Message boards :
Number crunching :
How to Switch to the New Official App?
Message board moderation
Author | Message |
---|---|
Send message Joined: 7 Jul 08 Posts: 47 Credit: 13,629,944 RAC: 0 |
When the new official app is released, how do we switch to it from Milksop's optimized app? Thanks. --Bill |
Send message Joined: 26 May 08 Posts: 1 Credit: 17,772,148 RAC: 0 |
Just delete the app_info file in your Milkyway project folder and restart BOINC. It should download automatically then.. |
Send message Joined: 17 Feb 08 Posts: 363 Credit: 258,227,990 RAC: 0 |
When the new official app is released, how do we switch to it from Milksop's optimized app? Simply delete the app_info.xml file in the project folder and restart boinc... however, i'm not sure why you would want to do that... It's highly unlikely that the new official app will be faster. Join Support science! Joinc Team BOINC United now! |
Send message Joined: 1 Oct 08 Posts: 106 Credit: 24,162,445 RAC: 0 |
Simply delete the app_info.xml file in the project folder and restart boinc... however, i'm not sure why you would want to do that... Maybe it will be incompatible with some new WUs? But I really hope the factor 2 to 3 speedup Travis reported is not a comparison to the old official one. As you certainly know, one would get that by changing 5 lines of code or so. Or for new CPUs even just by using another compiler ;) |
Send message Joined: 30 Aug 07 Posts: 2046 Credit: 26,480 RAC: 0 |
Simply delete the app_info.xml file in the project folder and restart boinc... however, i'm not sure why you would want to do that... Well, the numbers i have from the new version (which has all your suggested improvements) are from using mpicc on our linux cluster. It's standing at 4x faster than the old app. I'm hoping this will be even better on different architectures and other compilers. Hopefully we'll see how it compares this afternoon. |
Send message Joined: 30 Aug 07 Posts: 2046 Credit: 26,480 RAC: 0 |
Actually I just noticed something i missed. Looks like its 8x-9x faster now after a small change :P |
Send message Joined: 4 Oct 08 Posts: 1734 Credit: 64,228,409 RAC: 0 |
Looking forwards to the release of the new client, and the flood of WUs to meet the expected demand. Question Travis, please: There is a little confusion with us crunchers as to your meaning of the speed up of your new Linux tested client. What client is the new client's speed up compared to, the old slow client (Windows version for me) of Milksop's fast version? Is this compared the old Linux client, which was certainly slow under Windows? Or is the speed up compared the equivalent you agreed Milksop could release, as this was between 50x and 60x as fast as the old Windows client? Happy crunching, and is the work destined for the new client still compatible with the Milksop client (the Windows version in my case)? Finally, when (in UTC terms) is the afternoon timing for the client release you refer to? |
Send message Joined: 28 Aug 07 Posts: 31 Credit: 86,152,236 RAC: 0 |
Thanks for the updates, Travis. Especially the one in another thread about what's new in new version. This somehow answer my question about progress of project... As I see it, "Optimalized" version enables us to perform what we have been doing whole year withing a very few days. I'm still not sure if it is good news or bad news. New app version is aiming at both - doing better science and doing it faster. BOINC Project specifications and hardware requirements |
Send message Joined: 20 Sep 08 Posts: 1391 Credit: 203,563,566 RAC: 0 |
Thanks for the heads up there Travis, it's appreciated. Interesting times, we will see what we see.... Don't drink water, that's the stuff that rusts pipes |
Send message Joined: 4 Oct 08 Posts: 1734 Credit: 64,228,409 RAC: 0 |
|
Send message Joined: 30 Aug 07 Posts: 2046 Credit: 26,480 RAC: 0 |
Looking forwards to the release of the new client, and the flood of WUs to meet the expected demand. I'm comparing the old (slow) version of the application on one of our linux boxes to the new one we'll be uploading today. I don't have milksoup's to compare to. Dave's going to compile it under windows to see how the old client compares to our new one. I'll compare OSX sometime today. New workunits are NOT backwards compatible with previous app versions, there's a bit of new stuff in there and it will have it's own new compiler, so older application versions will not work with it. We're going to be releasing it as a new application (not a new version of the old one) and it will have it's own running validator/assimilator/etc. We'll keep both running concurrently until we've gotten all the bugs sorted out with the new one, then we'll stop generating work for the old application and kill the old assimilator after the remaining work gets back. I don't know if it will be as fast as milksoup's but i've implemented all his suggestions so i'm hoping it will be close. We'll be releasing the code so hopefully people will take a look at it and give us feedback so we can further improve it's performance. |
Send message Joined: 6 Sep 07 Posts: 66 Credit: 636,861 RAC: 0 |
We're going to be releasing it as a new application (not a new version of the old one) and it will have it's own running validator/assimilator/etc. We'll keep both running concurrently until we've gotten all the bugs sorted out with the new one, then we'll stop generating work for the old application and kill the old assimilator after the remaining work gets back. How about then adding options to the Milkyway Preferences to allow users which applications to run? TIA |
Send message Joined: 4 Oct 08 Posts: 1734 Credit: 64,228,409 RAC: 0 |
I think Travis states that once the bugs/kinks are out of the new faster official client code, and things are seen to run smoothly (without errors, etc). Also, as the WUs for this new official client are different from the old slow client, and for the Milksop/Crunch3r code, then the WUs for the new faster official client the old WUs will no longer be issued. That means the old slow client and Milksop/Crunch3r's code will no longer work. By asking for the new WUs to work with the old code (Milksop's I guess) it would complicate things too much. The Admins have devised a direction and they should stick to it. Perhaps later, coders could look at the new source code and tweak it for more speed? |
Send message Joined: 30 Aug 07 Posts: 2046 Credit: 26,480 RAC: 0 |
We're going to be releasing it as a new application (not a new version of the old one) and it will have it's own running validator/assimilator/etc. We'll keep both running concurrently until we've gotten all the bugs sorted out with the new one, then we'll stop generating work for the old application and kill the old assimilator after the remaining work gets back. I thought it was possible to select which applications you would run client side, that was part of the reason we made it a different application. |
Send message Joined: 6 Sep 07 Posts: 66 Credit: 636,861 RAC: 0 |
I thought it was possible to select which applications you would run client side, that was part of the reason we made it a different application. Indeed, but other projects do that through the project's preferences in the user's account page. I don't see something like this in my Milkyway account page. Or do you mean something else? TIA |
Send message Joined: 27 Aug 07 Posts: 647 Credit: 27,592,547 RAC: 0 |
I thought it was possible to select which applications you would run client side, that was part of the reason we made it a different application. The option to run test apps is there: http://milkyway.cs.rpi.edu/milkyway/prefs.php?subset=project If the optimized app would get the "test app" flag it would work I guess? Lovely greetings, Cori |
Send message Joined: 6 Sep 07 Posts: 66 Credit: 636,861 RAC: 0 |
The option to run test apps is there: http://milkyway.cs.rpi.edu/milkyway/prefs.php?subset=project In a crude way, yes. But I'd rather not run "test" applications on my systems. Likewise, I'd rather not see this option to mean something other than what it says. But other projects allow the user to specify which applications to run. For example, IBERCIVIS: Resource share 100 Run test applications? no Default computer location school Color scheme for graphics Tahiti Sunset Max frames/second for graphics 1 Maximum CPU % for graphics 1 Run only the selected applications bless-docking: no iter: no materiales128: no materiales16: yes materiales24: yes materiales32: no materiales48: no materiales64: no nanoluz: yes neurosim: yes protomol: yes |
Send message Joined: 30 Aug 07 Posts: 2046 Credit: 26,480 RAC: 0 |
I'm looking around trying to find where I can enable this option but i'm not seeing anything... I'll email the boinc projects list to get some info (but any response will probably take awhile). |
Send message Joined: 27 Aug 07 Posts: 647 Credit: 27,592,547 RAC: 0 |
I'm looking around trying to find where I can enable this option but i'm not seeing anything... I'll email the boinc projects list to get some info (but any response will probably take awhile). You may want to ask Rytis from PrimeGrid - they have also several different apps to chose from. ;-) Lovely greetings, Cori |
Send message Joined: 28 Aug 07 Posts: 52 Credit: 8,353,747 RAC: 0 |
I'm looking around trying to find where I can enable this option but i'm not seeing anything... I'll email the boinc projects list to get some info (but any response will probably take awhile). http://boinc.berkeley.edu/trac/wiki/BetaTest http://boinc.berkeley.edu/trac/wiki/PrefsImpl |
©2024 Astroinformatics Group