge 2011-12-04T08:52:53 *** Sir_Ragnarok has joined #aichallenge 2011-12-04T08:52:53 *** a1k0n has joined #aichallenge 2011-12-04T08:52:53 *** durarara has joined #aichallenge 2011-12-04T08:52:53 *** Insti has joined #aichallenge 2011-12-04T08:52:53 *** hkraal has joined #aichallenge 2011-12-04T08:52:53 *** Kaji has joined #aichallenge 2011-12-04T08:52:53 *** Redgis has joined #aichallenge 2011-12-04T08:52:53 *** pgpaskar_ has joined #aichallenge 2011-12-04T08:52:53 *** acieroid has joined #aichallenge 2011-12-04T08:52:53 *** Clex has joined #aichallenge 2011-12-04T08:52:53 *** amstan__ has joined #aichallenge 2011-12-04T08:52:53 *** deafferret has joined #aichallenge 2011-12-04T08:52:53 *** dai-ra has joined #aichallenge 2011-12-04T08:52:53 *** bqf has joined #aichallenge 2011-12-04T08:52:53 *** tobiassjosten has joined #aichallenge 2011-12-04T08:52:53 *** timos has joined #aichallenge 2011-12-04T08:52:54 *** BenJacks1n has joined #aichallenge 2011-12-04T08:52:54 *** Elderwol1 has joined #aichallenge 2011-12-04T08:52:54 *** luizribe` has joined #aichallenge 2011-12-04T08:52:54 *** xathis has joined #aichallenge 2011-12-04T08:52:54 *** jstemmer has joined #aichallenge 2011-12-04T08:52:54 *** delt0r_ has joined #aichallenge 2011-12-04T08:52:54 *** olexs has joined #aichallenge 2011-12-04T08:52:54 *** g0llum has joined #aichallenge 2011-12-04T08:52:54 *** antimatroid has joined #aichallenge 2011-12-04T08:52:54 *** Conorach has joined #aichallenge 2011-12-04T08:52:54 *** marijnfs has joined #aichallenge 2011-12-04T08:52:54 *** kire has joined #aichallenge 2011-12-04T08:52:54 *** foRei has joined #aichallenge 2011-12-04T08:52:54 *** Palmik has joined #aichallenge 2011-12-04T08:52:54 *** analyst74 has joined #aichallenge 2011-12-04T08:52:54 *** pairofdice has joined #aichallenge 2011-12-04T08:52:54 *** ejls_ has joined #aichallenge 2011-12-04T08:52:54 *** artart78 has joined #aichallenge 2011-12-04T08:52:54 *** spacebat_ has joined #aichallenge 2011-12-04T08:52:54 *** Accoun has joined #aichallenge 2011-12-04T08:52:54 *** Lithosphere has joined #aichallenge 2011-12-04T08:52:54 *** sorki has joined #aichallenge 2011-12-04T08:52:54 *** nplus has joined #aichallenge 2011-12-04T08:52:54 *** heinrich5991 has joined #aichallenge 2011-12-04T08:52:54 *** GeorgeJ has joined #aichallenge 2011-12-04T08:52:54 *** jnwhiteh has joined #aichallenge 2011-12-04T08:52:54 *** juuso_ has joined #aichallenge 2011-12-04T08:52:54 *** MikeyG has joined #aichallenge 2011-12-04T08:52:54 *** Minthos has joined #aichallenge 2011-12-04T08:52:54 *** klutometis has joined #aichallenge 2011-12-04T08:52:54 *** AlliedEnvy has joined #aichallenge 2011-12-04T08:53:04 *** f4hy has joined #aichallenge 2011-12-04T08:53:04 *** sigh has joined #aichallenge 2011-12-04T08:53:05 *** oele2 has joined #aichallenge 2011-12-04T08:53:05 *** raemde has joined #aichallenge 2011-12-04T08:53:05 *** modafinil has joined #aichallenge 2011-12-04T08:53:05 *** Zepp has joined #aichallenge 2011-12-04T08:53:05 *** Raimondi has joined #aichallenge 2011-12-04T08:53:14 *** AntDroid has joined #aichallenge 2011-12-04T08:53:14 *** noctopuss has joined #aichallenge 2011-12-04T08:53:14 *** bretep has joined #aichallenge 2011-12-04T08:53:14 *** Euphony has joined #aichallenge 2011-12-04T08:53:14 *** keith_ has joined #aichallenge 2011-12-04T08:53:14 *** ssclift has joined #aichallenge 2011-12-04T08:53:14 *** cwc has joined #aichallenge 2011-12-04T08:53:17 *** Thinq has joined #aichallenge 2011-12-04T08:53:17 *** bugnuts has joined #aichallenge 2011-12-04T08:53:17 *** cbad has joined #aichallenge 2011-12-04T08:53:17 *** antonh has joined #aichallenge 2011-12-04T08:53:17 *** kurti has joined #aichallenge 2011-12-04T08:53:17 *** cyphase has joined #aichallenge 2011-12-04T08:53:17 *** McLeopold has joined #aichallenge 2011-12-04T08:53:17 *** Israfel has joined #aichallenge 2011-12-04T08:53:17 *** AndrewBC has joined #aichallenge 2011-12-04T08:53:17 *** zostay has joined #aichallenge 2011-12-04T08:53:20 *** AntDroid has quit IRC (Changing host) 2011-12-04T08:53:20 *** AntDroid has joined #aichallenge 2011-12-04T08:53:45 *** magikmw has joined #aichallenge 2011-12-04T08:53:45 *** j3camero has joined #aichallenge 2011-12-04T08:53:45 *** epicmonkey has joined #aichallenge 2011-12-04T08:53:45 *** treeform has joined #aichallenge 2011-12-04T08:53:45 *** ikaros has joined #aichallenge 2011-12-04T08:53:45 *** Wraithan has joined #aichallenge 2011-12-04T08:53:45 *** mleise has joined #aichallenge 2011-12-04T08:53:45 *** jmcarthur has joined #aichallenge 2011-12-04T08:53:45 *** dom7b5 has joined #aichallenge 2011-12-04T08:53:45 *** dr- has joined #aichallenge 2011-12-04T08:53:45 *** rmmh has joined #aichallenge 2011-12-04T08:53:45 *** Extrarius has joined #aichallenge 2011-12-04T08:53:45 *** rofer has joined #aichallenge 2011-12-04T08:53:45 *** murr4y has joined #aichallenge 2011-12-04T08:53:45 *** Chris_0076 has joined #aichallenge 2011-12-04T08:53:45 *** emilk has joined #aichallenge 2011-12-04T08:53:45 *** jix has joined #aichallenge 2011-12-04T08:53:45 *** Zannick has joined #aichallenge 2011-12-04T08:53:45 *** Seeker` has joined #aichallenge 2011-12-04T08:53:45 *** Kommander has joined #aichallenge 2011-12-04T08:53:45 *** rwest has joined #aichallenge 2011-12-04T08:53:45 *** bugnuts2 has joined #aichallenge 2011-12-04T08:53:45 *** magiik has joined #aichallenge 2011-12-04T08:53:45 *** Apophis has joined #aichallenge 2011-12-04T08:53:45 *** bysin has joined #aichallenge 2011-12-04T08:53:45 *** rexou has joined #aichallenge 2011-12-04T08:53:45 *** welterde has joined #aichallenge 2011-12-04T08:53:45 *** gazpachoking has joined #aichallenge 2011-12-04T08:53:45 *** mansfeld has joined #aichallenge 2011-12-04T08:53:45 *** Alexer has joined #aichallenge 2011-12-04T08:53:45 *** Bluedgis has joined #aichallenge 2011-12-04T08:53:45 *** ruser has joined #aichallenge 2011-12-04T08:53:45 *** suid has joined #aichallenge 2011-12-04T08:53:45 *** ivan`_ has joined #aichallenge 2011-12-04T08:53:45 *** rintaun has joined #aichallenge 2011-12-04T08:53:45 *** sh4wn has joined #aichallenge 2011-12-04T08:53:45 *** staalebk has joined #aichallenge 2011-12-04T08:53:45 *** CIA-53 has joined #aichallenge 2011-12-04T08:53:45 *** Katiska has joined #aichallenge 2011-12-04T08:53:45 *** lahwran has joined #aichallenge 2011-12-04T08:53:45 *** aperiodic has joined #aichallenge 2011-12-04T08:53:45 *** smly- has joined #aichallenge 2011-12-04T08:59:45 *** Ttech has joined #aichallenge 2011-12-04T09:00:10 *** sigh has quit IRC (Remote host closed the connection) 2011-12-04T09:02:02 *** ChanServ has joined #aichallenge 2011-12-04T09:02:03 *** wolfe.freenode.net sets mode: +o ChanServ 2011-12-04T09:02:36 *** Fluxid is now known as Guest73419 2011-12-04T09:02:36 *** Ttech is now known as Guest64586 2011-12-04T09:13:57 do people have a max search path? 2011-12-04T09:14:01 and if so what do you use? :P 2011-12-04T09:16:28 *** keith_ has quit IRC (Ping timeout: 265 seconds) 2011-12-04T09:21:25 *** ejls_ is now known as ejls 2011-12-04T09:23:58 9 out of 10 games timed out. wtf! this internet connection sucks. how can i properly test my bot now. 2011-12-04T09:27:31 So, I finished last and my mu increased. How does that work? 2011-12-04T09:29:17 *** treeform has quit IRC (Remote host closed the connection) 2011-12-04T09:29:39 bretep: what's your bot's name? 2011-12-04T09:30:54 peterb 2011-12-04T09:31:10 Might be because I played against the 2th and another decent bot 2011-12-04T09:32:23 *** UncleVasya has joined #aichallenge 2011-12-04T09:32:38 bretep: well, you tied for 3rd with 4 people 2011-12-04T09:32:49 and one was higher ranked than you 2011-12-04T09:33:28 well, I guess only mu matters for that 2011-12-04T09:33:38 So, if you royally lose together with a few others you can rise in rank. :p 2011-12-04T09:33:48 the 3 people you tied with have slightly higher mu 2011-12-04T09:33:52 Misery loves company 2011-12-04T09:33:54 Or something 2011-12-04T09:33:56 so you did better than expected :) 2011-12-04T09:34:13 pguillory 0.18 away from xathis 2011-12-04T09:34:42 uh oh :D 2011-12-04T09:35:03 err... 0.16 2011-12-04T09:36:49 xathis: Did you compete in Planet Wars? 2011-12-04T09:36:57 *** TheLinker has joined #aichallenge 2011-12-04T09:37:22 UncleVasya: no i did not know of the contest back then 2011-12-04T09:38:05 good :) 2011-12-04T09:44:33 * antimatroid is struggling to get decent food collection going on a 200x200 map 2011-12-04T09:44:47 *** mj41 has joined #aichallenge 2011-12-04T09:44:56 xathis: does your map go alright on 200x200? 2011-12-04T09:45:00 bot* 2011-12-04T09:45:25 olexs, I've downloaded tools.zip but AntsTesting 0.5 still works on my machine. 2011-12-04T09:45:41 antimatroid: i never tried, but i don't think we'll get such big maps for 2 players 2011-12-04T09:46:07 i'm currently testing mine on a 3 player map that's close to that 2011-12-04T09:47:55 antimatroid: Is there any information about what will be after that contest? Another one, Zeta or something else? 2011-12-04T09:48:05 *this contest 2011-12-04T09:48:09 no such information exists 2011-12-04T09:48:17 as far as i know 2011-12-04T09:48:17 thanks 2011-12-04T09:48:28 i think most people would be interested on getting zeta going 2011-12-04T09:49:00 what i really want is official tcp servers for old contest games 2011-12-04T09:49:04 and possibly a few others 2011-12-04T09:50:18 probably keep the one user account policy but allow people to reset their ranking for any game? 2011-12-04T09:50:19 *** kincody has joined #aichallenge 2011-12-04T09:50:29 i think that might work best 2011-12-04T09:51:16 ie. doesn't really matter if someone doesn't update for a worse bot and it's in your own interest to update it if you've improved your bot 2011-12-04T09:55:31 *** marijnfs has left #aichallenge 2011-12-04T09:56:01 The other side is that some people who like to share their code after the end of the contests will not share it anymore. 2011-12-04T09:57:17 *** NotABug has joined #aichallenge 2011-12-04T09:59:20 And some bots sourcers for tron and pw already shared so there will be a mess in ranks among newbies: someone developing from scratch, someone patching open sourced bots. 2011-12-04T09:59:54 *** mikewintermute has joined #aichallenge 2011-12-04T10:00:27 but again, from the other side, it might increase the value of currently going 'main' contest: people can be sure that everyone starts from zero. 2011-12-04T10:02:04 yeah i think it would be fair game to share code at that point 2011-12-04T10:02:09 some people might become stiffer but whatever 2011-12-04T10:02:21 it would be great for beginners and people who want to just play around/test ideas 2011-12-04T10:02:46 *** grwip has joined #aichallenge 2011-12-04T10:03:11 *** mikewintermute has quit IRC (Client Quit) 2011-12-04T10:08:40 *** noctopuss has quit IRC (Ping timeout: 265 seconds) 2011-12-04T10:17:42 Someone made a good but public? 2011-12-04T10:19:28 *** zyberkiddy has joined #aichallenge 2011-12-04T10:20:23 bretep: Is lpace 18 good enough? 2011-12-04T10:21:32 *** ahoj has joined #aichallenge 2011-12-04T10:21:45 antimatroid: there are to "player %num%" strings in the cell maps. What's the difference? 2011-12-04T10:22:10 I don't know if I like it that everyone can upload a bot that has the skill of the 18th. 2011-12-04T10:24:09 I thought by public you mean compiled binary-file to test against your bot locally. 2011-12-04T10:24:58 I can give a link to .exe but I know that sharing code is against rules. 2011-12-04T10:27:33 someone released their code? 2011-12-04T10:27:43 bretep: To make it clear, our conversation with _antimatroid was about previous contests. 2011-12-04T10:27:59 *** zyberkiddy has left #aichallenge 2011-12-04T10:28:03 * antimatroid is confused 2011-12-04T10:28:30 antimatroid: I think bretep saw we're talking about shared code and was awared. 2011-12-04T10:28:36 *** zyberkiddy has joined #aichallenge 2011-12-04T10:29:43 binaries can be disassembled to figure out the exact strategies used are, .jar files can even be decompiled.. I wouldn't share them 2011-12-04T10:29:47 can someone provide links to profile of users(top prefered) which uses diffusion for exploration? 2011-12-04T10:30:24 *** Hexren has joined #aichallenge 2011-12-04T10:30:26 *** Hexren has left #aichallenge 2011-12-04T10:30:36 xathis I think? his ants look pretty diffuse to me 2011-12-04T10:31:14 Minthos: i'm not using diffusion 2011-12-04T10:31:14 xathis? i dont think so.. he tends to go diagonally only.. 2011-12-04T10:31:41 *** Vetox has joined #aichallenge 2011-12-04T10:31:43 zyberkiddy: that's just a side effect of how i implemented my pathfinding 2011-12-04T10:31:50 Hey guys 2011-12-04T10:31:57 hi 2011-12-04T10:32:01 I ran into a little problem 2011-12-04T10:32:04 IIRC my currently uploaded bot uses diffusion for exploration but it's not a top bot 2011-12-04T10:32:30 jix: where is it? 2011-12-04T10:32:34 jix: let me check ur plays 2011-12-04T10:32:35 When I click on play_one_game.py nothing happens 2011-12-04T10:32:35 Vetox: go ahead 2011-12-04T10:32:52 http://aichallenge.org/profile.php?user=4501 2011-12-04T10:32:59 Vertox: may be python problem 2011-12-04T10:33:06 Vetox: it's a python script, you need to run it with a python interpreter 2011-12-04T10:33:15 play_one_game.PY ?? 2011-12-04T10:33:19 I installed python from their website. 2011-12-04T10:33:39 it should be play_one_game.sh or .cmd 2011-12-04T10:34:14 its .cmd accidently typed .py 2011-12-04T10:34:28 and I have python 3.2.2 2011-12-04T10:35:16 Using windows? 2011-12-04T10:35:20 yes 2011-12-04T10:35:28 Vetox: try to add pause to the end of play_one_game.cmd 2011-12-04T10:35:41 I think you need to add python to the path then 2011-12-04T10:35:47 python 3.2.2 wont work afaik, you need 2.6.x or 2.7.x 2011-12-04T10:36:00 take cmd using run and then browse(cd) upto ur tools folder.. then rum manually typing name.. You can see error message 2011-12-04T10:36:04 it works fine with 3.2 2011-12-04T10:36:12 olexs: hi. 2011-12-04T10:36:19 UncleVasya: hey 2011-12-04T10:36:29 olexs: I've downloaded tools.zip but AntsTesting 0.5 still works on my machine. 2011-12-04T10:36:35 Open the cmd file in notepad or something like htat 2011-12-04T10:36:50 UncleVasya: okay... could you give me a link? 2011-12-04T10:36:55 Opened it and added "pause" to the end 2011-12-04T10:36:56 If there is no 'python' before playgame.py add it there 2011-12-04T10:37:17 yes, I'll upload it now. 2011-12-04T10:37:22 thanks 2011-12-04T10:38:02 The first word is python "%~dp0playgame.py" 2011-12-04T10:38:07 python "%~dp0playgame.py" 2011-12-04T10:38:14 janzert: any ideas on the proportion of larger maps we intend to use? 2011-12-04T10:38:34 Seems good, what error does it give? 2011-12-04T10:38:35 i'm unsure about whether to go for cheaper stuff that works well on larger maps 2011-12-04T10:38:42 or just let my bot not perform as well on larger maps 2011-12-04T10:39:11 @bretep Not sure. the cmd prompt opens and closes in a second 2011-12-04T10:39:12 Vetox: You have no gotten any error messages recently, so here's a random one just to let you know that we care. 2011-12-04T10:40:23 Open cmd and go to it manually, then you're certain it won't close :v 2011-12-04T10:40:36 Am uninstalling 3.2.2 and installing 2.7.2 2011-12-04T10:40:37 vetox, sounds like you need to add python to path or call python.exe with a direct path 2011-12-04T10:40:38 ok 2011-12-04T10:40:39 *** mleise has quit IRC (Quit: Leaving.) 2011-12-04T10:40:43 shift-rightclick in the folder - "Open command prompt here" 2011-12-04T10:40:46 Vetox: don't 2011-12-04T10:40:49 saves you the "going there manually" 2011-12-04T10:41:00 Open CMD and try python 2011-12-04T10:41:13 I guess it isn't in the path variable 2011-12-04T10:41:13 just make it c:\python32\python instead of python 2011-12-04T10:41:28 Vetox: hmm, have you saved file with 'pause' before launching again? 2011-12-04T10:41:35 Yes 2011-12-04T10:41:44 just type cmd in the address bar and press enter when you are in the tools folder 2011-12-04T10:41:47 It is needed specially for you to see the error message. 2011-12-04T10:41:52 *** mleise has joined #aichallenge 2011-12-04T10:42:59 *** ahoj has quit IRC (Ping timeout: 265 seconds) 2011-12-04T10:44:01 Python is not recognized as an internal or external batch command....or something similar 2011-12-04T10:44:17 Yeah, 2011-12-04T10:44:45 *** Garf has joined #aichallenge 2011-12-04T10:44:57 python is not set in the Path environment variable 2011-12-04T10:44:59 http://docs.python.org/using/windows.html got to 3.3.1. Excursus: Setting environment variables 2011-12-04T10:45:10 Vetox: It seems like you need to add python directory to your PATH in the system settings. 2011-12-04T10:45:26 How do I do that? 2011-12-04T10:45:42 See link :) 2011-12-04T10:46:06 Oops. Didnt realize the link was for me :P 2011-12-04T10:46:15 If you don't want to bother with that just call python with c:\python32\python 2011-12-04T10:46:17 olexs: http://ompldr.org/vYmwxNg/AntsTesting_0_5_a.7z 2011-12-04T10:46:26 *** yoden has joined #aichallenge 2011-12-04T10:46:41 (or wherever your python.exe is) 2011-12-04T10:46:48 UncleVasya: thank you, I'll try that 2011-12-04T10:47:42 btw, many thanks for the guy who adviced me to use that uploader :) 2011-12-04T10:48:35 Vetox: Wnen you'll solve your problem try Ants Testing too ;) 2011-12-04T10:48:42 Sure 2011-12-04T10:49:30 UncleVasya: it crashed when I clicked "start", "ar" is not a valid integer value 2011-12-04T10:49:40 yeah, omp is a lot better than the file factory hell :) 2011-12-04T10:49:48 You tried cell maps? 2011-12-04T10:49:53 even with unicorns 2011-12-04T10:50:00 I enabled all maps 2011-12-04T10:50:05 So I have to set the path where my python.exe is right? 2011-12-04T10:50:28 yep, I forgot. It exactly what i was doing before was involved into this conversation :) 2011-12-04T10:50:41 so cell maps don't work? :) 2011-12-04T10:51:01 temporary condition. 2011-12-04T10:51:16 try other types now. 2011-12-04T10:51:16 PYTHONPATH=C:\Python27 2011-12-04T10:51:40 I set this in System Environment variables but still doesnt work PYTHONPATH=C:\Python27 2011-12-04T10:51:53 Is there a chance of getting the gcc for c++11 updated to 4.6? 2011-12-04T10:52:19 There are 2 strings with players count in them and I must ask antimatroid which one to use. 2011-12-04T10:52:20 Vetox, grrrrar. 2011-12-04T10:52:26 'python' is not recognized as an internal or external command, operable program or batch file. 2011-12-04T10:52:36 Still get the same error....... 2011-12-04T10:52:45 Now in feeling kinda dumb.......... 2011-12-04T10:52:48 if it is his maps 2011-12-04T10:53:40 Add it to path, not pythonpath 2011-12-04T10:53:46 and make sure it's the right place 2011-12-04T10:54:00 Vetox, you must append that to path, not pythonpath 2011-12-04T10:54:33 olexs: Do another maps work for you? 2011-12-04T10:54:42 Path should be a long string, just add ;C:\yourpythondir 2011-12-04T10:55:02 UncleVasya: running it on maze maps only right now, looks good so far 2011-12-04T10:56:49 against mine? :) 2011-12-04T10:57:27 I did a few test plays against yours earlier, mine doesn't do too badly actually 2011-12-04T10:57:29 Oh I got it 2011-12-04T10:57:48 There is already a Path Variable....I should append c:\python27 to it 2011-12-04T10:57:52 .......... 2011-12-04T10:57:54 ... 2011-12-04T10:58:11 the version that is on aichallenge currently should at the very least make it into the top100 2011-12-04T10:58:30 it just takes ages -.- 2011-12-04T10:58:45 Now i got it working :) 2011-12-04T10:58:59 Incorrect number of bots for map. Need 4, got 5 2011-12-04T10:59:04 I get this error now :( 2011-12-04T10:59:12 Figure it out :) 2011-12-04T10:59:53 Hehe...Done and done :) 2011-12-04T11:00:00 Thanks guys :D 2011-12-04T11:00:04 You don't like the speed of the server while rising. When your bot is going down, you want it to be as slow as possible :D 2011-12-04T11:01:13 Vetox: Ants Testing makes a correct play_one_game.cmd by itself. You use only the nice gui. 2011-12-04T11:01:16 You can remove all but two bots and add --fill option 2011-12-04T11:01:31 So you don't have to constantly change the amount of bots when you change maps 2011-12-04T11:01:50 are those servers in server_stats page shows physical separate servers or no. of instances of games in servers? 2011-12-04T11:02:00 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has joined #aichallenge 2011-12-04T11:10:08 Hurrah!!!!! Finally got mt tutorial.cmd working!!!! 2011-12-04T11:11:30 grats 2011-12-04T11:13:15 *** Guest73419 is now known as Fluxid 2011-12-04T11:13:52 *** jbroman has joined #aichallenge 2011-12-04T11:16:15 olexs: fixed. Uploading. 2011-12-04T11:18:21 *** amstan__ is now known as amstan 2011-12-04T11:18:28 *** Vetox has quit IRC (Quit: Page closed) 2011-12-04T11:19:08 *** contestbot_ is now known as contestbot 2011-12-04T11:22:35 olexs: http://ompldr.org/vYmwxbw/AntsTesting_0_5_b.7z 2011-12-04T11:23:50 *** kire has quit IRC (Quit: Leaving) 2011-12-04T11:26:29 *** NotABug has quit IRC (Ping timeout: 265 seconds) 2011-12-04T11:32:07 *** amstan has quit IRC (Ping timeout: 240 seconds) 2011-12-04T11:39:49 *** TheLinker has quit IRC (Quit: Bye) 2011-12-04T11:39:59 *** zyberkiddy has left #aichallenge 2011-12-04T11:40:57 *** NotABug has joined #aichallenge 2011-12-04T11:42:37 *** keith_ has joined #aichallenge 2011-12-04T11:43:09 "This usually means they won all of those initial games (and probably got no 2-player games given that they don't change your ranking as much as winning an 8 would)" Ummm, 2 player games change your rank the fastest. No? 2011-12-04T11:43:25 *** TheLinker has joined #aichallenge 2011-12-04T11:43:52 well, winning an 8-player game against 7 strong opponents should increase your rank a lot 2011-12-04T11:43:52 http://forums.aichallenge.org/viewtopic.php?f=25&t=2036 2011-12-04T11:44:14 Yes, but if you have a strong opponent in 1v1 it increases more 2011-12-04T11:44:49 Atleast that's what I got when I read the TrueSkill docs 2011-12-04T11:52:53 finally, the new version gets a clean win over the long-undefeated "champion"... http://paste.aichallenge.org/mBNAJ/ 2011-12-04T11:56:15 *** Rob__ has joined #aichallenge 2011-12-04T11:58:28 very interesting map! 2011-12-04T11:59:11 *** Rob__ has quit IRC (Client Quit) 2011-12-04T12:07:05 *** NotABug has quit IRC (Ping timeout: 265 seconds) 2011-12-04T12:08:46 olexs: can you post the map somewhere? 2011-12-04T12:09:08 I found it had to retrieve it from the replay data. 2011-12-04T12:09:21 *retrieve map 2011-12-04T12:11:12 *** u_ has joined #aichallenge 2011-12-04T12:19:41 *** jacob_strauss has joined #aichallenge 2011-12-04T12:20:22 *** avdg has joined #aichallenge 2011-12-04T12:21:10 UncleVasya: http://paste.aichallenge.org/TcVIK/ here you go, raw from this paste is same as map file 2011-12-04T12:21:38 I got it out of a replay on ants paste as well, don't know where it originates from, but it's interesting for tests 2011-12-04T12:21:39 thank you 2011-12-04T12:24:43 *** avdg has quit IRC (Ping timeout: 252 seconds) 2011-12-04T12:26:27 Damn python 2011-12-04T12:26:42 slow? 2011-12-04T12:26:49 Very 2011-12-04T12:27:25 Now I just want to wrtie a bot in another language 2011-12-04T12:27:30 write 2011-12-04T12:31:06 pairofdice: are you aware that you can use pypy on the official servers? 2011-12-04T12:31:19 *** UncleVasya has quit IRC (Ping timeout: 252 seconds) 2011-12-04T12:33:11 No more pypy, I'm currently #1 in pypy and want to keep that spot :v 2011-12-04T12:34:29 :) 2011-12-04T12:35:44 aichallenge: McLeopold epsilon * r4a2883c / (sql/2_generate_matchup.sql sql/opponent.sql): fix max games bug - http://git.io/c15ZMw 2011-12-04T12:35:45 aichallenge: McLeopold epsilon * rc98c078 / ants/visualizer/js/Util.js : Merge branch 'epsilon' of github.com:aichallenge/aichallenge into epsilon - http://git.io/Y2oaCQ 2011-12-04T12:35:54 *** UncleVasya has joined #aichallenge 2011-12-04T12:37:43 I don't really want to figure pypy out 2011-12-04T12:38:54 aichallenge: McLeopold epsilon * r871de4d / (sql/2_generate_matchup.sql sql/opponent.sql): fix max games bug bug - http://git.io/FE7zGA 2011-12-04T12:39:56 pairofdice: "figure out"? best case is "rename .py to .pypy and reupload" 2011-12-04T12:40:57 if you're using python3 that won't work, of course. downloading and installing pypy should be a breeze as well though, and after that you just run "pypy MyBot.py" instead of "python MyBot.py" 2011-12-04T12:41:33 *** Surya has joined #aichallenge 2011-12-04T12:42:07 pypy supports most of python3 2011-12-04T12:42:29 it does? wow, i guess i'm behind on pypy development. time to go check it out again :) 2011-12-04T12:42:34 Ahh 2011-12-04T12:42:40 *** TheLinker has quit IRC (Quit: Bye) 2011-12-04T12:42:57 Okay, it runs 'considerably' faster with pypy 2011-12-04T12:44:05 *** TheLinker has joined #aichallenge 2011-12-04T12:44:55 That was less painfull than expected 2011-12-04T12:44:59 thanks 2011-12-04T12:46:44 antimatroid: I've read some part of your conversation about problem with timeout. Have you solved it? 2011-12-04T12:47:53 I'm interested because it seems I have near the same problem. 2011-12-04T12:49:55 pairofdice: you're welcome. and yes, pypy is nice for some light number crunching, which ants is (unless you count battle) ;) 2011-12-04T12:58:15 *** praveen has joined #aichallenge 2011-12-04T13:03:30 *** avdg has joined #aichallenge 2011-12-04T13:03:38 *** kincody has quit IRC (Ping timeout: 265 seconds) 2011-12-04T13:15:02 *** olexs1 has joined #aichallenge 2011-12-04T13:16:35 *** cowbandit has joined #aichallenge 2011-12-04T13:17:04 So Gnome 3 has been unmasked on Gentoo and I am shocked. 2011-12-04T13:17:23 *** olexs has quit IRC (Ping timeout: 252 seconds) 2011-12-04T13:17:59 No shutdown / hibernate option? I have to an additional layer in the audio pipeline (pulseaudio)? I cannot use Compiz? 2011-12-04T13:19:53 But hey, I can set my online status to either busy or available with two clicks 2011-12-04T13:23:57 *** keith_ has quit IRC (Ping timeout: 265 seconds) 2011-12-04T13:27:26 *** UncleVasya has quit IRC (Read error: Connection reset by peer) 2011-12-04T13:30:21 *** praveen has quit IRC (Quit: Page closed) 2011-12-04T13:37:50 *** pairofdice has quit IRC (Quit: leaving) 2011-12-04T13:44:59 *** amstan has joined #aichallenge 2011-12-04T13:44:59 *** ChanServ sets mode: +o amstan 2011-12-04T13:48:15 *** dvladim has joined #aichallenge 2011-12-04T13:53:27 *** mleise has quit IRC (Quit: Leaving.) 2011-12-04T13:53:30 *** Sion has joined #aichallenge 2011-12-04T13:53:49 Hi 2011-12-04T13:54:30 Is there any admins from aichallenge.org here? 2011-12-04T13:55:06 hi 2011-12-04T13:55:20 Sion: hi 2011-12-04T13:55:55 @amstain hi 2011-12-04T13:55:56 Sion: I have no idea what you mean. 2011-12-04T13:56:28 *** ChanServ sets mode: -o amstan 2011-12-04T13:57:07 I mean, Is there any admins from aichallenge.org here? so they can help me with my problem. 2011-12-04T13:57:16 Sion: yes, we're admins 2011-12-04T13:57:20 Sion: what's the problem? 2011-12-04T13:58:48 ok good, I have tryed to sign up, and it seams like it is working, but i cant log in. 2011-12-04T13:59:12 did you recieve a confirmation email? 2011-12-04T14:00:06 yes, and i followes the link and ome to a site that sayed that i can log in now 2011-12-04T14:00:38 *** dvladim has quit IRC (Ping timeout: 240 seconds) 2011-12-04T14:01:07 Sion: so have you tried logging in? 2011-12-04T14:01:18 yes 2011-12-04T14:01:23 and? 2011-12-04T14:01:31 not working 2011-12-04T14:01:35 how so? 2011-12-04T14:02:03 it says login failed 2011-12-04T14:02:36 *** cwc has quit IRC (Ping timeout: 265 seconds) 2011-12-04T14:03:04 Sion: maybe you forgot your password, have you tried that password recovery? 2011-12-04T14:03:49 yes and i got this message: The user name supplied was not found in our records. 2011-12-04T14:04:19 I have also tried to make another acount, but that didnt work either 2011-12-04T14:04:50 *** cyphase has quit IRC (Read error: Connection reset by peer) 2011-12-04T14:05:02 Sion: check the other tab 2011-12-04T14:06:32 Sion: ping me if you still have issues 2011-12-04T14:14:48 *** mleise has joined #aichallenge 2011-12-04T14:15:25 ok, i think im gona be fine now 2011-12-04T14:15:33 bye 2011-12-04T14:16:06 *** Sion has left #aichallenge 2011-12-04T14:20:58 McLeopold: on the map spec - it does not say it but is it possible to have a cell which is not visible from anywhere on the map (min land distance^2 > 77) 2011-12-04T14:21:23 *** cyphase has joined #aichallenge 2011-12-04T14:23:31 *** ikaros_ has joined #aichallenge 2011-12-04T14:30:36 <_flag> McLeopold: Have the map generator/generators that are going to be used for the final contest been decided? 2011-12-04T14:30:56 *** dom7b5 has quit IRC (Ping timeout: 248 seconds) 2011-12-04T14:34:46 *** dvladim has joined #aichallenge 2011-12-04T14:35:23 *** TheLinker has quit IRC (Remote host closed the connection) 2011-12-04T14:39:06 *** mikewintermute has joined #aichallenge 2011-12-04T14:39:21 *** mikewintermute has quit IRC (Client Quit) 2011-12-04T14:43:06 *** bmh has joined #aichallenge 2011-12-04T14:49:49 *** xathis has quit IRC () 2011-12-04T14:50:27 *** Scryer has joined #aichallenge 2011-12-04T14:50:57 *** epicmonkey has quit IRC (Read error: Operation timed out) 2011-12-04T14:51:56 *** Scryer has quit IRC (Client Quit) 2011-12-04T14:57:12 *** Accoun has quit IRC () 2011-12-04T14:59:05 Are there any past maps that could be usefull for testing combat? 2011-12-04T15:00:18 bretep: http://paste.aichallenge.org/TcVIK/ this one is good for 1-on-1 testing, pretty straightforward 2011-12-04T15:00:31 http://paste.aichallenge.org/mBNAJ/ game example 2011-12-04T15:01:32 *** dvladim has quit IRC (Quit: Konversation terminated!) 2011-12-04T15:02:07 Any maps with something like stationary bots as opponents? 2011-12-04T15:02:42 ai state visualization? what's that 2011-12-04T15:03:12 rmmh: https://github.com/j-h-a/aichallenge/blob/vis_overlay/VIS_OVERLAY.md 2011-12-04T15:04:03 I kinda want to write a real-time python visualizer 2011-12-04T15:04:10 how would you download that map? 2011-12-04T15:04:18 There's a java version that is real time 2011-12-04T15:04:25 I know, I just don't like it 2011-12-04T15:04:37 cowbandit: click on "view raw" and save that as a map file 2011-12-04T15:04:47 ah kk thanks 2011-12-04T15:14:23 If I click here on raw I get a replay file; http://paste.aichallenge.org/wDdqr/ 2011-12-04T15:14:29 How can I extract the map? 2011-12-04T15:18:28 *** Accoun has joined #aichallenge 2011-12-04T15:20:07 bretep: take the first link, with the empty map 2011-12-04T15:20:32 or if you want the map from the replay, you'll need to get it out of the json data 2011-12-04T15:20:55 *** dom7b5 has joined #aichallenge 2011-12-04T15:21:38 McLeopold: ants paste needs a function to download a map file from the replay, I think :) 2011-12-04T15:21:53 Agreed! 2011-12-04T15:22:16 all the data is there, it just needs to be put together in text and prefixed with "players n, rows x, cols y" 2011-12-04T15:24:52 next thing I want to implement is recognition of idling enemies (timed out or crashed), my battle system should kill them with no losses given the information. maps like this one can come in helpful 2011-12-04T15:28:17 olexs1: if you write a replay to map converter in python, I'll hook it up 2011-12-04T15:28:56 hm... my python is quite rusty, but I'll give it a try 2011-12-04T15:29:24 bugnuts: it is acceptable to have a large water area, so that a tile may not ever be visible to a bot 2011-12-04T15:29:40 McLeopold: that just looking at latest=1 isn't/wasn't a bug 2011-12-04T15:29:42 _flag: no, we will still accept map generators 2011-12-04T15:29:43 *** kire has joined #aichallenge 2011-12-04T15:29:52 janzert: how so? 2011-12-04T15:29:52 it's suppose to look at the whole population of available opponents 2011-12-04T15:29:53 I expect that will break some peoples exploration code in an entertaining way 2011-12-04T15:30:15 janzert: that included deactivated bots, so the game_limit was too low 2011-12-04T15:31:04 actually that change shouldn't make any difference 2011-12-04T15:31:25 is the activated/deactivated thing working properly? I missed the moment when my last submission deactivated itself after 3 days, but its game rate didnt noticeably change 2011-12-04T15:31:36 it is still looking at the whole population to get the count for the percentage number 2011-12-04T15:32:18 or I guess you got that in the second commit 2011-12-04T15:32:28 and the "n players still ahead" indicator still appeared in my profile, which I assume means that the bot was still in queue despite being deactivated 2011-12-04T15:32:38 but anyway it was deliberate that it looks at the full population for the hard cutoff 2011-12-04T15:33:11 in either case I now think the hard cutoff should just be removed, but first I plan to cut it in about half 2011-12-04T15:33:31 sorry if anyone is currently on my server. I need the CPU for a while 2011-12-04T15:35:01 McLeopold: Have you done anything with using 'worms' to generate maps? I made a proof of concept (generates 1/4 a map then tiles it - would need more work to allow rotational symmetry etc): http://pastebin.com/WBDkKFFj 2011-12-04T15:36:14 janzert: when doing a quick check against xathis, about 7 bots were within 30 ranks, and 7 were at > 1000, for the first 14 2011-12-04T15:36:30 the game limit was 1 game below the max 2011-12-04T15:36:49 most of the population was at 0 or 1, deactivated bots 2011-12-04T15:37:06 so this was making the game limit far to large 2011-12-04T15:37:07 basically the soft push to even out games is working so well that the hard cutoff takes out way to many opponents 2011-12-04T15:37:17 *** treeform has joined #aichallenge 2011-12-04T15:37:25 the soft push is the modified match quality? 2011-12-04T15:37:31 right 2011-12-04T15:37:48 ok, matches looked really bad, and it was the hard cutoff 2011-12-04T15:38:00 I'm okay with it going away 2011-12-04T15:38:20 i had a shot at xathis earlier but my bot crashed in the middle :( 2011-12-04T15:38:35 Extrarius: post some sample maps on paste for me 2011-12-04T15:39:15 aichallenge: janzert epsilon * r2d123e5 / (sql/2_generate_matchup.sql sql/opponent.sql): Lower hard game limit percentage to 5% with plan to completely remove - http://git.io/zfQnrw 2011-12-04T15:39:25 olexs1: if you are deactivated, the n players ahead will still count down 2011-12-04T15:39:41 you may still get a game if you match well with someone else, just not a seed 2011-12-04T15:39:54 then, if you reactivate, you'll probably get chosen fairly quickly 2011-12-04T15:40:20 okay 2011-12-04T15:41:06 McLeopold: I have some posts on there somewhere, but I'll post some from current working version (that one is an older version) 2011-12-04T15:41:24 janzert: the more even the matches are, the less difference that percentage makes 2011-12-04T15:42:13 hmm? 2011-12-04T15:42:38 the higher the percentage the smaller the pool of opponents to make an even match 2011-12-04T15:42:54 janzert: if the 5th percentile and the 12.5th percentile have the same game count... 2011-12-04T15:43:13 if everyone has the same game count... 2011-12-04T15:43:18 right 2011-12-04T15:43:33 if only we could :) 2011-12-04T15:43:34 How easy is it to fork aichallenge's code base to come up with ones own competitions? 2011-12-04T15:43:47 Wraithan: one click on github :) 2011-12-04T15:43:58 McLeopold: helpful. 2011-12-04T15:44:14 also notice that it's the percentile number of games + 1 2011-12-04T15:44:25 You must know python, mysql, php and be fairly good and stream programming 2011-12-04T15:44:32 I meant, is it designed with reuse for other competitions in mind? 2011-12-04T15:44:41 janzert: oops, did I leave that in? 2011-12-04T15:44:50 that's purposeful too :) 2011-12-04T15:45:25 Extrarius: import win32console? wtf? 2011-12-04T15:45:28 the percentile is an upper bound on the number of players removed from use 2011-12-04T15:45:48 i.e. if everyone had the same number of games everyone would still be available for play 2011-12-04T15:45:54 otherwise no one would be 2011-12-04T15:46:46 *** ikaros_ has quit IRC (Quit: Ex-Chat) 2011-12-04T15:48:23 ok, my TCP server is back up. Now running as a user 'jail' instead of someone with actual powers. 2011-12-04T15:51:40 *** cowbandit has quit IRC (Quit: Page closed) 2011-12-04T15:52:18 http://aichallenge.org/visualizer.php?game=161159&user=4513 2011-12-04T15:52:33 haha 2011-12-04T15:55:13 McLeopold: http://paste.aichallenge.org/f2NMQ/ 2011-12-04T15:57:52 McLeopold: http://paste.aichallenge.org/r3QfG/ 2011-12-04T15:58:16 *** delt0r_ has quit IRC (Ping timeout: 240 seconds) 2011-12-04T16:04:58 *** ltriant has joined #aichallenge 2011-12-04T16:08:21 *** kire has quit IRC (Quit: Leaving) 2011-12-04T16:09:57 *** TheLinker has joined #aichallenge 2011-12-04T16:10:57 *** delt0r_ has joined #aichallenge 2011-12-04T16:12:28 *** jacob_strauss has quit IRC (Quit: Page closed) 2011-12-04T16:24:46 *** Palmik has quit IRC (Remote host closed the connection) 2011-12-04T16:25:53 *** Antimony has joined #aichallenge 2011-12-04T16:27:12 *** Redgis has quit IRC (Quit: ... mains libres) 2011-12-04T16:28:08 *** treeform has quit IRC (Remote host closed the connection) 2011-12-04T16:30:30 *** ssclift has quit IRC (Ping timeout: 265 seconds) 2011-12-04T16:31:46 *** treeform has joined #aichallenge 2011-12-04T16:35:23 *** Banoth has joined #aichallenge 2011-12-04T16:46:09 *** xar0l has joined #aichallenge 2011-12-04T16:49:34 aichallenge: McLeopold epsilon * r06711af / website/visualizer.php : show all user's stderr for admins - http://git.io/JSw9pA 2011-12-04T16:51:14 *** xar0l has quit IRC (Quit: Page closed) 2011-12-04T16:51:18 can someone tell me if they can see their own stderr for an official game? 2011-12-04T16:53:06 *** jejeje has joined #aichallenge 2011-12-04T16:53:51 *** jejeje has left #aichallenge 2011-12-04T16:54:19 *** treeform_ has joined #aichallenge 2011-12-04T16:54:43 *** jojojo has joined #aichallenge 2011-12-04T16:57:52 *** treeform has quit IRC (Ping timeout: 240 seconds) 2011-12-04T17:00:52 McLeopold; do we have any idea of what proportion of maps will be above 150x150 in the final contest? 2011-12-04T17:01:25 it's important to know whether a bot is going to need to work well on large maps in the final contest or not 2011-12-04T17:05:28 antimatroid: 0% 2011-12-04T17:05:35 ? 2011-12-04T17:05:44 i thought max map size was 200x200? 2011-12-04T17:05:45 oh, maybe higher 2011-12-04T17:05:56 antimatroid: area is limited to 25000 2011-12-04T17:06:15 which is 158.1^2 2011-12-04T17:06:32 that's including water yes? 2011-12-04T17:06:36 yes 2011-12-04T17:06:43 each dimension is limited to 200 2011-12-04T17:06:48 hmmm okay 2011-12-04T17:06:54 is there a minimum map size? 2011-12-04T17:06:58 yes 2011-12-04T17:07:04 900 area per player 2011-12-04T17:07:12 and what's the minimum starting path length between hills? :P 2011-12-04T17:07:16 20 2011-12-04T17:07:19 what about min dimension? 2011-12-04T17:07:27 no min dimension 2011-12-04T17:07:31 antimatroid: you're running out of free variables ;) 2011-12-04T17:07:31 wow, I leave for a few days, I'm 6th 2011-12-04T17:08:06 well, i guess there is a min 2011-12-04T17:08:17 900/200 2011-12-04T17:08:17 based on area, yeah 2011-12-04T17:08:24 1800/200 2011-12-04T17:08:31 2 player min :) 2011-12-04T17:08:33 derp :P 2011-12-04T17:08:41 so it has to be at least as wide as the view radius 2011-12-04T17:08:44 that's what i wanted :P 2011-12-04T17:09:16 *** Redgis has joined #aichallenge 2011-12-04T17:09:17 i don't think my "painting" vision thing works otherwise 2011-12-04T17:09:35 and i use translation values in turn 0 to set up a lot of stuff 2011-12-04T17:10:23 http://aichallenge.org/visualizer.php?game=161906&user=871 aww come on 2011-12-04T17:12:20 can this cutoff be made like, 10 turns longer? it's not the first time I see it end a game literally 2-3 turns before a hill is taken 2011-12-04T17:13:11 also, "not razing hills" is a bad cutoff reason for that game, since the only hills still there have huge hives underneath, and it's impossible to raze a single hill with a hive under it 2011-12-04T17:17:02 lol 2011-12-04T17:19:41 just take it over, there is a way to raze a hill with a big hive 2011-12-04T17:19:44 Do the ants in the hiv count for the 90% cutoff, btw? 2011-12-04T17:20:10 (well, if the enemy allows it) 2011-12-04T17:20:32 avdg: I can't possibly step on it, since every turn it spawns a new ant 2011-12-04T17:20:55 if it moves, you can take over 2011-12-04T17:21:35 it can't move, because it gets killed by my ants standing ready to take the hill 2011-12-04T17:21:37 * bretep notes if I have a load of ants in the hive and surrounded, do not let them move 2011-12-04T17:21:38 *** cyphase has quit IRC (Ping timeout: 240 seconds) 2011-12-04T17:22:23 well, they can still kill each other by taking over places 2011-12-04T17:22:26 every turn, one ant gets spawned. even if I let it live and move away, next one will spawn immediately, since the hive is still full 2011-12-04T17:22:49 not if you put an ant on top of it 2011-12-04T17:22:49 I can still kill them all and take the hive afterwards, but not if the cutoff terminates the game mid-fight 2011-12-04T17:23:10 avdg: sadly, you're wrong 2011-12-04T17:23:26 bots should get bonuses if they run extra-fast 2011-12-04T17:23:27 if you watch the game carefully, my ants do step on it after killing the one ant 2011-12-04T17:23:30 in the matchmaking 2011-12-04T17:23:34 Ant on top gets a collision with the new ant I think 2011-12-04T17:23:35 *** cyphase has joined #aichallenge 2011-12-04T17:23:47 raze comes before spawn 2011-12-04T17:24:05 then how come I didn't raze that hill? 2011-12-04T17:24:28 did you actually move an ant on top of it? 2011-12-04T17:24:36 watch the game slowly. every single turn. 2011-12-04T17:25:03 Yes he did 2011-12-04T17:25:16 and yet he still spawned on top of my ant 2011-12-04T17:26:40 *** speakeasy has joined #aichallenge 2011-12-04T17:26:53 *** Antimony has quit IRC (Ping timeout: 252 seconds) 2011-12-04T17:27:00 http://aichallenge.org/visualizer.php?game=158597&user=3473 2011-12-04T17:27:06 that's probably somewhat of a preview of the finals 2011-12-04T17:27:17 me, pguillory, A, a1k0n, xathis 2011-12-04T17:28:49 according to specs, actually, I should have razed that hill. but it did not happen 2011-12-04T17:29:20 since raze should come before spawn... except apparently it does not 2011-12-04T17:30:19 BenJacks1n: If I make my bot not melt, I'll see you in the finals ;) 2011-12-04T17:31:57 I didn't mean to presume to still be in the top 10 then 2011-12-04T17:31:59 I meant more style of game 2011-12-04T17:32:39 I should probably start writing hill finding and combat 2011-12-04T17:32:48 raze is after attack and collision 2011-12-04T17:33:54 *** Antimony has joined #aichallenge 2011-12-04T17:34:06 McLeopold: there is no collision (exactly one of my ants is moving in per turn), and no attack (no enemy ants except the ones that get spawned right on top of me) 2011-12-04T17:34:23 *** goffrie has joined #aichallenge 2011-12-04T17:35:25 technically, the hill should be razed at turn 819 2011-12-04T17:36:05 hive counts are considered for cutoffs 2011-12-04T17:36:20 not sure they should be 2011-12-04T17:36:26 imperfect play can lose even with a hive 2011-12-04T17:36:27 olexs1: if you move an ant onto another ant, that is a colission and they both die 2011-12-04T17:36:45 McLeopold: but I don't move on top of other ants... or do I?.. 2011-12-04T17:36:52 *** g0llum has quit IRC (Read error: Connection reset by peer) 2011-12-04T17:37:30 the collision happens before the attack 2011-12-04T17:37:35 damn it... could it really be that I move onto the enemy ant instead of killing him first 2011-12-04T17:37:53 crap :) thanks for clearing that up 2011-12-04T17:37:59 olexs1: it's a known issue, you can't raze a hill if the hive is not empty 2011-12-04T17:38:07 you can 2011-12-04T17:38:10 as long as the hill owner doesn't move off 2011-12-04T17:38:12 if they move off it for you 2011-12-04T17:38:18 many people will 2011-12-04T17:38:21 my bot probably would 2011-12-04T17:39:04 McLeopold: were you doing the top-20-by-mu lists? 2011-12-04T17:40:09 no, but I can 2011-12-04T17:40:21 right... so basically I need to keep one space near a hill free for the enemy to step aside and die there, and at the same time move onto the hill 2011-12-04T17:40:51 olexs1: maybe, if the bot cooperates, but I think it was smart enough not to move 2011-12-04T17:41:28 well I need to at least give him the choice :) 2011-12-04T17:41:56 yeah, I'll consider something to prevent the cutoff, if it makes sense and the other agree 2011-12-04T17:44:25 the only other thing you could do is not collect food so that maybe the cutoff won't trigger 2011-12-04T17:44:35 or start suiciding yourself to turn it off 2011-12-04T17:44:47 *** speakeasy has quit IRC (Ping timeout: 240 seconds) 2011-12-04T17:49:55 not if they aren't moving 2011-12-04T17:50:02 whoops, was scrolled up :P 2011-12-04T17:52:40 *** Banoth has quit IRC (Ping timeout: 265 seconds) 2011-12-04T17:52:47 *** mj41 has quit IRC (Ping timeout: 240 seconds) 2011-12-04T17:57:23 *** Jak_o_Shadows has joined #aichallenge 2011-12-04T17:58:58 *** bretep has quit IRC (Ping timeout: 265 seconds) 2011-12-04T18:03:33 *** treeform_ has quit IRC (Remote host closed the connection) 2011-12-04T18:17:59 *** jstrong has joined #aichallenge 2011-12-04T18:27:17 *** Antimony has quit IRC (Ping timeout: 248 seconds) 2011-12-04T18:30:52 *** foRei has quit IRC (Quit: Bye) 2011-12-04T18:32:33 *** treeform has joined #aichallenge 2011-12-04T18:38:11 *** Garf has quit IRC (Quit: Make a new plan, Stan!) 2011-12-04T18:40:59 aichallenge: McLeopold epsilon * rf143f24 / (19 files): removed bad cell maze maps - http://git.io/ww3QQg 2011-12-04T18:41:01 aichallenge: McLeopold epsilon * rcba97dd / (14 files in 3 dirs): remove map other maps - http://git.io/YY3xdA 2011-12-04T18:50:15 *** cyphase has quit IRC (Quit: http://www.cyphase.com/) 2011-12-04T18:50:54 McLeopold: cool query for http://forums.aichallenge.org/viewtopic.php?f=25&t=2036&sid=0536aea6a3fc613f267b4a0ebf2024c0 2011-12-04T18:56:33 *** treeform has quit IRC (Remote host closed the connection) 2011-12-04T18:57:00 *** treeform has joined #aichallenge 2011-12-04T19:02:36 *** Jak_o_Shadows has quit IRC (Read error: Connection reset by peer) 2011-12-04T19:04:05 *** Conorach has quit IRC (Ping timeout: 252 seconds) 2011-12-04T19:07:15 *** Antimony has joined #aichallenge 2011-12-04T19:15:51 *** ikaros has quit IRC (Quit: Ex-Chat) 2011-12-04T19:25:23 *** mleise has quit IRC (Read error: Operation timed out) 2011-12-04T19:29:32 *** Surya has quit IRC (Quit: Page closed) 2011-12-04T19:31:51 *** cyphase has joined #aichallenge 2011-12-04T19:39:05 *** Redgis has quit IRC (Quit: ... mains libres) 2011-12-04T19:40:00 *** Anilm3 has joined #aichallenge 2011-12-04T19:43:51 *** bmh has quit IRC (Ping timeout: 265 seconds) 2011-12-04T19:45:15 *** Anilm3 has quit IRC (Quit: Lost terminal) 2011-12-04T19:53:26 *** jab_bott has joined #aichallenge 2011-12-04T20:01:53 are the workers 64-bit? 2011-12-04T20:04:34 rmmh: yes 2011-12-04T20:17:56 *** grwip has quit IRC (Remote host closed the connection) 2011-12-04T20:20:34 *** bhasker has joined #aichallenge 2011-12-04T20:22:05 ohhh 2011-12-04T20:22:09 dmj111 is 26th 2011-12-04T20:22:35 i think they discovered the contest late, they are someone to look out for 2011-12-04T20:22:47 hehe, they had to move from python to c++ 2011-12-04T20:24:01 *** AntDroid_ has joined #aichallenge 2011-12-04T20:25:56 evening 2011-12-04T20:31:23 *** bmh has joined #aichallenge 2011-12-04T20:31:43 contestbot: seen janzert 2011-12-04T20:31:43 amstan: janzert was last seen in #aichallenge 4 hours, 45 minutes, and 49 seconds ago: otherwise no one would be 2011-12-04T20:34:21 contestbot: seen mcleopold 2011-12-04T20:34:21 amstan: mcleopold was last seen in #aichallenge 2 hours, 49 minutes, and 45 seconds ago: or start suiciding yourself to turn it off 2011-12-04T20:41:55 *** Antimony has quit IRC (Ping timeout: 244 seconds) 2011-12-04T20:46:11 *** Euphony has quit IRC (Ping timeout: 265 seconds) 2011-12-04T20:56:41 contestbot: seen j3camero 2011-12-04T20:56:41 amstan: j3camero was last seen in #aichallenge 6 weeks, 1 day, 5 hours, 43 minutes, and 5 seconds ago: If you are finding that some of the testing tools suck, please consider making better ones. We generally release whatever tools at the beginning, but the contestants always end up making better ones. 2011-12-04T20:57:51 except this time we had mleise :P 2011-12-04T20:59:03 amstan: heard from jeff at all if he is planning to come? 2011-12-04T20:59:13 janzert: i haven't gotten any replies 2011-12-04T20:59:19 :( 2011-12-04T20:59:28 from anyone but bmh.. so.. 2011-12-04T20:59:48 but from past experiences, he might pop in here at any time 2011-12-04T21:00:02 so i'm hoping for that 2011-12-04T21:00:06 yep 2011-12-04T21:04:35 McLeopold: ping 2011-12-04T21:08:29 antimatroid around? 2011-12-04T21:10:25 *** Antimony has joined #aichallenge 2011-12-04T21:10:29 <_flag> So there seem to be very few games on the official server played on either random_walk maps or maze maps, are they being phased out? 2011-12-04T21:12:47 _flag: it's just that right now the generators for those have yet to be updated to always produce good maps and a large number of the cell maps have been added in the meantime 2011-12-04T21:15:09 *** olexs1 has quit IRC (Quit: Leaving.) 2011-12-04T21:15:16 <_flag> janzert: Okay, thanks. Do you think the contest will end as planned or might it be extended like last year? 2011-12-04T21:15:33 _flag: it was never extended 2011-12-04T21:15:36 <_flag> Because it's getting awfully close and it would be nice to have the final maps ready 2011-12-04T21:15:37 ? it has never been extended that I'm aware of 2011-12-04T21:15:42 <_flag> Really? 2011-12-04T21:15:46 nope 2011-12-04T21:15:46 _flag: never 2011-12-04T21:15:49 hmm, to resubmit or not to resubmit 2011-12-04T21:15:55 <_flag> I thought pw was a month or so longer than intended? 2011-12-04T21:16:03 nope 2011-12-04T21:16:18 it was a month or so longer than this one, but that was planned from the start of it 2011-12-04T21:16:19 <_flag> Okay, nevermind then 2011-12-04T21:16:47 *** avdg has quit IRC (Quit: Leaving.) 2011-12-04T21:17:34 i also emphasize that you shouldn't make expectations how maps should be beyond the game specs 2011-12-04T21:17:57 if some maps don't respect those, that's our bad, and you can expect we'll fix it, but nothing more 2011-12-04T21:18:47 *** keith__ has joined #aichallenge 2011-12-04T21:19:15 <_flag> Obviously I wouldn't do that intentionally, but there is always a possibility that if I only test on the maps that are avaliable that I would never know if I am making a mistake that might hurt me on other maps 2011-12-04T21:27:03 *** aarossig has joined #aichallenge 2011-12-04T21:28:30 the solution: test on all possible maps 2011-12-04T21:29:49 *** avdg has joined #aichallenge 2011-12-04T21:36:31 *** Antimony_ has joined #aichallenge 2011-12-04T21:37:12 *** Antimony has quit IRC (Ping timeout: 244 seconds) 2011-12-04T21:37:15 *** Antimony_ is now known as Antimony 2011-12-04T21:45:03 seen antimatroid 2011-12-04T21:46:12 @seen antimatroid 2011-12-04T21:46:12 janzert: antimatroid was last seen in #aichallenge 48 minutes and 21 seconds ago: except this time we had mleise :P 2011-12-04T21:46:29 *** gkhjh has joined #aichallenge 2011-12-04T21:51:03 a1k0n: thanks for resubmitting ;-) 2011-12-04T21:51:10 :| 2011-12-04T21:51:17 also re: test on all possible maps, I did make such a script 2011-12-04T21:51:20 takes quite a while to run... 2011-12-04T21:55:10 *** BenJacks1n is now known as BenJackson 2011-12-04T21:55:17 where did that 1 come from 2011-12-04T21:56:11 I've become pessimistic about being able to get combat working before the end of the contest 2011-12-04T21:57:00 I was pretty pessimistic about my last couple of submits 2011-12-04T21:57:10 not winning as reliably as I'd like 2011-12-04T21:57:27 but suddenly I break into the top 10 and I'm excited about hacking again ;-) 2011-12-04T21:58:52 i think i'm done with my bot 2011-12-04T22:00:37 *** Antimony has quit IRC (Ping timeout: 248 seconds) 2011-12-04T22:06:06 *** gkhjh has quit IRC (Quit: Page closed) 2011-12-04T22:08:30 hence the resub? 2011-12-04T22:08:34 yeah 2011-12-04T22:08:50 this version is measurably stronger than the last one but i'm not sure whether it's the strongest of all 2011-12-04T22:09:22 mine was a huge improvement over the earlier ones but despite having some variation in my own bots it's probably overspecialized 2011-12-04T22:10:26 I'm worried about combat too - I've taken so long to get anything half-decent in terms of exploration and food-gathering - combat, defence, and hill-attacking seems like a lot in the short-time remaining :/ 2011-12-04T22:11:54 a1k0n: what did you change? 2011-12-04T22:12:04 I'm in the middle of porting it to C and that's taking longer than I expected (my C is a bit rusty and I forgot how much effort error-checking and good memory management takes) 2011-12-04T22:12:12 use c++ 2011-12-04T22:12:22 at least you can encapsulate the memory management 2011-12-04T22:12:27 even if you don't use any other features 2011-12-04T22:13:10 *** keith___ has joined #aichallenge 2011-12-04T22:14:48 I wanted to use objective-c, but that's not available so I wrote some obj-c-like behaviour in C, including a FOR_IN macro and alloc/init/dealloc style functions on each object type - didn't bother with reference counting though as I can easily know at an object-level when things are done with for most cases 2011-12-04T22:15:47 *** jstrong has quit IRC (Ping timeout: 240 seconds) 2011-12-04T22:16:00 *** amstan_ has joined #aichallenge 2011-12-04T22:16:00 *** ChanServ sets mode: +o amstan_ 2011-12-04T22:16:59 implementing lists and queues and the like in C was good for brushing up on my C skills :D 2011-12-04T22:19:57 *** keith___ has quit IRC (Ping timeout: 265 seconds) 2011-12-04T22:31:27 *** yoden has quit IRC (Quit: Leaving.) 2011-12-04T22:33:06 *** Antimony has joined #aichallenge 2011-12-04T22:35:21 *** jstrong has joined #aichallenge 2011-12-04T22:46:38 *** Kommander has quit IRC () 2011-12-04T22:46:47 *** Antimony has quit IRC (Ping timeout: 240 seconds) 2011-12-04T22:49:30 *** dvladim has joined #aichallenge 2011-12-04T22:54:36 *** Antimony has joined #aichallenge 2011-12-04T23:11:25 *** avdg has quit IRC (Quit: Leaving.) 2011-12-04T23:13:13 *** antimatroid1 has joined #aichallenge 2011-12-04T23:13:13 *** antimatroid has quit IRC (Read error: Connection reset by peer) 2011-12-04T23:14:27 *** amstan_ has quit IRC (Ping timeout: 240 seconds) 2011-12-04T23:14:36 * amstan was looking at a1k0n's strategy post: http://forums.aichallenge.org/viewtopic.php?f=24&t=2044 2011-12-04T23:14:45 yes... i know some of these words... 2011-12-04T23:14:49 *** dvladim has quit IRC (Ping timeout: 252 seconds) 2011-12-04T23:14:55 many of them are irrelevant. the idea is simple 2011-12-04T23:15:16 yes, but i have no idea what you are saying 2011-12-04T23:15:23 :( 2011-12-04T23:15:31 my post is way too long 2011-12-04T23:15:47 a1k0n: maybe add pictures 2011-12-04T23:16:25 i remember your postmortem, i really liked that animation which showed who could get to certain squares the fastest 2011-12-04T23:17:06 yeah, like ben's too on the forums 2011-12-04T23:20:09 *** jab_bott has quit IRC (Quit: Page closed) 2011-12-04T23:24:17 *** nalaginrut has joined #aichallenge 2011-12-04T23:24:19 *** AdrianG has joined #aichallenge 2011-12-04T23:29:14 *** Antimony_ has joined #aichallenge 2011-12-04T23:31:41 *** Antimony has quit IRC (Ping timeout: 252 seconds) 2011-12-04T23:31:54 *** Antimony_ is now known as Antimony 2011-12-04T23:40:06 *** AdrianG has quit IRC () 2011-12-04T23:40:11 *** antimatroid1 has quit IRC (Ping timeout: 252 seconds) 2011-12-04T23:43:21 *** Antimony has quit IRC (Ping timeout: 245 seconds) 2011-12-04T23:58:27 *** delt0r_ has quit IRC (Ping timeout: 240 seconds)