Message boards :
Number crunching :
Optimised AP v5.00 - initial release
Message board moderation
Previous · 1 · 2 · 3 · 4 · 5 . . . 9 · Next
Author | Message |
---|---|
Leaps-from-Shadows Send message Joined: 11 Aug 08 Posts: 323 Credit: 259,220 RAC: 0 |
No but I'm wondering why the 5.00 you have running aren't showing up as ap_5.00r69_SSE3.exe, instead of astropulse_5.00_windows_intelx86.exe Testing the standard v5 app at Beta. I haven't gotten any new Astropulse units on Main yet. ...and the extra entry disappeared. Must have been some sort of fluke - I don't have it set to keep apps in memory when suspended. Cruiser Gateway GT5692 L-f-S Edition -Phenom X4 9650 CPU -4GB 667MHz DDR2 RAM -500GB SATA HD -Vista x64 SP1 -BOINC 6.2.19 32-bit client -SSE3 optimized 32-bit apps |
Josef W. Segur Send message Joined: 30 Oct 99 Posts: 4504 Credit: 1,414,761 RAC: 0 |
Possible least aggravation way: Jason's approach should be good. The problem is that the ap_4.37 builds have a slightly changed checkpoint file format, so trying a restart with a checkpoint from an earlier version is an error. I failed to think about that at the right time. Those who have done SETI Beta work will recognize the 4.37 number as the first attempt to have built-in radar removal. The Windows build did not actually do so when run with BOINC, making its results fully compatible with earlier versions. Joe |
Josef W. Segur Send message Joined: 30 Oct 99 Posts: 4504 Credit: 1,414,761 RAC: 0 |
No but I'm wondering why the 5.00 you have running aren't showing up as ap_5.00r69_SSE3.exe, instead of astropulse_5.00_windows_intelx86.exe BOINC has a feature; it automatically keeps an app in memory when suspended if it has not yet checkpointed. That's intended for applications which never checkpoint, but does also apply to other apps prior to their first checkpoint. Joe |
Byron S Goodgame Send message Joined: 16 Jan 06 Posts: 1145 Credit: 3,936,993 RAC: 0 |
Jason's approach should be good. The problem is that the ap_4.37 builds have a slightly changed checkpoint file format, so trying a restart with a checkpoint from an earlier version is an error. I failed to think about that at the right time. I'm not liking the sound of that...lol...setting NNT until there's more info on what we're doing. |
Cosmic_Ocean Send message Joined: 23 Dec 00 Posts: 3027 Credit: 13,516,867 RAC: 13 |
Jason's approach should be good. The problem is that the ap_4.37 builds have a slightly changed checkpoint file format, so trying a restart with a checkpoint from an earlier version is an error. I failed to think about that at the right time. I restarted a task that started with the 4.35 opti and am using the 4.37 opti and it's working just fine. No errors. Been running on 4.37 for about an hour now. An error should have shown up by now. Linux laptop: record uptime: 1511d 20h 19m (ended due to the power brick giving-up) |
OzzFan Send message Joined: 9 Apr 02 Posts: 15691 Credit: 84,761,841 RAC: 28 |
Jason's approach should be good. The problem is that the ap_4.37 builds have a slightly changed checkpoint file format, so trying a restart with a checkpoint from an earlier version is an error. I failed to think about that at the right time. For the second time tonight, I'm saying "me too" to Cosmic Ocean's post(s). |
jason_gee Send message Joined: 24 Nov 06 Posts: 7489 Credit: 91,093,184 RAC: 0 |
Excellent, I hope it goes likewise for the majority of users... I only built the executables, and it was a little mean to me, but then again that's more likely my weird setup due to testing development builds... All good, carry on! "Living by the wisdom of computer science doesn't sound so bad after all. And unlike most advice, it's backed up by proofs." -- Algorithms to live by: The computer science of human decisions. |
Byron S Goodgame Send message Joined: 16 Jan 06 Posts: 1145 Credit: 3,936,993 RAC: 0 |
|
jason_gee Send message Joined: 24 Nov 06 Posts: 7489 Credit: 91,093,184 RAC: 0 |
Absolutely... expect some performance boost. "Living by the wisdom of computer science doesn't sound so bad after all. And unlike most advice, it's backed up by proofs." -- Algorithms to live by: The computer science of human decisions. |
Leaps-from-Shadows Send message Joined: 11 Aug 08 Posts: 323 Credit: 259,220 RAC: 0 |
Initial estimates using CPU time and progress percentage (02:38:40 and 10.064%) point to increased efficiency - ~25 hours with v5 SSE3 optimized versus ~29 hours with old v4 SSE3 optimized on Cruiser. Lookin' good so far... I still have one 4.36-branded Astropulse unit in my queue, so we'll see how efficient the new v4 SSE3 optimized app is compared to the old v4 SSE3 optimized app as well. Cruiser Gateway GT5692 L-f-S Edition -Phenom X4 9650 CPU -4GB 667MHz DDR2 RAM -500GB SATA HD -Vista x64 SP1 -BOINC 6.2.19 32-bit client -SSE3 optimized 32-bit apps |
[B^S] madmac Send message Joined: 9 Feb 04 Posts: 1175 Credit: 4,754,897 RAC: 0 |
|
Byron S Goodgame Send message Joined: 16 Jan 06 Posts: 1145 Credit: 3,936,993 RAC: 0 |
|
[B^S] madmac Send message Joined: 9 Feb 04 Posts: 1175 Credit: 4,754,897 RAC: 0 |
|
Richard Haselgrove Send message Joined: 4 Jul 99 Posts: 14650 Credit: 200,643,578 RAC: 874 |
First message states advance users I am not advanced just a beginner With a joining date of Feb 9 04, I think you put yourself down if you still feel you're a beginner. On the other hand, it's never too late to learn - I discovered new things about BOINC while preparing that package that surprised even the oldest and wisest heads amongst us. The 'Advanced Users' warning in my post is really just a legal disclaimer to cover my backside: it's designed to at least warn the eager young puppies in the community that they should stop, think, and do the upgrade at a time when they can give it their full attention - don't just rush in mindlessly. The process is actually quite simple: * download the files * unpack them from the archive * join the new app_info to the relevent bit of your old one * put the files in the right directory * restart BOINC You've installed optimised apps before, so most of that is familiar: the only remotely complicated part is joining the two bits of app_info together. Have a look at the opening post again, and follow the link to Ned Ludd's description of the splicing process: it's really quite painless. If all else fails, have a go at it yourself, and send me the resulting file by PM so I can proof-read it for you. I say 'send it by PM', because I think it can be quite confusing for people to have a whole variety of app_info examples scattered across these boards on an "it works for me" basis - I would urge volunteers, however well-meaning their intentions, to avoid adding to the plethora of such examples unless they've really got something new to add. |
Cosmic_Ocean Send message Joined: 23 Dec 00 Posts: 3027 Credit: 13,516,867 RAC: 13 |
Hey, I've been around for years and didn't become an advanced user until about 6 weeks ago. I definitely agree that it's never too late to learn. There are lots of instructions floating around for how to change over to optimized applications. The problems I ran into were from doing actions before reading the rest of the instructions. The one thing that still hasn't been very clear in the instructions for installing the AP optimized application is that if, for some reason, you still want to use the stock MB app., you still have to reference the stock MB app in the app_info.xml file. However, if you're going opti on AP, you might as well do it for MB, too, since that cuts the crunch time about in half. You really should do both..I would recommend it. Linux laptop: record uptime: 1511d 20h 19m (ended due to the power brick giving-up) |
Leaps-from-Shadows Send message Joined: 11 Aug 08 Posts: 323 Credit: 259,220 RAC: 0 |
Some good news and some bad news from Cruiser... Bad news: Looks like I'm going to have my first 30+ (31.04 according to current CPU time/percentage calculations) hour Astropulse unit using the v4.37 optimized app. It actually seems to be slower than the v4.35 optimized app. This is on a 4.36-branded work unit that crunched about two hours on the old app and the rest on the new app. I will be confirming on my last 4.36-branded work unit, done completely with the v4.37 optimized app. It started crunching a little over an hour ago, and the current estimate is 33.94 hours. Definite thumbs-down on that... Good news: My first Astropulse work unit crunched with the v5 optimized app is on track for a completion time of 24.8 hours. Woohoo! Two thumbs up! Edit: As of approximately 16:00 UTC, the hybrid work unit from the bad news entry ended with Found 30 single pulses and 30 repeating pulses, exiting error. Task details. The second unit under bad news (the last one branded v4.36) is still crunching. My machine also downloaded another five Astropulse work units, along with a whole bunch of Multibeams. That gives Cruiser a total of 10 v5.00-branded Astropulse units in its queue. Cruiser Gateway GT5692 L-f-S Edition -Phenom X4 9650 CPU -4GB 667MHz DDR2 RAM -500GB SATA HD -Vista x64 SP1 -BOINC 6.2.19 32-bit client -SSE3 optimized 32-bit apps |
Richard Haselgrove Send message Joined: 4 Jul 99 Posts: 14650 Credit: 200,643,578 RAC: 874 |
Bad news: That's exactly why it would be helpful to have a range of different CPUs in the testing fleet. Sorry it didn't work out this time - everything got caught up in the rush following the project's release of v5.00 - but keep knocking on the door at KWSN and maybe you can help spot little problems like that before release next time. ... Found 30 single pulses and 30 repeating pulses, exiting error. Task details. That isn't an error: it's actually an optimisation in the stock app, a planned exit route designed to avoid wasting CPU cycles on tasks that aren't worth it. It's possible that that WU is contaminated with exactly the sort of radar interference that led to the development of the v5.00 app in the first place. Don't shed any tears over it. |
1mp0£173 Send message Joined: 3 Apr 99 Posts: 8423 Credit: 356,897 RAC: 0 |
when will it be available for us beginners so that we can do both MB and APs? The problem is that you need the right app_info.xml for the correct optimized multibeam, and the right app_info.xml for the correct optimized AP. Merging two XML files into one is trivially easy -- it's harder to explain than to do. Just tack the two together in your favorite editor, remove the "end" tag from the first one and the "beginning" tag from the second one. If you're having trouble with that, it's because you think it should be harder. |
[B^S] madmac Send message Joined: 9 Feb 04 Posts: 1175 Credit: 4,754,897 RAC: 0 |
|
Logan Send message Joined: 26 Jan 07 Posts: 743 Credit: 918,353 RAC: 0 |
Logan did a xml file for me so that I can do both AP and MB, so I have downloaded the file and see what is the difference, then see if I can save it as an xml file. Mine seems to open in outlook not notepad. Dont open from outlook, save it first, and then, if you want, can open it with notepad or winpad, but you only needs to replace the old one with that file. Logan. BOINC FAQ Service (Ahora, también disponible en Español/Now available in Spanish) |
©2024 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.