[ View Thread ] [ Post Response ] [ Return to Index ] [ Read Prev Msg ] [ Read Next Msg ]

BGonline.org Forums

The Michael Depreli Bot Comparisons - Anyone want to do the speed study?

Posted By: eXtreme Gammon
Date: Tuesday, 15 February 2011, at 5:03 a.m.

In Response To: The Michael Depreli Bot Comparisons (Nack Ballard)

As Keith pointed the speed graph come my own testing. All speeds are based on the analyze of 2 sessions (one money session and a 25 point match). One exception is BG-blitz where the speed was determined based on rollout speed compare to XG's

First of all it is important to point that it depends greatly of the processor used. Snowie is single threaded, this means it can only use one core of your processor, while all the other programs will use all the cores/thread available. I used my core i7 920 for that analyze which is 4 cores and 8 threads.

Now the speed of analyze is one thing and speed of Rollout is another. I think the number you were referring are about rollouts. For rollout, yes, XG edge seems to be more than on a plain analyze of a match. This is due to the inner working of each program which details will probably be boring to most users and are, anyway, not something I disclose.

I am actually surprised no one took over that speed analyze, I expected when I put that chart together that someone will redo the test for analyze speed (and maybe for Rollout speed). I do not have time to spare to do that. Also I'd rather have this test made by someone else but me as it carries more weight coming from another source than an involved party.

So, is anyone interested to do such speed study?

I would say the things to test are:

  • analyze speed in a specific level (the same than what i did)
  • rollout speed in 1-ply (0-ply for Gnubg)
  • rollout speed in 3-ply (2-ply for Gnubg)

    The type of processor to test are:

  • one core (example Pentium 4)
  • two core (example Core 2 duo)
  • four core (example Core 2 Quad)
  • four core with hyper threading (example Core i7)

    I think all other configuration can be derived from these tests.

    This should be much faster to do than the tremendous effort that was needed for the Depreli study (both in term in rollout and the compiling of the results).

    Any volunteers?

  • Messages In This Thread

     

    Post Response

    Your Name:
    Your E-Mail Address:
    Subject:
    Message:

    If necessary, enter your password below:

    Password:

     

     

    [ View Thread ] [ Post Response ] [ Return to Index ] [ Read Prev Msg ] [ Read Next Msg ]

    BGonline.org Forums is maintained by Stick with WebBBS 5.12.