Author: Eugene Nalimov
Date: 21:55:13 04/20/02
Go up one level in this thread
Vincent, please cool up. I am not discussing Intel vs. MS compilers -- I can say a lot, but will remain silent for obvious reasons. I am not saying that mine, yours, or somebody's else EGTB code "per se" is speed-critical. I just noticed that based on what I know *I* would not say the programming in your chess program was "very professionally done". I am glad to hear that you at last do not violate *my* copyright anymore. I kept all the 4 (or 5? -- that's in my office) messages I got from you, and all my replies, and in no one of them you are asking permission to use my code. In some of the messages I explained new ideas you never thought about (your words, not mine). And you don't have to pay that 10% speed penalty in the multithreading programs even on x86. Once again: it took me exactly 15 seconds to figure out solution to your problem after you mentioned it on CCC. Sorry I did not mentioned it earlier, it looked obvious to me, so I forget about it till you write your claims once again. I suggest you a deal -- I explain to you how it can be done, you think about it a little, and if I am right you'll never write on CCC/RGCC the nonsence you are writing regularly, e.g. that Diep has the best possible evaluation in the world, that all but you understand nothing in the (chess) programming, that you are writing flawless professional quality code, etc. I respect you as a chessplayer. It looks that you are good programmer. But your social skills can be improved. Regards, Eugene On April 21, 2002 at 00:24:46, Vincent Diepeveen wrote: >On April 21, 2002 at 00:06:06, Eugene Nalimov wrote: > >>First, some time ago *you* wrote that you have to use 2 copies of EGTB cache and >>decompression tables because you uses processes, not threads. I would not call >>such the design "very professional done". > >You are not nice here. I understand why. You try to distract the people. >the EGTB code has 0% influence onto the running time of the speed >test people. In openings position you don't need EGTBs!!!!!! > >99% of the chessgame the EGTBs are unimportant. A few years ago >when most programs sucked ass in endgames, they were very important. > >Now they aren't!!!! > >They are a thing to have simply. If you have it, it gets less attractive. > >>Second, it took me exactly 15 seconds to find out how you can easily get back >>those 10% of the speed in your program. > >You seem to not know at all what a chess engine is and you can't >program either. > >No way to get the estimated 10% back with threads *anyhow*. > >It's an extra instruction simply to the processor, extra indirection >simply. and that where the poor processors already get stressed >to the limit considering how many instructions they gotta try to fit in >just 64KB L1 code cache. > >Perhaps it's more than 10%, well i never measured. It's simply >EXTRA instructions which you burn. > >Every array lookup you need it. It is sick to just think about >rewriting a program in order to burn more instructions!! > >>Third, of course I don't have access to the source of your chess program, but I >>can bet that your TB indexing code is exact copy of mine -- and any expert will >>say so, too. > >You seem still in dream land. i have option to compile in the code, >the IA64 version i shipped to you has old nalimov code, >but you can't say anything about that, >as far as i know you deleted it because you can't >legally test it because you are under NDA. > >I also have my own EGTBs and it's sharing the cache. > >EGTBs are not so important at tournament play, they were when i >didn't have them (paderborn februari 1999) and diep sucked in endgames, >now if it gets to an EGTB endgame the result is already known by >the engine nearly what the game will get, so the chance EGTBs >influence the outcome of the game has become factor 1000 less >when talking about far endgame. > >Usually i run without them at home. At tournaments i only >turn them on 'in case that they are needed', simply not risking >problems. > >>And forth, I gave permission to use my TB access code to *anyone* who asked. You >>are the only programmer who did not bother her/himself even to ask. > >You did not even reply man, despite that the question >was shipped twice. > >Apart from that i don't mind a legal battle anyhow, because first >the code was given free then later authorship was claimed at it, >very INCONSEQUENT. > >I can proof that you received twice a request you didn't answer too. > >Got witnesses there. > >Further your EGTBs suck so much ass from programming viewpoint that in >a possible commercial version of DIEP i won't support them anyhow. > >It's very bad programmed code which is slow indexing the tables >and it needs a lot of RAM for the indices too. > >Apart from that i stepped away from the believe that DTM is the >way to go. w/d/l is less space and can get compressed way better. > >Yes i DID ask permission to use the compression and got a reply >to that too and the promise made there i will keep to the author. > >Not the EGTB code you wrote is clever, it is the compression >code that is clever. > >You do as if your code is holy. It isn't. Only the compression >code is interesting to use for my EGTBs. I have all 5 men at 1 cdrom, >you need 7.5GB. > >I am amazed you want to discuss EGTBs here now that i post >intel c++ compiler is faster you start about EGTBs. > >You tried another argument discrediting being multiprocessor. >You got it wrong. EGTBs matter SHIT for speed of my program. > >WHO THE FUCK CARES FOR THAT. I had them turned off running the >test of course, because intel c++ can't compile buggy code very >well! > >>Regards, >>Eugene > >>On April 20, 2002 at 23:40:16, Vincent Diepeveen wrote: > >>>On April 20, 2002 at 23:15:59, Robert Hyatt wrote: >>> >>>>On April 20, 2002 at 23:14:30, Kevin Strickland wrote: >>>> >>>>>On April 20, 2002 at 20:19:21, Slater Wold wrote: >>>>> >>>>>>On April 20, 2002 at 15:31:02, Eugene Nalimov wrote: >>>>>> >>>>>>>I would not tell "However to all standards, programming of DIEP is very >>>>>>>professionally done" about the programmer who needlessly duplicates tens of >>>>>>>megabytes of data only because he did not figured out how to efficiently use >>>>>>>threads instead of processes. >>>>>>> >>>>>>>Or about the programmer who included others' code into his program without the >>>>>>>permission. >>>>>>> >>>>>>>Eugene >>>>>> >>>>>>Ouch. I guess it was only a matter of time......... >>>>>> >>>>> >>>>>Also note no reply from Vincent... that really must have hurt. I felt it here. >>>>> >>>>>My only question is what code did he use in his program that he didn't get >>>>>permission? If it was from a program like Crafty and I was Robert I would have a >>>>>hard time letting one compete in tournaments with a program that included even >>>>>one line of my code. >>>>> >>>>>Interesting question but doubtful I would ever get a serious response. >>>>> >>>>>Kevin. >>>> >>>> >>>>Just speculation, but I would suspect Eugene was referring to the egtb.cpp >>>>code... that probes his tablebases... >>> >>>He has no rights to anything. He is simply frustrated guy who >>>can't program! >>> >>>Also he doesn't know shit from threads vs processes. >>> >>>I can easily share the shared memory and start another process. >>> >>>Making it multithreaded means i need to rewrite the entire fucking code, >>>only a megabyte or 2 and also get 10% slower.
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.