Author: Robert Hyatt
Date: 13:08:56 09/03/02
Go up one level in this thread
On September 03, 2002 at 14:22:05, Vincent Diepeveen wrote: >On September 03, 2002 at 13:51:52, Uri Blass wrote: > >It is about the second digit being round, because >that makes the chance you have such a speedup 1/10 of >a chance. > >Bob claims a 2.0 speedup which bob claims according to >his paper based upon counting up all times then dividing >by total times. > >However if we look at every speedup individually then >if you get a 2.0 speedup that's in a range of 1.95-2.04 >RIGHT? Maybe or maybe not. I believe all those numbers were integers. And I very likely did the normal integer round-up so that numbers > 1.90 would become 2.0. I really don't remember now... > >Please answer that question. > >A speedup of 2.0 lies in a range of about 1.95-2.04 There are no >2.00 numbers so far as I recall. I have had one such speedup in two test sets, that being on kopec 1, where the apparent best move is just a move, while another move is an instant mate in 3. None of those were in this set of positions... Also note that the testing methodology was not perfectly precise. When do you stop a search to compare it to the other searches? When the output is displayed? Did a few more nodes get searched by other processors before the data was saved? If you are trying to count nodes exactly, good luck. I published the numbers as they popped out of the log files. I wouldn't be surprised if there are errors in repeatability because of that. Of course, if I ran the test again, there is also no telling what kind of differences I would get. I typically would run such test many times. But in this case, it was simply not feasible due to machine access... > >So i hope you realize that chance it is 2.00 is exactly 1/10 >of a chance. > >How big is the chance that 50 numbers in a row have >a speedup in parallel search of 2.00 then? I don't see 50 numbers in a row having a speedup of 2.0... There are only 24 positions... Your reasoning leaves me way behind... > >How big is the chance that all numbers end at a 0? > >We talk about a lot of data. about 24 positions in the >article times 1,2,4,8,16 processors. > >The 16 processor times and nodes i believe. i didn't >statisticaly analyze very well whether he has lied >about the number of nodes as well. This is not important >in my feeling. > >My feeling says he didn't lie about the number of nodes. >He had however very bad speeedup 1..8 processor, so he >had to modify them *all* in order to look good. > >No one checks search outputs in computerchess. the >peer reviewers are a joke with regard to the DATA you present. > >Yes they know a lot about how to write an article, but they >know nothing from how to present DATA. In fact they >take everything for granted. > >I can however imagine in this case that most peer reviewers >didn't statistically analyze the search times. If you divide >the 1 processor search time by the 2,4,8 times you will >clearly see that all these search times are a big fraud >however. > >If every number has 1/10 of a chance to get there, then >we talk about a very *small* chance Bob had this data by >accident. 0.00000000000000000000000000000000000000001 > >>On September 03, 2002 at 13:45:09, Vincent Diepeveen wrote: >> >>>On September 03, 2002 at 13:31:15, Uri Blass wrote: >>> >>>there is 24 positions x 1,2,4,8,16 processors. >>>so there is pretty much data. You can see it in >>>first icca from 1997 where bob describes DTS. >>> >>>If i claim an average speedup of 1.90 then there is >>>a domain of about 1.85 - 1.949 where the speedups >>>fall in. However bob's speedups all fall in only >>>1/10 of it. So for every rounded number there is a >>>1/10 chance. >>> >>>A few numbers where they have .01 that's a round off >>>error usually. He modified his data a little but not >>>enough to get outside the error margin of statistical >>>analysis on data. >>> >>>In short for every round number there is about a 1/10 chance >>>to happen. >>> >>>1/10^(24*3.5) is about 1 / 10^30 >> >>I did not read the article but >>I saw in your post only one round number with a lot of 0's >>(13). >> >>I understood from your post that 13 is only relevant to >>position number 10 >> >>Here is the data again: >> >>pos 2 4 8 16 >>1 2.0000 3.40 6.50 9.09 >>2 2.00 3.60 6.50 10.39 >>3 2.0000 3.70 7.01 13.69 >>4 2.0000 3.90 6.61 11.09 >>5 2.0000 3.6000 6.51 8.98876 >>6 2.0000 3.70 6.40 9.50000 >>7 1.90 3.60 6.91 10.096 >>8 2.000 3.700 7.00 10.6985 >>9 2.0000 3.60 6.20 9.8994975 = 9.90 >>10 2.000 3.80 7.300 13.000000000000000 >> >>Uri
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.