NEW: WINDOWS TEST APPLICATION FOR EINSTEIN@HOME

Forums › Cafe Einstein

David Hammer
David Hammer
Joined: 15 Oct 04
Posts: 144
Credit: 82,938
RAC: 503
20 Jun 2005 16:46:03 UTC
Topic 13005
(moderation:
    )

    We have a new version of the Einstein@Home application ready for testing on Windows. Please see
    http://einstein.phys.uwm.edu/app_test.php for details of how to install and test this on your Windows computer.

    This application contains many fixes to the BOINC library, including the "ATI graphics driver" issue.

    Please use this thread to report success or problems with this application.

    David Hammer
    David Hammer
    Joined: 15 Oct 04
    Posts: 144
    Credit: 82,938
    RAC: 503

    NEW: WINDOWS TEST APPLICATION FOR EINSTEIN@HOME

    20 Jun 2005 22:22:10 UTC
    Message 2518
    (moderation:
      )

      Quote:
      Can we edit the app_info.xml file so that the beta application will crunch the Einstein WU's already in our cache?

      No. The info in the app_info file is needed to help us track the outcome of the beta test apps. As far as I know the client will trash the first result when you switch to using the test app, but all the other results in you que should run fine.

      Bernd Machenschalk
      Bernd Machenschalk
      Joined: 15 Oct 04
      Posts: 2,684
      Credit: 25,950,161
      RAC: 34,820

      Maybe the warning on top of

      20 Jun 2005 22:28:09 UTC
      Message 2519 in response to message 2518
      (moderation:
        )

        Maybe the warning on top of the beta test page is not prominent enogh there - any better idea?

        BM

        BM

        Bernd Machenschalk
        Bernd Machenschalk
        Joined: 15 Oct 04
        Posts: 2,684
        Credit: 25,950,161
        RAC: 34,820

        RE: Looks like this new

        21 Jun 2005 8:47:00 UTC
        Message 2520
        (moderation:
          )

          Quote:
          Looks like this new test application took me about 11% longer to crunch than the standard version.

          Sounds strange. Neither the copiler options nor the "science code" that is used for the current Workunits has changed. Is this measured or estimated time you are referring to? CPU time or real time? Anything else changed on this machine?

          Did anyone else see this?

          BM

          BM

          Bernd Machenschalk
          Bernd Machenschalk
          Joined: 15 Oct 04
          Posts: 2,684
          Credit: 25,950,161
          RAC: 34,820

          RE: while the code may not

          21 Jun 2005 10:42:52 UTC
          Message 2521
          (moderation:
            )

            Quote:

            while the code may not have changed, it comes with a pdb. i assume it is debug code which would expalin the longer crunch time (though i have not checked if the 4.79 was debug or release code). cannot comment on the actual time it takes but the estimated time went up a bit.

            The "official" App downloads a pdb, too. It contains information for boincs Stackwalker, but no code that is executed. There have been some additions to the science code, so the code is not precisely the same, but these additions are not used or triggered by the current Workunits.

            BM

            BM

            Bernd Machenschalk
            Bernd Machenschalk
            Joined: 15 Oct 04
            Posts: 2,684
            Credit: 25,950,161
            RAC: 34,820

            RE: I just installed the

            21 Jun 2005 12:30:45 UTC
            Message 2522
            (moderation:
              )

              Quote:

              I just installed the beta test version (just after finishing a WU). I suspended the task - and got a new WU (of course) ;-)
              I 'stopped' the project ('no new work') and then suspended the second WU and the stopped the client.
              After starting version 0.03 I got the following messages for *BOTH* WUs:
              2005-06-21 09:14:24 [Einstein@Home] WU refers to nonexistent app_version: einstein 479
              2005-06-21 09:14:24 [Einstein@Home] Can't link workunit in state file

              --> all your WUs with a different version are cancelled!

              Pitty. But "stopping" the project is a good idea if you are running a newer client - try to get "no new work" for Einstein@Home, finish and report the Results you already got, then install the test App and resume the project - this should make you switch to a new App without losing a single Result.

              BM

              BM

              Bernd Machenschalk
              Bernd Machenschalk
              Joined: 15 Oct 04
              Posts: 2,684
              Credit: 25,950,161
              RAC: 34,820

              RE: One question...when E@H

              21 Jun 2005 12:36:52 UTC
              Message 2523
              (moderation:
                )

                Quote:
                One question...when E@H releases its new non-beta app, will 0.03 be replaced with it?

                I'm not sure I understand you - do you ask if, at some point, the current Beta App 0.03 will become the official one? Then the answer would be: this depends, e.g. on the outcome of this Beta test. Or do you ask if when we put up a new "official" Windows App there will continue to be a Beta test App with the version number 0.03? Then the answer is: probably not.

                BM

                BM

                Bernd Machenschalk
                Bernd Machenschalk
                Joined: 15 Oct 04
                Posts: 2,684
                Credit: 25,950,161
                RAC: 34,820

                Ok, that's good enough - I'll

                21 Jun 2005 19:02:33 UTC
                Message 2524
                (moderation:
                  )

                  Ok, that's good enough - I'll have a closer look on what changed. Probably wasn't intentional, though.

                  BM

                  BM

                  Bernd Machenschalk
                  Bernd Machenschalk
                  Joined: 15 Oct 04
                  Posts: 2,684
                  Credit: 25,950,161
                  RAC: 34,820

                  RE: RE: Ok, that's good

                  22 Jun 2005 8:35:37 UTC
                  Message 2525
                  (moderation:
                    )

                    Quote:
                    Quote:

                    Ok, that's good enough - I'll have a closer look on what changed. Probably wasn't intentional, though.

                    BM

                    I installed 0.03 on my office PC with an Intel P4 2GHz yesterday morning and had no significant difference in duration!

                    That's good to know.

                    We try to keep the WUs the same size, but of course there are always some fluctuations. I wonder if it could be that people seeing differences here have just by accident gotten a bit longer ones at first, but 10% is relatively large and there seem to be a quite a number of them...

                    Anyway, then please continue to report your differences in crunching time, even and especially if it is zero, so I can get a more significant picture.

                    Quote:
                    Yesterday evening I installed 0.03 on my home pc with an AMD 1700+. I have not finished a WU but the estimated time is 10h instead of 9h (aprox. 11% more).
                    Is it related to AMD CPUs?

                    I don't think so. It's probably more related to the (nominal) change of the platform ("x86-windows" to "anonymous"). I wouldn't give too much on the estimations, at least not before having finished a couple of Results on a new platform or after a benchmark run.

                    Quote:
                    Graphics with ATI now working fine (even normally not used due to CPU consumption).

                    Well, if you have an OpenGL accelerated (ATI) card, then the graphics shouldn't hold up the CPU (very much, i.e. noticably).

                    BM

                    BM

                    Bernd Machenschalk
                    Bernd Machenschalk
                    Joined: 15 Oct 04
                    Posts: 2,684
                    Credit: 25,950,161
                    RAC: 34,820

                    Ok, so far I can confirm that

                    22 Jun 2005 12:59:10 UTC
                    Message 2526
                    (moderation:
                      )

                      Ok, so far I can confirm that there is indeed a slowdown, which may show up on some machines more than on others, and is has to do solely with the version of boinc we are using, not with our science code and not with the compiler. Seems that the fixes put in lately slow it down noticably. I will continue to investigate this.

                      BM

                      BM

                      Bernd Machenschalk
                      Bernd Machenschalk
                      Joined: 15 Oct 04
                      Posts: 2,684
                      Credit: 25,950,161
                      RAC: 34,820

                      RE: Just a note to clarify,

                      22 Jun 2005 16:31:58 UTC
                      Message 2527
                      (moderation:
                        )

                        Quote:
                        Just a note to clarify, You're talking about the server side BOINC code, and not the core client versions that the end users are running, right? I don't want some users thinking that they can roll back their core client and see their speeds increase. ;)

                        Sorry, this needs some clarification indeed.

                        The boinc code can roughly be divided in "Server", "Client" and "Library". The Server code (which, in turn, has many parts) is what's running on our server, the Client code is what the "Boinc Core Client" you download is built from, and the Library code ist what is linked into the Applications (and the Core Client as well) that are downloaded to your machine by the Client. The slowdown happens in the latter part, the boinc library, which the user comes to see only as undistinguishable part of the Applications. The Application consist mainly of two parts: the "Worker" or science code which does the actual science work and number crunching, and the boinc library, which makes it fit into the rest of the boinc framework (e.g. handles communication with the core client, encapsulates system-dependent stuff like filename handling and some graphics stuff asf.).

                        So as a fix for the current slowdown we don't need to change something on our server and you don't need to change the BOINC Core Client on your machines, but we need to build new Apps when we found and fixed the problem.

                        Does this make it a bit clearer?

                        BM

                        BM

                        Comment viewing options

                        Select your preferred way to display the comments and click "Save settings" to activate your changes.

                        Forums › Cafe Einstein