[09:05] Join: anonymous joined #corewars [09:05] brx? [09:05] Part: anonymous left #corewars [09:50] Join: Boo joined #corewars [11:16] Join: fiveop joined #corewars [11:50] MSG: Quit: leaving [12:02] Join: Boo joined #corewars [12:52] MSG: Quit: leaving [13:15] Join: Fluffy joined #corewars [13:25] I hate compilers. How can it expected me to have more than 1 GiB of RAM :-( [13:25] *exptect [13:25] *expect [14:00] Join: Roy joined #corewars [14:00] * Roy tests [14:02] Hi Roy [14:02] You are testing what? [14:10] If IRC still works, we got a new server and firewall today (moving the company) [14:11] And it gave a security warning when starting mIRC, MS said: "This program is being blocked, to unblock talk to the administrator" [14:11] But it still works anyway... :) [14:12] hmm [14:14] I think, today is quite a productive day. I've produced more than 185000 lines of code today [14:15] Unfortunately I won't know in the near future, whether it really compiles :( [14:18] Hehe.. [14:18] Tomorrow I might be able to check the WinPython version of PyCorewar [14:19] :) [14:47] But some pc's have more then a Gb of RAM... fast mars ;-) [14:50] Join: sascha joined #corewars [14:52] Fluffy: did you generate the code? :P [14:57] fiveop: yes [14:59] MSG: [15:00] Roy: I know, but I only have 768 MiB of RAM and GCC needs about 1.2 GiB to compile the generated code [15:01] MSG: Ping timeout: 252 seconds [15:01] swap? [15:02] I've never needed it during the last years. That's why I only have 512 MiB of swap [15:02] GCC still says, that this is not enough [15:02] ...... [15:02] I've got 1,5 GiB swap [15:02] * Fluffy shrugs [15:03] Mem: 775048k total, 765436k used, 9612k free, 38108k buffers [15:03] Swap: 1461900k total, 4k used, 1461896k free, 502880k cached [15:06] And how often it happens, that you really need that much swap space? [15:08] hm [15:09] I guess it didn't even happen while I was playing a 3d game with a huge amount of textures [15:09] but [15:09] you never know and wth should I do with all that space on my harddrive otherwise? :P [15:16] Hmm thats true.. [15:30] MSG: Quit: leaving [15:38] Join: Fluffy joined #corewars [16:05] MSG: [16:09] Join: anonymous joined #corewars [16:11] hello [16:11] Part: Fluffy left #corewars [16:12] You be brx for me [16:12] please all be brx [16:12] Wow! [16:12] Part: anonymous left #corewars [16:13] Join: Fluffy joined #corewars [16:36] Join: anonymous joined #corewars [16:36] Gosh! [16:37] I miss brx [16:37] too you miss brx [16:37] hi [16:37] :) [16:39] Part: anonymous left #corewars [17:12] Join: sascha joined #corewars [17:21] Join: bvowk joined #corewars [17:22] Hi bvowk :) [17:22] * bvowk uses a remote root exploit on fluffys machine and then creates files with insulting names in the home directories.. [17:22] bvowk: go on [17:25] yeah? :) [17:25] I'm only saying it because it seems like the thing to do, I cleaned up someones unix box.. [17:25] hmm [17:25] * Fluffy hates his c-compiler [17:26] I guess when they got onto it and realized it was a 10 year old machine that had nothing but some math papers on it they were pretty disgusted and left a bunch of nasty filenames :) [17:26] yeah? [17:26] my compiler refuses to compile my code [17:26] It wants to have more memory [17:27] which I don't have [17:27] it seems to want more? [17:27] what on earth are you compiling that a modern box doesn't have enough ram? [17:28] the only times I've run into that is when I was compiling computer generated C files of doo [17:28] m [17:28] a big switch statement with about 6500 cases [17:28] thats it?! [17:28] exactly what I've done [17:28] I had one with a few million cases and I only needed 1G [17:28] turn your optimiazation level down [17:28] done [17:28] er.. [17:28] still wanted 1 GiB of ram [17:28] cry? [17:29] done, too [17:29] can you add a mother huge swapfile? [17:29] what compiler are you using? [17:30] gcc 4.0.2 [17:30] hrm. [17:30] I've generated a big switch-statement for opcode/modifier/amode/bmode [17:30] :) [17:30] with Python [17:30] I'm so proud *sniffle* [17:30] I wanted to make sure, that this is too much even for modern computers [17:31] how much code for each case? [17:31] how did you generatethe code for each case? [17:31] some lines, all all 185000 lines, less than 4 MiB of code [17:31] because you might be able to squeak through if you make a separate file of a pile of inline functions [17:31] fiveop: wrote a little python script [17:32] it'll end up the same in the end, however, you'll have broken it down to little bitesized bits for the compiler [17:32] won't need to alloc quite so much [17:32] bvowk: will try that tomorrow [17:32] you could prolly get away with putting them in the same file actually.. as long as they were separate function calls. [17:33] in any case PyCorwar is at the moment way faster than exhaust [17:34] almost as fast as exhaust-ma [17:34] and I still have only added more special cases for the most frequent insns [17:35] Hmm i don't think that the massive use of inline function will solve Fluffy's memory-probs.. [17:35] sascha: At least I can try :) [17:36] Engineering by probiering ;-) [17:36] hey! [17:36] very nice fluffy! [17:37] Yesterday I was quite sure, that the generate code would be more than 1 MiB [17:37] sascha: it'll help, the problem is that right now its attempting to work with all the code in a single go during the initial compilation.. [17:37] but today I'm no longer that sure [17:37] if you break it out, it'll compile each function separately, and the switch will be manageable.. [17:37] my current estimation is about 280 kiB, which current processors should be able to handle quite well [17:37] at link time it'll stitch it all back together again, but using them as little blocks instead of huge swaths of code [17:38] that might prevent some optimisation! ;) [17:38] it might, but it'll allow compilation! :) [17:39] fiveop: I'm already unrolling the loop the loop/switch as much as possible, because the compiler can't optimize it properly [17:39] fiveop: So it shouldn't be an issu [17:39] e [17:39] you could also try nesting the switches [17:39] ie, if your opcode is between foo and bar, you're passed off to another switch function for final hackery [17:40] if you broke it up into 4 nested switches it might not be so bad.. [17:40] you might want to put all the high volume instructions under the first branch.. [17:40] as then your branch prediction will be right more often [17:41] bvowk: At the moment I'm trying exactly that [17:42] willvarfas list of opcode freqs has 272 entries [17:42] it should be enough to handle theses insn first [17:42] and then the rest "normally" [17:42] in two or three switches [17:43] hrm. [17:43] 1st switch: top 50 % of all insns. [17:43] so you've got nested switches? are they all in the same switch body, if so you might want to just bust them out to separate functions. [17:43] 2nd switch: whatever [17:44] 3rd switch: the stupid rest (div/mod/...) [17:44] bvowk: No, I'm planning for 3 switches in a row [17:45] instead of the usual "case: ... break;" I use "case: ... continue;" [17:46] hrm [17:46] try gcc 3 :) [17:46] yes, I know [17:46] it use significantly less memory for compilation of big stuff [17:46] (I just tried w/ 4.0 and 3.4.4 [17:46] when I switched from gcc 3.3 to gcc 4.0 the performance of the 88-MARS fell from 9.3 to 8.2 MIPS [17:47] of course, I have 2GB of ram [17:47] that sucks. [17:48] As soon as PyCorewar faster than exmars/exhaust I'll be content enough [17:48] and make it version 0.2.0 [17:48] the I have to care about the remaining problems with the parser [17:49] which shouldn't take much time [17:49] joonas wrote a wc replacement in redforth last night using the new version of exmars-streams :) [17:49] eek [17:49] he was twiddling with making a bignum library so he could calculate big primes :) [17:49] You've mentioned his name [17:49] scary huh? [17:49] eek. [17:50] * bvowk scurries away to hide [17:50] Join: Nenad joined #corewars [17:50] Hi Nenad [17:50] Hi! [17:52] hi nenad [17:58] 26872 [17:58] opps [17:58] oops even [17:58] ? [17:59] no, it is not a prime [17:59] thats the number of 94nop warriors I've saved [17:59] nenad was asking [18:02] where did you get than many? [18:02] *that [18:02] I made them :P [18:02] are they any good? [18:02] I've got like 150,000 nano warriors too.. [18:02] some of them.. thats where MWR came from [18:03] hmm [18:03] Join: sf joined #corewars [18:03] Hello [18:03] Hi sf! [18:03] hi sf [18:03] hi sf [18:04] back in a minute [18:04] going to change m keboard [18:05] MSG: Client Quit [18:07] Join: sf joined #corewars [18:07] lisp? [18:07] MSG: Client Quit [18:07] hey SF... how's that bridge working out for you? [18:07] you have a lisp? [18:07] :) [18:07] Join: sf joined #corewars [18:09] bvowk: It should be spelled "lithp" [18:11] I'm still waiting for the bridge to be delivered [18:12] Can I pay cash on delivery? [18:13] MSG: Read error: Connection reset by peer [18:15] Join: sf joined #corewars [18:15] There are some stranges ads next to r.g.cw on google groups today [18:20] Join: Mizcu joined #corewars [18:20] Hi Mizcu [18:20] h [18:21] Hi Mizcu [18:21] o.k maybe bbl [18:21] * sascha waves [18:21] * sf waves [18:21] Part: sascha left #corewars [18:22] * Nenad waves [18:40] * Fluffy waves [18:40] hmm ... stupid me [18:40] * sf waves [18:40] He left a while ago [18:40] :) I thought that you were leaving [18:40] Oh, I thought you was going [18:40] no, not yet [18:40] :) [18:40] hrm. [18:40] doing anything interesting sf? [18:41] Eating lockets, drinking Cherry Coke and writing for our project [18:41] lockets? [18:41] are you a metalivore? [18:41] polymorphed into a xorn are we? [18:42] Yes, I eat metal [18:42] * bvowk used to drink coke.. but I gave it up and lost like 70lbs a few years ago.. [18:42] now I dont like the taste [18:42] It's popular in civilised coutries [18:42] bvowk: What's 70 lbs in kilogramm? [18:43] and now that much sugar (as my sugar consumption is now nearly 0) makes me wildly hyperactive [18:43] somewhere about 30kg [18:43] Fluffy: multiply by 0.454 [18:43] Join: Fizmo joined #corewars [18:44] hi [18:44] sf: thanks [18:44] hi Fizmo [18:44] hi Jens [18:44] Hi Fizmo [18:44] sent an update of LJ T on 94t [18:45] should score much better ;-) [18:45] MSG: Read error: Connection reset by peer [18:45] Nenad, you'll get soon some competition on 94t ;-) [18:45] hehehehe [18:46] Join: sf joined #corewars [18:46] wb sf [18:46] Lockets are a popular honey and lemon medicated sweet, which make cherry coke taste strange [18:47] interesting.. [18:48] You probably have them there with a different name [18:48] We have a twix chocolate bar. [18:48] we do too.. [18:48] bletch [18:48] In europe they are called Raider [18:48] raider? [18:49] aren't you in europe btw? :) [18:49] Yes, same colours for the logo and everything, different name [18:49] * bvowk ducks [18:49] We used to have something called Opal Fruits [18:49] Those were renamed Starburst so we match the rest of Europe :( [18:50] sf, not any more [18:50] raider, that is [18:50] hrmm, here it is called twix [18:50] Oh? [18:50] twix here, as well [18:50] they changed the name several years ago [18:51] Oh [18:51] http://www.uni-koeln.de/studenten/al/info/47/0503.html [18:51] maybe because uk people are very conservative ;-) [18:52] It's called Twix. We called it Twix. You should call it Twix too :P [18:52] nice score, Fizmo. [18:52] Where? [18:52] 94t [18:52] Join: DrewHK joined #corewars [18:52] hello [18:53] Hi, DrewHK [18:53] oh, so quickly [18:53] wooow :-) [18:54] twix likely means something naughty in some common european language :) [18:54] almost 15 points better [18:54] optimaxed? [18:54] just crudely [18:54] I should probably do that, too... I've got a nice 94t benchmark :) [18:55] however, haven't optimized any warriors yet [18:55] have just a quickhacked fsh [18:55] 94nop with changed imp constants [18:55] and just some few scanner optimized for 94t [18:56] I've got a combination of that and all of my good warriors - those that I have and those that I haven't sent to the hill [18:56] cool [18:56] Join: Twix joined #corewars [18:56] Raider is a goner!! [18:56] Part: Twix left #corewars [18:57] Hi Roy [18:57] looool [18:57] Nenad: I should do a 94t infinite hill [18:58] :) i would like that [18:58] Join: Roy joined #corewars [18:58] Hi all :) [18:58] as long as no warriors are available for a hill new player will hardly try it [18:58] Hi Twit [18:58] Hi Roy [18:58] Sorry, Twix :) [18:58] Hehe twat.. [18:58] Hi Roytwix [19:01] Fizmo: when is the next rf round going to be organized [19:01] ? [19:01] and who is organizing it? [19:03] * sf votes for a round which requires bvowk's new opcodes [19:04] ummmm [19:05] everybody is very welcome who would like to organize next round [19:05] :) [19:05] I thought that it was already assigned [19:05] just send me a note that you are willing to do a round [19:06] well [19:06] hmmm [19:06] I am interested. [19:06] if the others agree to it [19:07] I can do one but not the next one [19:07] Do you have an idea Nenad? [19:07] I can get one soon :) [19:07] Chris's site has a page of ideas [19:07] yes, you're right. There where someone who wanted to do the next round, but I think we both lost a bit the view on it after CSEC2005 [19:08] sf: yes, but it needs urgently an update because most of the ideas were already used [19:10] Fizmo: so, who is it? [19:11] the next organizer? [19:12] MSG: Ping timeout: 252 seconds [19:15] Join: sf joined #corewars [19:15] My webpage needs an update too :( [19:15] That one who give me a note first ;-) [19:15] * Nenad gives Fizmo a note [19:17] Chris: this is for you http://www.themusicdepot.com/acatalog/Product_55295-N_lg.jpg [19:17] * Roy returns.. [19:18] talking about www.ibm.com/software/lotus ? :-) [19:18] I just heard the most amazing story btw in the subway.. [19:19] Nenad: deal :-) [19:19] sf: thanks :-) [19:19] :) [19:20] Roy: what story? [19:20] Somebody was telling that their company is trying to brain-control wasps with electric stimulans in the directional-organ, so they could move the wasp around in any way, remote controlled... and they wanted to also put a tiny camera on the wasps so they could 'sell' it.. [19:21] What is the lifespan of a wasp? [19:21] Noo idea... [19:21] What makes a wasp the most suitable insect? [19:21] that is true, not a story [19:21] But it sounded plausable.. [19:21] Bit freakish though :-D [19:22] http://www.newscientist.com/data/images/ns/9999/radiocontrolledhuman.mpg [19:22] they have already succeeded somewhat controlling mouses [19:22] If they can do it with humans.. [19:22] http://www.newscientist.com/article.ns?id=dn7829 [19:23] hmm, a lot of wasps, with little radios, character recognition programs, and i never need to search for my books [19:23] MSG: [19:24] Radio controlled spy wasps... Al Qauda better invest in flytraps [19:24] or birds [19:24] or viruses that attack insects [19:25] wasp-flue :) [19:25] birds capable of digesting camera parts :) [19:25] Or just humans, easiest :P [19:25] even better, birds that integrate those cameras into their system after they eat the wasps [19:25] MSG: Ping timeout: 252 seconds [19:28] Join: sf joined #corewars [19:30] * Fizmo Bot announce a website update of http://www.corewar.info/discuss.htm [19:32] * Roy Bot announces his departure [19:32] * Roy Bot waves [19:32] * Nenad waves to Roy [19:32] * Roy waves back [19:33] MSG: Quit: Dammit the spy-wasp spotted me! [19:33] MSG: Read error: Connection reset by peer [19:34] Join: sf joined #corewars [19:34] MSG: Remote host closed the connection [19:35] * Fizmo waves too late [19:36] Join: sf joined #corewars [19:36] MSG: Remote host closed the connection [19:37] Join: sf joined #corewars [19:37] MSG: Remote host closed the connection [19:40] * Fizmo Bot announce a website update of http://www.corewar.info/ihp.htm [19:42] Join: fatalc joined #corewars [19:43] wb [19:43] Hi fatalc [19:43] Hi Nenad [19:43] :) [19:46] hrm.. fatalc [19:46] Did I ask you about webpage statistics? [19:46] Am I still here? [19:46] ? [19:46] I dunno, are you? [19:47] Please someone say something, am I still here? [19:48] yes, u r [19:49] * fatalc votes out [19:50] * fatalc votes Fizmo out [19:50] Argh, this is worse than bersirc :( [19:51] Join: sf joined #corewars [19:51] MSG: Quit: Physical Wreck IRC webscript: http://www.pswclan.com/cgi-bin/irc/irc.cgi [EOF] [19:56] I've got to go now... See you all later! Bye! [19:56] * Nenad waves [19:56] * Fizmo waves [19:56] MSG: [19:57] * sf tries to wave, but doesn't know if it will work :( [19:57] I'm sorry fizmo, but you've been voted out of the tribe :) [19:57] you'll need to leave the channel immediately :) [19:59] Join: sf[1] joined #corewars [19:59] MSG: Read error: Connection reset by peer [20:00] I would like to change my vote [20:00] ok, time to leave now [20:00] * Fizmo waves [20:03] MSG: [20:03] Join: sf[2] joined #corewars [20:03] * bvowk was only kidding! [20:03] MSG: Read error: Connection reset by peer [20:04] I was just in the middle of a conversation with Fizmo :( [20:04] Stupid connection [20:04] shouldn'ta voted him out then [20:05] I tried to change my vote :( [20:05] I didn't realise what the vote button does :( [20:16] It isn't possible to have a faster than C clear [20:16] If it is faster than C it is a stone [20:16] Because it misses bits [20:19] MSG: Quit: "Help! I've been g:lined from my mIRC!!" Bersirc 2.2: less n00bs [ http://www.bersirc.org/ - Open Source IRC ] [20:21] Join: sascha joined #corewars [20:29] MSG: Read error: Connection reset by peer [20:31] Join: johnkw joined #corewars [20:31] my simulation was just as exact, that I do not now my results are good or not. [20:31] damn [20:31] bye all [20:32] MSG: [21:03] Most recent verson of PyCorewar can be found at http://corewars.jgutzeit.de/pycorewar/download/pycorewar-0.2.0.dev-20060315.tar.bz2 [21:03] *version [21:15] Join: Fluffy_ joined #corewars [21:18] MSG: Ping timeout: 252 seconds [21:25] see ya [21:33] Join: John joined #corewars [21:33] Hi [21:34] Hello John [21:35] Nice scores on 94nop Sascha :-) [21:35] Thx..after dat we all score better ;-) [21:36] Now I need to make something to get back on [21:36] Before I go away for the weekend [21:55] Join: sf joined #corewars [21:55] MSG: Remote host closed the connection [21:55] Join: sf joined #corewars [21:55] Hello [21:56] Neo: working now [22:00] Part: sf left #corewars [22:01] MSG: Quit: humhum [22:19] hey met? [22:20] www.math.ualberta.ca/~bvowk/files/exmars_streams.exe [22:20] its got working -c 0, -r 0 doesn't work, and some cleanups.. [22:20] should be better [22:22] Hi Bvowk [22:22] Great thanks :-) [22:22] Back in a mo [22:24] HWSNBM was playing with redforth and it last night.. [22:24] I still have to make a few more changes.. like warrior length and coresize should be stupidly expandable.. [22:25] and I'm thinking maybe we could make a switch to bump the max coresize up to 4B [22:31] 4B? [22:31] yeah.. 4 billion cells [22:31] Oh! [22:32] I can pass command line parameters to exmars using rounds [22:33] exmars optima.red -r 20 displays the top 20 optima numbers for the coresize for example. (it displays 1 per round) [22:33] thats cool [22:34] I'm still working on some programs [22:34] But I'm away for 5 days, busy preparing now :-( [22:36] I'm leaving shortly for a few weeks myself. [22:36] What are you up to bvowk? [22:36] MSG: Quit: leaving [22:37] * sascha waves [22:37] * John waves [22:37] Part: sascha left #corewars [22:40] You should turn on the nano evolver before you go bvowk ;-) [22:42] I'm hoping to have my 94nop evolver running before I leave [22:43] Auto submit? [22:46] likely not [22:47] I don't think I'll have enough time to a) get something good out b) test everything to make sure it doesn't suck first [22:47] Join: DrewHK joined #corewars [22:48] Hi Drew [22:50] hi [22:50] have you seen brx recently? [22:52] Someone found my webpage by searching for bvowk today! :-) [22:52] I think I saw him in yesterday's logs [22:52]