Download at: http://www.robocoderepository.com/BotDetail.jsp?id=2256
SilverFist is #1 with a rating of 2073! Congratulations PEZ and Axe! --David Alves
Thanks a lot David!! Just returned from a family dinner, where we was comemorating the buying of our (mine and my brothers) first boat, and when i arrived here... I am now comemorating our (mine and PEZ) first bot! Life can have funny coincidences, donīt it? U can imagine the size of the smile in my face... Even my best expectations didnīt prepared me to that. Sorry if my words sounds too much enthusiastics, but I FEEL GOOD!!!! -- Axe
I even saw it doing 2078. :-)
Congrats Axe and PEZ!-- Jonathan
Thanks! :) Iīm wordless! And winefull!!! -- Axe
This means that SilverSurfing is better than any other kind of surfing. Something I've suspected quite a while now. Now it's time you put together that page about the way you do surfing Axe. But... please wait until you've sobered up from all that wine. =) -- PEZ
Hic! :) Iīll pay my debts! See Axe/ToDoList...
This kind of words that iīm about to say might sound somewhat pedant (if they sound that way Iīll blame the wine!:)), but I wanna say these words, so there they go(hic!):
PS: Another good thing about SF is that it might wake-up ABC, since now he is challeged... -- Axe
Congratulations! Great score, just the wake-up call I needed ;). I wonder how many points it would score without the pre-saved data... -- ABC
Thanks. Glad u are back! But about pre-saved data... no difference i think. That data u saw, is from SSīs movement, with 17 bots in pre-saved (1 of them is WSCBotA, and almost all others are HOTs, and against them that data wouldnīt make any diff, i assure u). Actually that pre-saved had gone in package because I was rushing to a dinner with my family (also i forgot to set SFīs colors to yellow and blue)... Actually SS have an old and un-solved bug: it needs at least 1 bot data in the data dir, otherelse it crashes unable to create the dir... But u had raise an interesting point: How would it perform with real pre-saved data? ;) -- Axe
Ah, ok, I thought you had used the new BeeWT gun. Even greater acomplishment then. I'll see what I can do with Shadow, I just noticed I didn't even have the compiler configured since my lastest hardware upgrade, now I have some catching-up to do... :) -- ABC
Now I also understood your question...
I was surprised with it... Maybe u remember, it was u that convinced me to not use pre-saved... and the testings made that time shown not a big diff. after all (iirc about 5 points)... -- Axe
Just saw it for the first time... wow, that's a good bot. -- Alcatraz
ABC, so you thought much of this was due to WikiTargeting data in SilverFist? Just hold your horses. A SilverFist WT will arrive soon, at a repository near you. =) -- PEZ
Yep, but I also already knew your "normal" gun is amazingly good too, those TC results don't lie... Can't wait to see where SilverFist WT will end up, that'll be a tough challenge, trying to dethrone it without pre-saved data. ;) -- ABC
Yeah. Though the deal it that Axe and I will remove SilverFist once one of our own bots is the main challenger. Or any another bot dethrones it. Whichever happens first. And, you can always preload your bot with gun data. Or maybe you can't? =) Vic is a genius! -- PEZ
My gun would need around 1MB per opponent... :) -- ABC
Genius? moi? recognition! finally! (jk =^) ... anyway, this WT compression scheme only works for GF guns, I'm afraid. I have basically ruined my own chances with my SOOL gun. That too can't be compressed because you can't identify the SuperNodes in a log. At least, I don't think so. Fortunately my next gun (the binary tree gun) will be able to use saved data. Only it will save segmentation data and not real GF data. --Vic
SilverFist WT released! -- PEZ
I got this error, not sure if it's SilverFist's fault, RoboRumble@Home's fault, or Robocode's fault. :-p --David Alves
Executing battles ... Fighting battle 0 ... wiki.SilverFist 1.1WT,kvk.HebusLeTroll 0.41 RESULT = wiki.SilverFist 1.1WT wins 4105 to 879 Fighting battle 1 ... davidalves.net.YALT 1.62,axeBots.Musashi 2.18 Preventing unknown thread wiki.SilverFist 1.1WT from access: (java.io.FilePermis sion C:\robocode\rrah\.robotcache\wiki.SilverFist_1.1WT.jar_\wiki\SilverFist.dat a read) RESULT = axeBots.Musashi 2.18 wins 3060 to 2084 Fighting battle 2 ... jekl.DarkHallow .80,axeBots.Okami 1.04 Preventing unknown thread wiki.SilverFist 1.1WT from access: (java.io.FilePermis sion C:\robocode\rrah\.robotcache\wiki.SilverFist_1.1WT.jar_\wiki\SilverFist.dat a read) RESULT = jekl.DarkHallow .80 wins 3120 to 2355 Fighting battle 3 ... davidalves.net.YALT 1.62,pe.SandboxLump 1.52 RESULT = davidalves.net.YALT 1.62 wins 3199 to 2301 Fighting battle 4 ... wiki.SilverFist 1.1WT,el33t.EL33tGangstarr2 2.0
I get messages like that too for many bots. But this is a bit different. Seems like SFWT threads are hanging around when other bots fight. Strange. But I don't think it impacts results. Or, hope not, that is. -- PEZ
Actually I just saw the above behaviour with messages about SFWT threads... Maybe Axe and / or Vic can figure out what's going on. -- PEZ
Are you using multiple threads? If so you're supposed to stop them all at the end of each round... --David Alves
Well, I am certainly not doing that. But maybe Axe is. Even though I doubt it. -- PEZ
Looking at some statistics it doesn't do any multithreading. I have no trouble with it yet... -- Jonathan
It happens on my desktop, but not on my laptop. (Both are running RR@H now) Desktop has java version 1.4.2_05-b04. I am using -Dsun.io.useCanonCaches=false, so it's not that error. Also, it seems to slow down my computer a lot... Normally my desktop turns in battles much faster than my laptop, but it's only running about half as fast right now. Neither computer is running anything else. --David Alves
I got the problem too now! There is really one thread too much! -- Jonathan
Yep, it's happening on my laptop too now. PEZ! Axe! What have you done to my computers?! :-) --David Alves
This is just wrong... each battle a new thread is spawned. SilverFist doesn't need to participate in that battle, yet even tries to load data from a bot it shouldn't know of! (but not always)
Fighting battle 9 ... mld.Moebius 2.0,davidalves.net.YALT 1.62 Preventing unknown thread wiki.SilverFist 1.1WT from access: (java.io.FilePermission /Users/jonathan/roborumble/.robotcache/wiki.SilverFist_1.1WT.jar_/wiki/SilverFist.data/mld.Moebius 2.0.gfs read)I have currently 6 threads too much, and it's running really slow. -- Jonathan
It crashed after SilverFist spawned yet another thread after some slooooow non-SF battles. Axe/PEZ? -- Jonathan
I vote for moving it to the pasteboard until you figure out what's wrong and fix it... But then again I'm biased because I want more battles for my bot. It's been on the rumble for 4 hours and still only has 300 battles.... --David Alves
Sorry guys, it crashed my computer and throws out of memory errors when running. Three votes seems like enough to me. I moved it to the paste board. -- jim
Just arrived at home... 1:44 am... Seems that SilverFist itīs winning in first round by Knock-Down! Cool, thatīs a powerfull punch indeed! :) Just kiddin, iīll try to take a look... -- Axe
Saw it... My fault it seems... Sorry people, in two hours a new version will be released... @PEZ: U are probably sleeping, but anyway: Iīll correct it, ok? My fault. -- Axe
Problems in packaging, nevermind, itīs fixed... (i think:)) -- Axe
Just out of curiousity, what was causing the error? --David Alves
I like the fact that SilverFist WT has only lost to two bots. CassiusClay and SilverSurfer. =) -- PEZ
The cause of the problem was a wrongly packed .jar, all the .java was ok, but apperently some of the .class werenīt recompiled... Just recompiled and re-packed.
Disapointing gain, eh? Strange... -- Axe
How do you mean a dissapointing gain? -- PEZ
![[Home]](/images/RoboWiki.png)