--- Log opened Wed Mar 03 00:00:56 2004
00:05:35 -!- Roy [~roy_van_r@kf-mss-cb01-00480.dial.kabelfoon.nl] has joined #corewars
00:05:36 #corewars: < Roy> Hi
00:06:06 #corewars: < Mizcu> Roy
00:10:21 #corewars: < brx> hi
00:13:42 #corewars: < Roy> Will is posting on the gridwar forum :) yay
00:17:11 #corewars: < brx> what is gridwar
00:18:58 #corewars: < brx> ah my optimizer is ready
00:19:25 #corewars: < Roy> www.gridwars.com
00:19:42 #corewars: < brx> and it is *FSCKING* slow
00:19:59 #corewars: < Roy> Its a bit like corewars, also a programming game for memory
00:20:12 #corewars: < brx> i will take a look!
00:20:28 #corewars: < Mizcu> brx: how many fights and how many opponents in the benchmark?
00:20:34 #corewars: < brx> and now watch perl creating perl and running a loop that will stall my pc for aeons!
00:20:52 #corewars: < brx> Mizcu: it's not taht..it's the patternmatching and the LONG NESTED LOOPS
00:21:00 #corewars: < brx> for(500 .. 1666){@template[$constants{step2}[3]]=~s[(?<=equ\s)\s*\S+][$_];for(1000 .. 3000){@template[$constants{step1}[3]]=~s[(?<=equ\s)\s*\S+][$_];for(200 .. 500){@template[$constants{dc2}[3]]=~s[(?<=equ\s)\s*\S+][$_];open(TW, ">$twfile") or die "Couldn't open $twfile for writing!\n";print TW @template;close TW;`$koth $twfile $rounds $testset`=~m[.*(?<!\d)(\d+)(?:.+?\d+\.\d+){3,3}(?!.*?\d)]s;if($1>$highscore){$highscore=$1;copy($twfile, "$warrior.be
00:21:07 #corewars: < brx> this is the  loop generated by my program
00:21:17 #corewars: < Mizcu> oorgh..
00:21:22 #corewars: < brx> omg this is such a hack
00:21:29 #corewars: < brx> and written in the *.red is this
00:21:33 #corewars: < brx> step1 equ opt(1000-3000)
00:21:33 #corewars: < brx> step2 equ opt(500-1666)
00:21:33 #corewars: < brx> dc2 equ opt(200-500)
00:23:44 #corewars: < Mizcu> Oh well, my computer runs 12*75 matches in a minute and the chance to get a succesfull combination out of 4 tested stepsized is already small enough, not mention something like 12 on dual papers.. 
00:24:03 #corewars: < brx> so basically it will do the innermost loop (2000*1166*300) times
00:24:31 #corewars: < brx> and considering the patternmathing involved... and the fact that the benchmarking program was also written by me.. in perl 
00:24:42 #corewars: < brx> this will take a few hundred years
00:24:47 #corewars: < brx> and it's only 3 constants
00:25:34 #corewars: < michal> what are you optimizing?
00:26:38 #corewars: < brx> constants
00:26:49 #corewars: < brx> the program's name is copti
00:27:09 #corewars: < Mizcu> and what kind of warrior will go into the fray?
00:27:22 #corewars: < brx> it doesn't matter.
00:27:27 #corewars: < brx> but it's a paper
00:27:35 #corewars: < brx> i am only testing it on a simple paper i wrote
00:27:49 #corewars: < brx> i will look for the results tomorrow
00:27:55 #corewars: < brx> it's running in the background atm :)
00:28:01 #corewars: < michal> what simulator are you using?
00:28:08 #corewars: < brx> pmars server
00:28:36 #corewars: < michal> exmars is 50% faster on average..
00:28:37 #corewars: < michal> or..
00:28:42 #corewars: < brx> believe me, it is not the simulator's fault. it's my programs fault or rather the concept of trying all possible combinations of constants
00:29:14 #corewars: < michal> can't you randomize constants a bit?
00:29:28 #corewars: < brx> it will traverse the inner loop 2000*1166*300 times. perl pattern matching and diskaccess will be >90% of the runtime
00:29:41 #corewars: < michal> I have my my mostly working.. setting it up is a bit tricky for now
00:29:47 #corewars: < michal> but you might give it a try
00:29:55 #corewars: < brx> why should i randomize.. like that? "opt(rand(100))" meaning try 100 random values? :)
00:30:08 #corewars: < Mizcu> with some elimination of absurd numbers you could get couple'o'numbers less
00:30:17 #corewars:  * Roy does it with random numbers :)
00:30:29 #corewars: < brx> that's why i am giving copti ranges in my red file
00:30:34 #corewars: < brx> 1000-3000 :)
00:30:39 #corewars: < brx> 500-1666
00:30:41 #corewars: < brx> 200-500
00:30:48 #corewars: < brx> i am only testing how long it will take
00:30:58 #corewars: < michal> I have my *mars* mostly working :)
00:31:30 #corewars: < Mizcu> but for example you can fit 1143, 2667 and 2001  in 1000-3000
00:31:42 #corewars: < brx> michal: :)
00:31:56 #corewars: < Mizcu> (1001 should be also absurd enough)
00:32:17 #corewars: < brx> i could test primenumbers :P
00:32:48 #corewars: < brx> aaah
00:32:54 #corewars: < brx> koth pmars and copti are on the top of top :)
00:33:13 #corewars:  * brx waits
00:33:26 #corewars: < brx> i will watch an anime
00:33:27 #corewars: < brx> ^^
00:33:43 #corewars: < michal> as for randomization, and syntax in the redcode file, I would do it like this: x equ opt (100 * rand (1000, 3000))
00:33:55 #corewars: < michal> but the anime will steal your cpu cycles..
00:34:01 #corewars: < brx> mhm michal 
00:34:24 #corewars: < brx> michal: i am NOT going to stare at copti's generated optimizer loop until i go to bed :)
00:34:28 #corewars: < brx> $loops .= q/for(/.qq/$constants{$_}[1] .. $constants{$_}[2]/.q/){@template[$constants{/.qq/$_/.q/}[3]]=~s[(?<=equ\s)\s*\S+][$_];/ for keys %constants;
00:34:31 #corewars: < brx> $loops .= q/open(TW, ">$twfile") or die "Couldn't open $twfile for writing!\n";print TW @template;close TW;`$koth $twfile $rounds $testset`=~m[.*(?<!\d)(\d+)(?:.+?\d+\.\d+){3,3}(?!.*?\d)]s;if($1>$highscore){$highscore=$1;copy($twfile, "$warrior.best");}/.q/}/x$cnum;
00:34:35 #corewars: < brx> this has generated the loop btw
00:34:37 #corewars: < brx> what an awful hack
00:34:46 #corewars: < brx> and stupid, i think i could have done it recursively
00:35:39 #corewars: < brx> and i also think i should have done it in lisp
00:38:06 #corewars: < brx> will anybody bet for copti NOT crashing before giving me a result?
00:38:49 #corewars: < Mizcu> how many rounds there are per each pmars run?
00:38:53 #corewars: < brx> only 10
00:38:56 #corewars: < brx> :)
00:38:59 #corewars: < Mizcu> ...
00:39:08 #corewars: < brx> it will take hours even with 10 rounds only
00:39:11 #corewars: < michal> oh, and that line noise reminds me:
00:39:21 #corewars: < michal> perl -e '$Q="^AQWEA%*ZXCRTYFGHVBNUIPXO(KL-QWE-TY#iOPASDFGH()L_-/ZJKlZXCVBNM<|1234S67890]"^"%a83~1RNa40ry5}7;v#;->k%3E*%Aq6+^.6N)9}-<j-5)CF;q]CzxuC>=5y,;!PZTQ[[s~RTU_|";print`$Q`'
00:39:30 #corewars: < michal> -- seen in my friend's .plan
00:39:36 #corewars: < brx> .plan?
00:40:00 #corewars: < michal> you know, when you `finger` someone
00:40:09 #corewars: < brx> has this entered an obsfucation contest?
00:40:23 #corewars: < Mizcu> i dont want to know much about that know..
00:40:29 #corewars: < Mizcu> 2nd know -> code
00:40:29 #corewars: < michal> I don't think so. Guess what it does?
00:40:38 #corewars: < brx> Mizcu: learn perl :)
00:40:45 #corewars: < brx> michal: you will tell us
00:41:54 #corewars: < michal> if I remember right, it executes some harmless commands [like `ls ~`] and sends output to his email
00:42:30 #corewars: < brx> lol
00:46:36 #corewars: < Mizcu> Who needs a perl obfuscator anyway
00:47:15 #corewars: < brx> :=
00:47:59 #corewars: < Mizcu> just throw it on 'net as it is like anyone is even going to bother trying ;)
00:59:00 -!- Roy [~roy_van_r@kf-mss-cb01-00480.dial.kabelfoon.nl] has quit []
00:59:32 #corewars: < michal> another wasted day
00:59:34 #corewars: < michal> good night :)
00:59:35 -!- michal [~michal@host-ip2-246.crowley.pl] has quit [Leaving]
00:59:37 #corewars: < brx> night :)
01:10:58 -!- brx [~as@pD9EA9B50.dip.t-dialin.net] has quit [Ping timeout: 180 seconds]
01:13:39 -!- Mizcu [Mizcu@dsl-hkigw4m4a.dial.inet.fi] has quit []
01:29:30 -!- AtnNn [~AtnNn@dsl-131-99.aei.ca] has quit [Ping timeout: 180 seconds]
02:23:56 -!- AtnNn [~AtnNn@dsl-143-110.aei.ca] has joined #corewars
02:27:30 -!- AtnNn [~AtnNn@dsl-143-110.aei.ca] has quit [Ping timeout: 180 seconds]
02:29:01 -!- AtnNn [~AtnNn@dsl-135-232.aei.ca] has joined #corewars
03:25:37 -!- bvowk [~bvowk@h24-67-137-90.ed.shawcable.net] has joined #corewars
03:25:38 #corewars: < bvowk> hrm
03:29:58 -!- bvowk [~bvowk@h24-67-137-90.ed.shawcable.net] has quit [using sirc version 2.211+ssfe]
04:08:39 -!- grabek [~grabek@nat-b.acn.pl] has quit [Ping timeout: 180 seconds]
04:10:29 -!- grabek [~grabek@nat-b.acn.pl] has joined #corewars
05:56:52 -!- grabek [~grabek@nat-b.acn.pl] has quit [Ping timeout: 180 seconds]
05:57:52 -!- grabek [~grabek@nat-b.acn.pl] has joined #corewars
06:15:13 -!- grabek [~grabek@nat-b.acn.pl] has quit [Ping timeout: 180 seconds]
06:16:16 -!- grabek [~grabek@nat-b.acn.pl] has joined #corewars
08:31:27 -!- willvarfa [~Will@arken-16-57-114.ip-pluggen.com] has joined #corewars
08:48:08 -!- will_varf [~Will@arken-16-57-58.ip-pluggen.com] has joined #corewars
08:51:14 -!- willvarfa [~Will@arken-16-57-114.ip-pluggen.com] has quit [Ping timeout: 180 seconds]
09:19:17 -!- willvarfa [~Will@arken-16-57-22.ip-pluggen.com] has joined #corewars
09:22:12 -!- will_varf [~Will@arken-16-57-58.ip-pluggen.com] has quit [Ping timeout: 180 seconds]
09:22:15 -!- willvarfa [~Will@arken-16-57-22.ip-pluggen.com] has quit [Lämnar]
09:27:04 -!- willvarfa [~Will@arken-16-57-22.ip-pluggen.com] has joined #corewars
09:50:19 -!- will_varf [~Will@arken-16-57-41.ip-pluggen.com] has joined #corewars
09:52:41 -!- willvarfa [~Will@arken-16-57-22.ip-pluggen.com] has quit [Ping timeout: 180 seconds]
10:21:31 -!- willvarfa [~Will@arken-16-57-77.ip-pluggen.com] has joined #corewars
10:24:23 -!- will_varf [~Will@arken-16-57-41.ip-pluggen.com] has quit [Ping timeout: 180 seconds]
10:53:39 -!- willvarfa [~Will@arken-16-57-77.ip-pluggen.com] has quit [Ping timeout: 180 seconds]
11:03:35 -!- Roy [~roy_van_r@guest-dhcp-281.si.hhs.nl] has joined #corewars
11:05:37 -!- Roy [~roy_van_r@guest-dhcp-281.si.hhs.nl] has quit [Read error: 54 (Connection reset by peer)]
11:37:51 -!- Roy [~roy_van_r@guest-dhcp-281.si.hhs.nl] has joined #corewars
11:37:57 #corewars: < Roy> Hi (again)
11:59:07 -!- willvarfa [~Will@arken-16-57-65.ip-pluggen.com] has joined #corewars
11:59:11 -!- willvarfa [~Will@arken-16-57-65.ip-pluggen.com] has quit [Lämnar]
13:10:41 -!- willvarfa [~Will@arken-16-57-52.ip-pluggen.com] has joined #corewars
13:10:44 #corewars: < willvarfa> hi folks
13:11:56 #corewars: < Roy> Hi will
13:12:11 #corewars: < Roy> I saw your post on gridwars, you planning on programming something?
13:13:50 #corewars: < willvarfa> well I thought about it
13:14:02 #corewars: < willvarfa> I don't really like the tools or the limitations
13:14:27 #corewars: < willvarfa> but I would like that server!
13:14:34 #corewars: < willvarfa> yourself?
13:18:08 -!- Mizcu [Mizcu@dsl-hkigw4m4a.dial.inet.fi] has joined #corewars
13:21:47 #corewars: < Roy> I tried a couple of strategies, nothing really worked well
13:21:52 #corewars: < Roy> But I'll keep trying :P
13:22:16 #corewars: < Roy> There are diff ways, using that cpp file you can use your own tools etc
13:26:22 -!- will_varf [~Will@arken-16-57-71.ip-pluggen.com] has joined #corewars
13:28:55 -!- willvarfa [~Will@arken-16-57-52.ip-pluggen.com] has quit [Ping timeout: 180 seconds]
13:30:38 #corewars: < will_varf> can you use cpp?  I thought you had to use CxC (12 variables etc)
13:32:56 #corewars: < will_varf> you posted in the forums?
13:33:28 #corewars: < Roy> Yep, just some simple questions
13:33:40 #corewars: < Roy> Well yes, you have to use CxC
13:33:53 #corewars: < will_varf> my gridwars crashes when I run my warrior; I can't see why
13:33:54 #corewars: < Roy> BUT you can use a cpp to run the .run file
13:34:07 #corewars: < will_varf> ah cool
13:34:19 #corewars: < Roy> Sometimes mine freezes, just a never-ending loop :)
13:34:33 #corewars: < Roy> I can send you one of my test warriors..? want it?
13:34:42 #corewars: < will_varf> compiling takes about 8 seconds on my P4 2.4Ghz 500MB-ram box; can't see a GA wrapper working through very many combinations at all
13:34:51 #corewars: < will_varf> yes please!
13:35:10 #corewars: < Roy> It isn't good! (only sometimes in 1on1)
13:35:38 #corewars: < Roy> But we could learn from eachother (tricks) could improve both warriors
13:45:38 -!- Roy [~roy_van_r@guest-dhcp-281.si.hhs.nl] has quit []
13:53:31 -!- will_varf [~Will@arken-16-57-71.ip-pluggen.com] has quit [Lämnar]
14:01:36 -!- willvarfa [~Will@arken-16-57-79.ip-pluggen.com] has joined #corewars
14:25:19 -!- will_varf [~Will@arken-16-57-89.ip-pluggen.com] has joined #corewars
14:28:18 -!- willvarfa [~Will@arken-16-57-79.ip-pluggen.com] has quit [Ping timeout: 180 seconds]
14:54:53 -!- will_varf [~Will@arken-16-57-89.ip-pluggen.com] has quit [Ping timeout: 180 seconds]
15:03:01 -!- willvarfa [~Will@arken-16-57-13.ip-pluggen.com] has joined #corewars
15:12:13 -!- brx [~as@pD9EAB639.dip.t-dialin.net] has joined #corewars
15:12:16 #corewars: < brx> w00t
15:12:24 #corewars: < brx> guess who has successfully finished school ^^
15:12:29 #corewars: < brx> hi guys ;)
15:13:05 #corewars: < willvarfa> for ever?
15:13:08 #corewars: < brx> no :(
15:13:11 #corewars: < brx> university.
15:13:14 #corewars:  * brx sighs
15:13:17 #corewars: < brx> no wait
15:13:20 #corewars:  * brx cheers
15:13:30 #corewars: < brx> i am looking forward to it
15:13:51 #corewars: < willvarfa> congratulations?
15:13:57 #corewars: < willvarfa> just had your last exam or something?
15:13:59 #corewars:  * brx nods
15:14:08 #corewars: < brx> well today i got the exam results!
15:14:13 #corewars: < willvarfa> ah cool!
15:14:19 #corewars: < willvarfa> what uni do you want to go to?
15:14:25 #corewars: < brx> i have passed, i can even fail the next oral exam as much as i want to
15:14:26 #corewars: < brx> i did it.
15:14:36 #corewars: < brx> near leipzig :)
15:14:41 #corewars: < willvarfa> very cool
15:14:48 #corewars: < willvarfa> congratulations!
15:15:00 #corewars: < willvarfa> how do you plan to celebrate?
15:15:34 #corewars: < brx> well this evening i am going to meet my collegues of sp6
15:15:56 #corewars: < brx> play some soccer, then go into a bar afterwards
15:16:19 #corewars: < brx> thank you :)
15:16:47 #corewars: < brx> (sidenote: copti is still benchmarking)
15:17:30 #corewars: < willvarfa> copti?
15:17:51 #corewars: < brx> constants optimizer. i wrote a perl program yesterday to optimize constants in warriors
15:18:08 #corewars: < brx> it has been working for >14 hours now
15:19:27 #corewars: < brx> omg. it's not even finished 20% of its work
15:20:12 #corewars: < willvarfa> lol
15:20:18 #corewars: < willvarfa> has it found anything interesting yet?
15:20:52 #corewars: < brx> it will give me the "optimal" combination of constants it finds.
15:20:55 #corewars: < brx> so. no.
15:21:04 #corewars: < brx> i don't even know what the highest score was
15:21:08 #corewars:  * brx sighs
15:23:00 #corewars: < willvarfa> but you will on sunday?
15:23:23 #corewars: < brx> i hope so.
15:25:32 #corewars: < willvarfa> best of luck :-)
15:25:40 #corewars: < willvarfa> it is always a waiting game
15:25:50 #corewars: < willvarfa> how many variables are you considering?
15:27:16 #corewars: < brx> only 3
15:27:22 #corewars: < brx> it was only a TEST warrior.
15:27:25 #corewars: < brx> a simple paper.
15:27:34 #corewars: < brx> i just wanted to.... time copti
15:27:42 #corewars: < brx> but this takes a massive amount of time
15:27:50 #corewars: < brx> i am considering a rewrite
15:29:05 -!- will_varf [~Will@arken-16-57-22.ip-pluggen.com] has joined #corewars
15:31:29 -!- willvarfa [~Will@arken-16-57-13.ip-pluggen.com] has quit [Ping timeout: 180 seconds]
15:32:52 #corewars: < will_varf> how would you rewrite?  what would you change?
15:33:24 #corewars: < brx> well i might consider writing a mars in c
15:33:41 #corewars: < brx> before that i could optimize the patternmatching in my perlprogram
15:34:14 #corewars: < will_varf> the actual invoking of a mars from perl does take a long time; can perl run c apps in-process?
15:34:35 #corewars: < brx> i don't think so actually
15:35:20 #corewars: < will_varf> are you testing each combination of variables against a benchmark?
15:35:27 #corewars: < brx> yup
15:35:33 #corewars: < brx> ah yes, that's another point
15:35:37 #corewars: < will_varf> which benchmark?
15:35:42 #corewars: < brx> either rule out absurd number
15:35:42 #corewars: < brx> s
15:35:56 #corewars: < brx> or resort to a random number solution
15:35:59 #corewars: < brx> syntax
15:36:22 #corewars: < brx> like "constant equ opt(100; 1000-2000)"
15:36:40 #corewars: < brx> test 100 numbers in the range of.
15:36:47 #corewars: < will_varf> it might be worth first fighting against a wimp; if the combination can't kill it in a single fight, why bother testing it against the benchmark?
15:36:51 #corewars: < brx> i think that's what roy does
15:36:55 #corewars: < brx> hm
15:37:08 #corewars: < will_varf> (that is what evolvers do)
15:37:32 #corewars: < brx> is it?
15:37:52 #corewars: < will_varf> also, as you progress through the benchmark, you can start to work out if the warrior can possibly be the best warrior if it wins all future fights; if it can't, you can abort testing it
15:38:01 #corewars: < brx> it does not sound like the optimal way to evolve.
15:38:27 #corewars: < brx> mhm
15:38:43 #corewars: < brx> well i need to duplicate some code then.
15:38:48 #corewars: < brx> but that's fine i suppose
15:39:03 #corewars: < will_varf> are you comfortable with c?
15:39:07 #corewars: < brx> sure
15:41:10 #corewars: < will_varf> then you might want to take a look at exMars, soon to be polished by Martin Ankerl
15:41:18 #corewars: < will_varf> (a gzip of progress is available)
15:41:43 #corewars: < will_varf> it is the fastest mars so far, written in C; it includes pmar's parser, which is a major boon
15:41:56 #corewars: < will_varf> it would make a good engine for an optimiser
15:42:14 #corewars: < will_varf> (and there is to be a ruby interface, if you want to write the higher controls in ruby instead)
15:43:22 #corewars: < brx> hmm
15:43:27 #corewars: < brx> thank you for the advise
15:43:32 #corewars: < brx> i will give it a shot
15:45:38 #corewars: < will_varf> it would be a fun project
15:45:41 #corewars: < will_varf> (or a dmars)
15:46:08 #corewars: < brx> dmars? :)
15:48:25 #corewars: < will_varf> distributed mars
15:49:07 #corewars: < will_varf> whilst brainstorming has lead to a straightforward definition, and I've written a local-loopback version (completely untested, but it compiles), no-one has yet written a distributed version
15:49:19 #corewars: < will_varf> just another fun project I was throwing down to you ;-)
15:49:23 #corewars: < brx> :)
15:49:33 #corewars: < brx> oooooh throwing doooooown.. i see ^^
15:50:51 #corewars: < will_varf> lol
15:51:15 #corewars: < will_varf> you students between school and uni need something to keep you inside eating pizza
15:52:01 #corewars: < will_varf> when you mentioned going out, seeing people, drinking (when not alone), playing football.. well, you obviously need projects!
15:52:20 #corewars: < brx> lol
15:52:23 #corewars: < brx> i see :)
15:52:36 #corewars: < brx> don't worry, i will have projects in the near future
15:52:39 #corewars: < brx> quite a few actually
15:52:45 #corewars: < brx> happy ^^
15:52:56 #corewars:  * brx bounces like a codemonkey would
15:53:57 #corewars: < will_varf> np
15:54:27 #corewars: < will_varf> a simple distributed version would be relatively easy
15:54:58 #corewars: < will_varf> a good distributed version that tried hard not to waste a cycle would be a real pure piece of software engineering, and worthy of bragging about (except when out socialising, obviously)
15:55:06 #corewars: < brx> well i am talking about some lisp projects :)
15:55:13 #corewars: < will_varf> lisp?
15:55:17 #corewars: < brx> ^^
15:55:21 #corewars: < brx> i could write a mars in lisp
15:55:46 #corewars: < will_varf> you probably could
15:56:00 #corewars: < will_varf> do you like lisp?
15:56:16 #corewars: < brx> yes
15:56:23 #corewars: < brx> although i don't have much experience with it yet
15:56:35 #corewars: < brx> though writing perlgenerating perlcode i longed for it
15:57:05 #corewars: < will_varf> bb 5mins
15:57:07 -!- will_varf [~Will@arken-16-57-22.ip-pluggen.com] has quit [Lämnar]
15:59:13 -!- willvarfa [~Will@arken-16-57-9.ip-pluggen.com] has joined #corewars
15:59:16 #corewars: < willvarfa> back
15:59:39 #corewars: < brx> wb
16:28:58 -!- will_varf [~Will@arken-16-57-97.ip-pluggen.com] has joined #corewars
16:31:57 -!- willvarfa [~Will@arken-16-57-9.ip-pluggen.com] has quit [Ping timeout: 180 seconds]
16:33:20 -!- bvowk [~bvowk@h24-67-137-90.ed.shawcable.net] has joined #corewars
16:33:22 #corewars: < bvowk> hrm
16:33:27 #corewars:  * bvowk rattles wil
16:36:04 #corewars: < brx> bvowk
16:36:26 #corewars: < bvowk> greets
16:41:14 #corewars:  * bvowk tests his greylisting magic
16:41:30 #corewars: < brx> ?
16:43:43 #corewars: < will_varf> hello
16:43:51 #corewars: < bvowk> spam control brx
16:43:55 #corewars: < bvowk> hey will.. how's it going
16:47:36 #corewars: < will_varf> busy day, busy day
16:47:38 #corewars: < will_varf> yourself?
16:47:50 #corewars: < bvowk> I think I'm going to go buy an ipod :)
16:54:14 #corewars: < will_varf> cool :-)
17:00:01 -!- willvarfa [~Will@arken-16-57-107.ip-pluggen.com] has joined #corewars
17:02:29 -!- will_varf [~Will@arken-16-57-97.ip-pluggen.com] has quit [Ping timeout: 180 seconds]
17:08:42 -!- willvarfa [~Will@arken-16-57-107.ip-pluggen.com] has quit [Lämnar]
17:30:10 -!- Mizcu [Mizcu@dsl-hkigw4m4a.dial.inet.fi] has quit []
18:29:38 -!- willvarfa [~Will@arken-16-57-54.ip-pluggen.com] has joined #corewars
18:32:20 #corewars: < willvarfa> hi
18:42:04 -!- michal [~michal@host-ip2-246.crowley.pl] has joined #corewars
18:49:00 -!- will_varf [~Will@arken-16-57-19.ip-pluggen.com] has joined #corewars
18:51:18 -!- willvarfa [~Will@arken-16-57-54.ip-pluggen.com] has quit [Ping timeout: 180 seconds]
18:53:24 -!- Metcalf [~John@217.158.137.194] has joined #corewars
18:53:27 #corewars: < Metcalf> Hi
18:55:28 #corewars: < will_varf> hello john
18:55:31 #corewars: < will_varf> how goes things?
18:55:50 #corewars: < Metcalf> HI Will.
18:56:09 #corewars: < Metcalf> Okay.  Just bought a CD writer for my laptop, so I'm in a good mood.
18:56:27 #corewars:  * Metcalf has never had a writer before
19:02:59 #corewars: < Metcalf> Just copied my corewars directory to CD.  takes up 200 meg and 20,000+ files
19:03:14 #corewars:  * Metcalf was astonished
19:04:36 #corewars: < will_varf> wow
19:04:39 #corewars: < will_varf> that is quite a lot
19:07:52 #corewars: < Metcalf> Lots of tested tweaks I assume... and benchmark results... and probably lots of duplicates.
19:09:03 #corewars: < will_varf> have you ever thought about entering gridwars?
19:14:17 #corewars: < Metcalf> No
19:14:53 #corewars: < Metcalf> The site took too long to navigate when I visited it.
19:15:07 #corewars: < Metcalf> And there was something which put me off, can't remember what exactly
19:17:28 #corewars: < will_varf> yeap, it doesn't feel quite *right* somehow
19:19:53 -!- willvarfa [~Will@arken-16-57-55.ip-pluggen.com] has joined #corewars
19:21:03 #corewars: < Metcalf> "
19:21:05 #corewars: < Metcalf> I have deduced that there must be 'Gods corewarrior'£
19:21:18 #corewars: < willvarfa> ?
19:22:49 -!- will_varf [~Will@arken-16-57-19.ip-pluggen.com] has quit [Ping timeout: 180 seconds]
19:23:02 #corewars: < willvarfa> bbs
19:24:27 #corewars: < Metcalf> The first infinite hill? http://tinyurl.com/3dx99
19:25:25 #corewars: < bvowk> '93?
19:26:24 #corewars: < bvowk> thats almost a decade ago
19:27:25 #corewars: < bvowk> erm.
19:27:30 #corewars: < bvowk> thats *MORE* than a decade ago
19:28:12 #corewars: < Metcalf> We have decimal decades over here bvowk...
19:28:45 #corewars: < Metcalf> no *clicking* today, thanks
19:30:29 #corewars: < jaska> bvowk uses base-16 decades?
19:36:12 #corewars: < Metcalf> Hmmm... Moore thinks Gisela 609 is a bomb-dodger
19:40:56 #corewars: < willvarfa> was it?
19:41:14 #corewars: < Metcalf> Erm...
19:43:14 #corewars: < Metcalf> No, it's a oneshot.
19:50:58 -!- will_varf [~Will@arken-16-57-22.ip-pluggen.com] has joined #corewars
19:52:39 -!- Roy [~roy_van_r@kf-mss-cb01-00133.dial.kabelfoon.nl] has joined #corewars
19:52:44 #corewars: < Roy> ello
19:53:27 #corewars: < brx> ello
19:53:39 #corewars: < Roy> Hi :P
19:53:57 -!- willvarfa [~Will@arken-16-57-55.ip-pluggen.com] has quit [Ping timeout: 180 seconds]
19:54:49 #corewars: < will_varf> hello Roy
19:54:55 #corewars: < Roy> hi
19:54:56 #corewars: < brx> Hi Hi :P
19:55:20 #corewars: < michal> hi
19:55:53 #corewars: < brx> hi michal 
19:55:57 #corewars: < Roy> yeah yeah, hi hi hi hi!
19:56:00 #corewars: < Metcalf> Hi
19:56:05 #corewars: < Metcalf> Time to go almost
19:58:16 #corewars: < Metcalf> Anyone know how to get hold of the complete contents of r.g.cw?
19:58:29 #corewars: < Metcalf> So I can read the lot!
19:58:31 #corewars:  * will_varf waves
19:58:33 #corewars:  * Metcalf waves
19:58:38 -!- Metcalf [~John@217.158.137.194] has quit [Leaving]
19:58:45 #corewars:  * michal waves
20:02:27 #corewars: < brx> bye!
20:02:38 #corewars: < will_varf> cya
20:12:30 !irc.koth.org *** Looking up your hostname...
20:12:30 !irc.koth.org *** Checking Ident
20:12:30 !irc.koth.org *** No Ident response
20:12:31 !irc.koth.org *** Found your hostname
20:12:32 !irc.koth.org *** Your host is irc.koth.org[216.231.108.170/6667], running version 2.8/hybrid-6.3.1
20:12:32 -!- Mode change [+i] for user jk_log
20:12:32 -!- jk_log [~jakub@134.gymko.ba.gtsi.sk] has joined #corewars
20:22:06 -!- willvarfa [~Will@arken-16-57-99.ip-pluggen.com] has joined #corewars
20:25:02 -!- will_varf [~Will@arken-16-57-22.ip-pluggen.com] has quit [Ping timeout: 180 seconds]
20:39:19 #corewars: < willvarfa> I have to go
20:39:24 #corewars: < willvarfa> I'll be away until next week
20:39:30 #corewars: < willvarfa> may the core/grid be with you!
20:39:31 #corewars: < Roy> Bye will! keep me updated ;)
20:43:09 #corewars: < grabek> he will
20:43:11 #corewars:  * willvarfa *waves*
20:43:13 -!- willvarfa [~Will@arken-16-57-99.ip-pluggen.com] has quit [Lämnar]
20:43:15 #corewars: < grabek> or rather: will will
20:45:10 #corewars: < Roy> will would wander with wisdom
20:45:36 #corewars: < grabek> surely
20:45:44 #corewars:  * grabek goes check his pork chops
20:46:20 #corewars: < bvowk> mmm... pork chops
20:46:34 #corewars: < grabek> mashed potatoes
20:46:44 #corewars: < grabek> and spinach with an egg
20:46:46 #corewars: < bvowk> I'll be by for dinner
20:48:04 #corewars: < grabek> i knew it
20:49:39 #corewars: < grabek> i am on my dinner quest
20:49:41 #corewars: < grabek> yummy
20:50:04 -!- Roy [~roy_van_r@kf-mss-cb01-00133.dial.kabelfoon.nl] has quit []
20:53:04 #corewars: < brx> grabek: 
20:53:06 #corewars: < brx> pork chips
20:53:08 #corewars: < brx> yuck
20:53:15 #corewars: < brx> copti been running for 19 hours straight
20:55:21 #corewars: < michal> maybe you should consider running it on a ramdisk?
21:03:51 #corewars: < grabek> chip yourself, brx :-)
21:06:06 #corewars: < brx> chipo?
21:06:11 #corewars: < brx> michal: !!!
21:06:16 #corewars: < brx> michal: !!!!!!!11
21:06:29 #corewars: < brx> omfg
21:06:43 #corewars: < michal> hm?
21:06:52 #corewars: < brx> <- stupid
21:06:56 #corewars: < brx> i thank thee michal
21:07:16 #corewars: < brx> (actually i have never found a use for ramdisks, but this is an eyeopener)
21:24:13 -!- michal [~michal@host-ip2-246.crowley.pl] has quit [Ping timeout: 180 seconds]
21:26:52 -!- michal [~michal@host-ip2-246.crowley.pl] has joined #corewars
22:06:47 -!- joonas [jpihlaja@kruuna.helsinki.fi] has joined #corewars
22:07:24 #corewars: < joonas> hello all
22:08:34 #corewars: < bvowk> hey joonas.. stick around.
22:08:38 #corewars: < joonas> ok
22:09:13 #corewars: < joonas> woot!
22:09:15 #corewars: < joonas> oops
22:10:37 #corewars: < grabek> woot
22:10:42 #corewars: < grabek> back to my books
22:10:50 #corewars: < grabek> forgive me, joonas 
22:11:06 #corewars: < joonas> ok, that's fine.  keep to those books.
22:11:24 #corewars: < joonas> tell us when you're done reading tho.
22:11:31 #corewars: < grabek> i have my exam on 16th
22:11:33 #corewars: < joonas> for good.
22:11:35 #corewars: < grabek> not much time
22:12:00 #corewars: < grabek> and loads of things to read
22:12:00 #corewars: < joonas> best of luck to you, then.
22:12:04 #corewars: < grabek> thank you
22:12:06 #corewars:  * grabek bows
22:12:12 #corewars:  * grabek waves unhapilly
22:12:38 #corewars:  * joonas tiptoes so as not to disturb lukasz
22:13:17 #corewars: < grabek> i can't hear loudest yell from Helsinki here, in Warsaw
22:13:22 #corewars: < grabek> don't bother :-)
22:13:25 #corewars: < grabek> cu later
22:13:28 #corewars: < joonas> bye
22:19:38 -!- michal [~michal@host-ip2-246.crowley.pl] has quit [Ping timeout: 180 seconds]
22:21:43 -!- michal [~michal@host-ip2-246.crowley.pl] has joined #corewars
22:23:12 #corewars: < joonas> brb
22:27:18 #corewars:  * bvowk rattles joonas
22:27:31 #corewars: < bvowk> christ.. I thought that guy would never leave
22:33:42 #corewars:  * bvowk rattles joonas again
22:41:05 #corewars: < joonas> rattles
22:41:19 #corewars: < joonas> who?
22:41:30 #corewars:  * joonas enjoys his coffee
22:59:19 -!- Jeff_K [~sascha@dial-195-14-250-144.netcologne.de] has joined #corewars
22:59:26 #corewars: < joonas> hi
22:59:37 #corewars: < Jeff_K> Hi
23:00:18 #corewars: < Jeff_K> I saw in the log's that brx is writing at an random-optimizer in Perl...Brx are you here ?
23:00:33 #corewars: < joonas> he must be sleeping.
23:00:39 #corewars: < joonas> was here all day I think.
23:00:46 #corewars: < joonas> (according to logs, at least.)
23:00:58 #corewars: < Jeff_K> Yeah, that is what the logs say ;-)
23:01:10 #corewars: < Jeff_K> Hmm, i didn't have his eMail-adress !
23:01:23 #corewars: < Jeff_K> Want to give him some tipps..
23:01:26 #corewars: < joonas> doesn't it show when he logs on to #corewars ?
23:01:36 #corewars: < joonas> well why not post them to r.g.c!
23:01:51 #corewars: < joonas> ?, I mean.  but ! is appropriate.
23:02:32 #corewars: < Jeff_K> Hmm, an nice article "thougt's of random-optimizng for CW is a nice idea too. !
23:02:58 #corewars:  * joonas has lots to say on the topic, but is too lazy to write them down.
23:03:40 #corewars: < joonas> you having more recent experience on the topic are an ideal person to write such an article.
23:04:40 #corewars: < Jeff_K> Same to me. While writing on optimax there are many ides arround but i can't implement not 10% of them, so give it away to all can be a nice idea. Will seems to be a good toolsmith too. !
23:05:21 #corewars: < joonas> heh
23:05:56 #corewars: < Jeff_K> O.K., but first i must relase optimax 1.0, and thats very hard. Push it from "works for me" to running stable at Linux and Windows is very hard .
23:06:01 #corewars: < joonas> personally, I think a lot more thought needs to go into reducing the effect of randomness when evaluating opponents.
23:06:10 #corewars: < joonas> yes.
23:06:26 #corewars: < joonas> fortunately you have willing beta testers!
23:06:32 #corewars: < joonas> for both platforms
23:08:06 #corewars: < Jeff_K> I'm working on something called "dynamic operators", they can calculate ranges where are effektic contants are. That works good for some kind of constants..for Example..fuse, start of djn-stream or decoy's lenght, but fail nicely if papersteps are to optimize.
23:08:32 #corewars: < joonas> intriguing
23:08:54 #corewars: < joonas> how do you calculate ranges?
23:09:39 #corewars:  * joonas just asked Jeff_K the secret to the family soup recipe :)
23:10:53 #corewars: < Jeff_K> For each Randry ( so called the generated warrior ) optimax stores the DNA, a string with all randomized constants. Then let them run for 100 Randry and the look in the top ten if there are constants that near by..And the fix them range or make it smaller..O.K, i give you an example.
23:12:10 #corewars: < Jeff_K> djn-stream of a clear, for mutate the bomb s/d/d for example. first the dynamic operator is "const equ !d(100-7900)
23:13:01 #corewars: < Jeff_K> then, after 100 Randry the DNA of the top ten has a range from 4700-6000. So the operator will trim his range to that...
23:13:46 #corewars: < joonas> I see.
23:13:47 #corewars: < Jeff_K> o.k, top ten is a little bit small, but with 25 or 40 it's works. But only for some kind of constants..you understand which i mean, the ones what more "linear".
23:13:55 #corewars: < joonas> yes.
23:14:19 #corewars: < joonas> that's a really interesting way of automatically finding good ranges.
23:15:19 #corewars: < Jeff_K> I hope that it works, but i have serious prioblem with the math behind. But i'm on the way. Another idea is multuistage-runs
23:15:40 #corewars: < joonas> ?
23:15:52 #corewars: < Jeff_K> First you have to define the operators !1(100-200), !2(2000-3000) and so on.
23:16:32 #corewars: < joonas> hm.. so the (100-200
23:16:33 #corewars: < joonas> ) are the ranges
23:16:39 #corewars: < joonas> what are the '!1' and '!2' bits?
23:17:00 #corewars: < Jeff_K> Prioritynumber, let me explain  
23:17:14 #corewars: < Jeff_K> All the operators will used in the first run
23:17:29 #corewars:  * joonas wonders what "runs" are
23:18:55 #corewars: < Jeff_K> where optimax creates 1000 Randrys. After that, he fixed the first operator, to his DNA-range, and all other operator will have the value of the best Randry so far. So youn cann find the best Value for operator 1. Then fix operator 1 and with this number and try it with 2.
23:19:29 #corewars: < joonas> I see.
23:19:43 #corewars: < joonas> and all this fixing of operators happens automatically?
23:20:45 #corewars: < Jeff_K> Yes, should be...i make this because optimax is very fast and it's a pity if you give it 8 hours time and the best is created after 2 hours. So a run can used the whole time better.
23:21:19 #corewars: < joonas> wow
23:21:22 #corewars: < Jeff_K> Are you familar with optimax. Has Fizmo told you what we doing ?
23:21:29 #corewars: < joonas> vaguely
23:21:33 #corewars: < joonas> not in so many words
23:21:57 #corewars: < Jeff_K> one moment, a send you a snippet..
23:22:17 #corewars: < joonas> I got the impresion that it was a program that did mini-hill kind of ranking and automatic randomised constants
23:22:44 #corewars: < joonas> fizmo told me about the many different stages that a candidate warrior has to go through to make it "on the hill"
23:22:48 #corewars: < joonas> that was very impressive
23:23:15 #corewars:  * bvowk is bad
23:23:22 #corewars: < joonas> when he told me about it there were three or four stages?
23:23:27 #corewars:  * joonas smites bvowk
23:23:37 #corewars: < bvowk> smites?!
23:23:44 #corewars: < joonas> smotes?!
23:23:45 #corewars: < Jeff_K> Yes, that is the point..2nd stage is only one warrior, 3 stage is a collection iof 10-12 Warrior..Then in 4 100 Warrior are waiting only for the best of them.
23:23:45 #corewars:  * bvowk is smoten
23:24:02 #corewars:  * bvowk just bought a 15GB ipod
23:24:08 #corewars: < joonas> bvowk: isn't that what angry gods to every other day
23:24:48 #corewars: < joonas> jeff_k: this is what fizmo told me.
23:25:12 #corewars: < joonas> sounds like a good plan to reduce the garbage candidates
23:25:14 #corewars: < Jeff_K> When you define the targetscore good you have a good chance that only the best Randry came to 4. With that you can create and benchmark 4-5 times more Randry as with the old methiod..
23:26:01 #corewars: < joonas> bvowk's got a cool method of automatically adjusting the targetscore so that the percentage that get to stage 4. is always 99%
23:26:04 #corewars: < Jeff_K> Yes, for Stone or so we have Phase 1 that test only that the randry will survive for thet 80000 cycles.. For suicidial you can defione the range..
23:26:15 #corewars: < joonas> well, the equivalent of stage for for ccai
23:26:36 #corewars: < joonas> equivalent of stage 4, I mean.
23:26:50 #corewars: < Jeff_K> We use dynamic target. Target is highscore - 10% ( or 5% )
23:26:59 #corewars:  * joonas hopes bvowk won't sm[io]te him for blabbing secrets
23:28:13 #corewars:  * joonas also hopes 15GB doesn' mean 15 GigaBucks
23:28:20 #corewars: < Jeff_K> ;-)
23:28:44 #corewars: < joonas> what happens to warriors that reach stage 4?
23:29:20 #corewars: < Jeff_K> They will fight against fizmo's fixed strategie hill...100 of the best warriors ever made !
23:30:03 #corewars: < bvowk> eh?
23:30:08 #corewars: < Jeff_K> When the Hill was new, scores of the Randry against fsh were 1:1 to the score for the 94nop.
23:30:32 #corewars: < joonas> heh, need a better fsh then. :)
23:30:32 #corewars: < Jeff_K> nowadays the fsh need some update..an 150er fsh scores 130 on 94nop.
23:30:38 #corewars: < Jeff_K> Yes.
23:30:55 #corewars:  * joonas would love to see the fsh
23:31:32 #corewars: < Jeff_K> Target of the fsh is to create good, but more important is durable warriors..so a fiywed number of warriors for each categorie is important for the durance of the warrior..
23:31:48 #corewars: < Jeff_K> s/fiywed/fixed/
23:32:11 #corewars:  * joonas thinks trying to approximate 94nop isn't good for durability
23:32:44 #corewars: < joonas> a benchmark should include specialists that are very good at killing the kind of warrior you are trying to make
23:32:54 #corewars: < Jeff_K> Hmm, that is what i'm working on these days...want to find indices for a durable warrior...
23:33:02 #corewars: < joonas> it doesn't matter whether those specialist have a good chance of being on 94nop or not.
23:33:59 #corewars: < joonas> because in the end, people will find warriors that are extremely good at killing your warrior, and then you'll be glad you had the specialists in your benchmark, for everyone else's similar warrior will be killed first.
23:34:08 #corewars: < Jeff_K> You can use such a specialist in phase 2..
23:34:46 #corewars: < Jeff_K> If you Warrior is good against all kind of warrior, one specialist can't thorw him form the Hill, ?
23:35:26 #corewars: < joonas> yes, that's true.  but say you have five warriors that are similar.
23:35:48 #corewars: < joonas> the one of those five that is trained against the worst outcome will survive the others
23:36:27 #corewars: < joonas> because the other may get more wins than the trained one, but the trained one will have the right balance between %W vs. %L vs. %T.
23:36:40 #corewars: < joonas> always low %L.
23:36:46 #corewars: < bvowk> joonas: my method is very simple, I test against the hardest warrior for 5 rounds, and set the score high enough that only warriors that will (should) make it on the hill will be tested again..
23:37:11 #corewars: < bvowk> as it goes, we hover around 99.97% cribdeath most times..
23:37:49 #corewars: < joonas> with 5 rounds you're lucky to get any meaningful result at all. IMO.
23:38:12 #corewars: < bvowk> sure you can, you'll know if your warrior is suicidal
23:38:40 #corewars: < Jeff_K> 5 Rounds are some kind of randomscore. 250 rounds is IMHO the min.
23:39:02 #corewars: < joonas> not to be cruel or anything, but if you need to test for suicidal candidate, you're doing it wrong.
23:39:35 #corewars: < joonas> (none of this applies to evolvers though, only optimising partly hand written warriors.)
23:39:46 #corewars: < bvowk> well, better warriors will win those 5 rounds more often than poorer warriors..
23:40:11 #corewars: < bvowk> its quicker to toss the ones that don't perform in those 5 rounds then it is to benchmark a potentially not so great warrior
23:40:43 #corewars: < joonas> I guess with 99.97$ cribdeath, that's got to be true.
23:40:47 #corewars: < joonas> %, even.
23:40:52 #corewars: < bvowk> I can quick test and toss 2000 or so warriors in the time I benchmark 
23:40:54 #corewars: < bvowk> 1
23:41:54 #corewars: < joonas> so you can tolerate 99.95% cribdeath?
23:42:06 #corewars: < joonas> (= 1/2000)
23:42:12 #corewars: < joonas> (= 1-1/2000, I mean.)
23:42:14 #corewars: < bvowk> yeah
23:42:31 #corewars: < joonas> 's all good then. :-D
23:42:58 #corewars: < bvowk> I still make good forward progress most times..
23:43:11 #corewars: < bvowk> I've been finding problems with dead end pools lately..
23:43:24 #corewars: < joonas> how do you mean?=
23:43:34 #corewars: < bvowk> but I think thats more related to my new instruction chaining code instead of the testing regimine..
23:43:50 #corewars: < bvowk> I'll end up with an entire pool that looks something like
23:44:02 #corewars: < bvowk>   17   spl.f  # -794 , $ 3459
23:44:02 #corewars: < bvowk>   18   mov.i  #    1 , <    1
23:44:02 #corewars: < bvowk>   19   djn.b  $   -1 , #-1817
23:44:02 #corewars: < bvowk>   20   djn.b  $   -1 , # -423
23:44:16 #corewars: < bvowk> a lameass imp is another popular one..
23:44:35 #corewars: < bvowk> and it'll stall there for a long long time.
23:46:04 #corewars: < joonas> looks like a mean warrior for an evolved one though.
23:46:06 #corewars: < bvowk> but I think its because my high mutation rate means that eventually the most common instruction combination will take over
23:46:20 #corewars: < bvowk> but that pool never seems to improve at all
23:46:27 #corewars: < bvowk> it'll always be just short of the hill
23:48:06 #corewars: < joonas> hm.. so let me get this straight: before the stochastic mods you never saw these stalls?
23:48:23 #corewars: < bvowk> not that bad
23:48:44 #corewars: < bvowk> but I made the pools deeper and skewed the results on successful mutations..
23:49:10 #corewars: < bvowk> s/pools/chains
23:49:15 #corewars: < bvowk> I made the chains deeper
23:49:24 #corewars: < joonas> ah
23:49:26 #corewars: < joonas> hm
23:49:36 #corewars: < joonas> how deep?
23:49:43 #corewars: < bvowk> and if a mutation is successful it adds wieght to that leg of the chain
23:50:01 #corewars: < bvowk> they're 4 instructions deep now instead of 2 like they were previously
23:50:03 #corewars: < joonas> ah, they're dynamically generated.
23:50:13 #corewars: < bvowk> kinda.
23:50:33 #corewars: < joonas> I was under the impression that you were gathering the stats from 94nop or koen.
23:50:44 #corewars: < bvowk> yes, they start that way
23:51:03 #corewars: < bvowk> and then there is a fudge factor for each leg of the chain now
23:51:21 #corewars: < joonas> tell me about chains and legs
23:51:44 #corewars: < bvowk> basicly its a huge collection of if's
23:52:19 #corewars: < joonas> what kind of conditions?
23:52:21 #corewars: < bvowk> I randomly select a start instruction, then from that instruction, we select an instruction likely to follow that combination
23:53:00 #corewars: < bvowk> if(previous instruction == foo && previous ins -1 == bar && previous ins -2 == baz....
23:53:11 #corewars: < bvowk> back 4 levels 
23:53:11 #corewars: < joonas> I see now.
23:54:39 #corewars: < bvowk> not very exciting
23:54:39 #corewars: < joonas> do you track effective addresses of the insns so they don't break when a warrior is mutated?
23:54:54 #corewars: < bvowk> it makes for really really really big tables of instructions tho
23:55:01 #corewars: < joonas> I can imagine. 
23:55:16 #corewars: < bvowk> it takes about 600MB of memory to compile the client if I use the full 94-nop from K
23:55:34 #corewars: < bvowk> no, I don't track the addresses..
23:55:58 #corewars: < bvowk> I've tried adding those tweaks several times.. but it ends up using up more cpu time than just making piles and piles of them
23:56:05 #corewars: < joonas> I have a feeling that that approach doesn't scale well past stones and papers.
23:56:15 #corewars: < bvowk> not really
23:56:29 #corewars: < bvowk> and its not even very good for those :)
23:56:32 #corewars: < joonas> not really, it does or doesn't scale?
23:56:34 #corewars: < joonas> oh
23:56:57 #corewars: < joonas> still, Skin and its ilk are very cool achievements.
23:56:59 #corewars: < bvowk> I wish my flag idea was less cpu hungry :(
23:57:19 #corewars: < bvowk> because that shows potiential to make really great warriors..
23:57:23 #corewars: < joonas> did you restrict the address range where the flags are taken care of?
23:57:37 #corewars: < joonas> (because if you do, to say 100 cells only, then it's much less cpu hungry.)
23:57:49 #corewars: < bvowk> the flags aren't the problem.. its the optimization logic behind them...
23:57:55 #corewars: < joonas> ah.
23:58:04 #corewars: < bvowk> and taking the warrior between corewars code -> symbols -> back
23:58:25 #corewars: < bvowk> perhaps I just not doing it right..
23:58:32 #corewars: < joonas> well, basically, I think you have to bite the bullet and evolve to the next plane of evolving. :-D
23:58:41 #corewars: < bvowk> heh
23:58:43 #corewars: < joonas> (and find bigger clusters!)
23:58:52 #corewars: < bvowk> yes... find bigger clusters
23:59:02 #corewars: < joonas> what, is that like a matra with you? ;)
23:59:05 #corewars: < joonas> mantra
23:59:14 #corewars: < bvowk> all those stupid chemists and physicists doing *REAL* work..
23:59:17 #corewars: < bvowk> how dare they :)
23:59:19 #corewars: < joonas> lol
23:59:41 #corewars: < joonas> so they finally figured out that the department has some serius HW :)
23:59:45 #corewars: < joonas> poor bvowk
--- Log closed Thu Mar 04 00:00:10 2004