MS Windows Beta Test App 4.62 available
5 Jun 2006 12:29:36 UTC
Topic 13348
(moderation:
I put a new Windows App for testing on the Beta Test Page.
We are still working on incorporatinging Akos' optimizations into an App that meets the requirements for our "official" Apps, but it turned out to take somewhat longer than we expected.
As we still need a new App to start the next run with, I made this intermediate release of a standard Windows App from the same codebase as the 4.58 Linux App.
Short tests showed a speedup of 3-4 compared to our current official Window App 4.37. This is not as fast as Akos' Apps with architecture-specific optimizations, but as this will probably be the official Windows App we start the next run with (if this test turns out to be successful), I strongly encourage you to test it beforehand.
BM
BM
MS Windows Beta Test App 4.62 available
)
The credit is claimed by the Core Client (and the scheduler), and granted by the validator, so primarily this doesn't have much to do with the App.
There is a way to tell the Client the amount of FLOPs that's actually used from the App, instead of using the client's benchmark result, but it's hard to implement a counting that works on all CPUs and it would just balance for the different benchmark results of different Clients, but does not much about the different efficiencies of the Apps on different CPUs.
So in the next run, starting in about two weeks, we will grant a rather fixed credit for each (valid) Task done. But that, again, is rather related to the scheduler and the validator, not to the App.
There will be no change regarding crediting for the rest of the current run.
BM
BM
RE: Does this mean that
)
I simply can't say that right now. It is covered in internal tests of the new setup which are not finished yet.
Akos' Apps are still based on the 4.37 App, which is pretty old code, especially the BOINC part. If it turns out we need e.g. fixes or features from newer BOINC, then you won't be able to continue using these Apps.
BM
BM
I see that the anonymous
)
I see that the anonymous platform scheme we use for public Beta Test is badly prepared for the transition to S5R1 we're doing now.
With that transition I'd consider the Beta Test to be over - and successful. The Apps we will use for S5R1 are almost the same than the ones we had in public Beta Test, the improved science code is actually identical, thus the new Apps will be precisely as fast as the Beta Test Apps.
Thank you for all your help, this was incredibly useful!
However the Beta Test Apps are lacking some features we found we need for the new run, and thus can't be used for the new Workunits. There is a bit of S4 work left, but no guarantee that the scheduler will try to give it to your machine. So
- if you want to keep your machines busy regardless of the type of work processed, just remove the app_info.xml file (and restart the client). You will download the new App for the new work, and the old official App for the old S4 work.
- if you prefer to do the old S4 work while it is available and risk your machine runs dry, keep the app_info.xml until there's no more S4 work left (we'll add an entry for that on the server status page).
- if you want to live risky and enjoy it, try the following: remove the app_info.xml, but keep a copy of the Beta App executables somewhre outside the project directory. Restart the client. After getting S4 work and having downloaded the old official App, stop the client again and replace the executables of the official App with the backups of the Beta Test App. Then start the client again. There's no guarantee that this will fool all clients on all machines, but it should work most of the time and finish the S4 work with the speed of the Beta Apps.
Have fun!
BM
BM