Message boards :
Number crunching :
Astropulse Errors-Optimized version 5
Message board moderation
Previous · 1 . . . 3 · 4 · 5 · 6
Author | Message |
---|---|
jason_gee Send message Joined: 24 Nov 06 Posts: 7489 Credit: 91,093,184 RAC: 0 |
... That will bear further investigation if that's all that's going on in your case, as those privileges should be inherited from the appropriate folder when the files are put in place :S Will have to get out the MS certification handbook to figure that one out ;D. Cheers, Jason "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. |
OzzFan Send message Joined: 9 Apr 02 Posts: 15691 Credit: 84,761,841 RAC: 28 |
... It would depend on if he did a move or a copy operation into the target folder and if its on the same drive or a different drive. If it was a move operation on the same drive, it should retain its original permissions. All other operations (move from different drive, copy from same drive, copy from different drive) should have inherited the target folder's permission settings. |
Scrooge McDuck Send message Joined: 26 Nov 99 Posts: 701 Credit: 1,674,173 RAC: 54 |
It would depend on if he did a move or a copy operation into the target folder and if its on the same drive or a different drive. That's right. I removed the optimized AP client and downloaded it again. I unpacked the client to a folder on the WinXP desktop and then moved the binaries (.exe) to the appropriate folder. The moved files retain (as OzzFan said), their original access rights. The inheritance of the target folder's permission settings don't work in this case. So one has to manually edit the access rights of the moved binaries and has to activate the "inherit access rights from upper objects" check box. (I only guessed the english checkbox description from the one of my german WinXP: "Berechtigungen übergeordneter Objekte auf untergeordnete Objekte, sofern anwendbar, vererben. Diese mit den hier definierten Einträgen mit einbeziehen") I never really liked this complex access right mechanism of Windows NT. The UNIX way of MacOS X, Solaris, Linux, ... is much easier to manage. |
Richard Haselgrove Send message Joined: 4 Jul 99 Posts: 14653 Credit: 200,643,578 RAC: 874 |
The alternative technique - holding down the Ctrl key to force a copy when dragging, or copy/paste with Ctrl-C / Ctrl-V, is much easier when there are complex permissions like this. Also, you get to keep the original download until you can be sure that everything is working properly! |
Scrooge McDuck Send message Joined: 26 Nov 99 Posts: 701 Credit: 1,674,173 RAC: 54 |
I wasn't aware of any different consequences to access permissions when moving instead of copying some files. Maybe it's possible to add a hint to the instruction file of optimized client releases for Windows that it is important to copy and never to move the contained files to avoid problems with file access permissions. |
Sharky Send message Joined: 31 Mar 07 Posts: 12 Credit: 235,260 RAC: 0 |
|
OzzFan Send message Joined: 9 Apr 02 Posts: 15691 Credit: 84,761,841 RAC: 28 |
Why does AP V5 5.03 took so long time ? Because they refined the searching process which requires more computing time. |
tullio Send message Joined: 9 Apr 04 Posts: 8797 Credit: 2,930,782 RAC: 1 |
Why does AP V5 5.03 took so long time ? Mi first AP V5 should take about 100 hours on my Opteron 1210 running Linux. My wingman has a Core2 DUO and will probably take even less. Tullio |
Sharky Send message Joined: 31 Mar 07 Posts: 12 Credit: 235,260 RAC: 0 |
|
OzzFan Send message Joined: 9 Apr 02 Posts: 15691 Credit: 84,761,841 RAC: 28 |
Sharky: Change your tags in your signature from [/url] [/url] to [img] [/img] |
Sharky Send message Joined: 31 Mar 07 Posts: 12 Credit: 235,260 RAC: 0 |
|
Josef W. Segur Send message Joined: 30 Oct 99 Posts: 4504 Credit: 1,414,761 RAC: 0 |
Why does AP V5 5.03 took so long time ? Just to clarify, those timings are for stock AP 5.00 and AP v5 5.03. IIRC the mods combined stock discussion into this thread without changing the title, so the post is on topic on that basis. The additional searching for repetitive pulses from data with negative dispersion can be expected to increase stock time by about 67%, the more complex data inserted for blanking can add another 30% or so. Joe |
speedimic Send message Joined: 28 Sep 02 Posts: 362 Credit: 16,590,653 RAC: 0 |
Just wann a say THANK YOU to anyone envolved in the making of the new optimized AP apps!! Looking at the completion times (and those of the stock app) puts a big smile on my face! BTW, my rigs seem to like it too - they only download AP wus... mic. |
TheMatrix Send message Joined: 20 May 99 Posts: 1 Credit: 711,233 RAC: 0 |
I seem to be observing some strange scheduler behavior when running astropulse and seti enhanced with cuda. The machine I see it on is a dual core AMD, with dual nvidia 8800's in it. It rips through seti wu's on the nvidia's of course and keeps working diligently on them when there is a single astropulse unit running on one of the cpu's. The trouble starts when a second astropulse work unit is downloaded and begins running on the second processor of the AMD. As soon as I have two astropulse units running on the cpu, it will finish up the gpu cuda units it has in reserve and then stops asking for anymore... When one of the astropulse units finally starts getting close to completion, the scheduler finally runs out and requests more work units, resulting in another burst of seti enhanced + cuda processing. This is all fine and good, but leaves my nvidia's idle most of the time, since astropulse units take so long to complete. If I were to guess, I would say there is a logic problem that only checks the number of processors available to see if it needs to get more work, and doesn't check for an aggregate of processors(cpu) + cuda devices(gpu) and then request more cuda-enabled work units. 3/10/2009 6:11:59 PM||Starting BOINC client version 6.4.7 for windows_x86_64 3/10/2009 6:11:59 PM||Processor: 2 AuthenticAMD Dual-Core AMD Opteron(tm) Processor 2210 [AMD64 Family 15 Model 65 Stepping 2] 3/10/2009 6:11:59 PM||OS: Microsoft Windows XP: Professional x64 Editon, Service Pack 2, (05.02.3790.00) 3/10/2009 6:11:59 PM||CUDA devices found Any thoughts on how I can keep cuda units running on my gpu's, while still handling dual astropulse units on my cpu's? |
Richard Haselgrove Send message Joined: 4 Jul 99 Posts: 14653 Credit: 200,643,578 RAC: 874 |
Welcome to the boards, and congratulations on such a thoughtful and well-presented first post. I would consider first of all the cache size setting in your preferences - the "Maintain enough work for an additional: 0.1 days" value in the 'Network usage' section of your Computing preferences page. I have seen cases where BOINC looks at a long-running task such as Astropulse, and sees that it will last long enough on its own to meet your requested work reserve - but isn't clever to realise that a single task can't be spread across three processing resources. Try setting that "Maintain enough work..." value to something just longer than Astropulse tasks usually take on your CPUs: you need to update the web-site page first, then go into BOINC Manager and update the SETI Project from there. [actually, it might be easiest to test this first directly from the Manager. Go into Advanced view, drop down the Advanced menu, and choose 'Preferences...'. You can set the same 'Additional work buffer' on the middle tab there. Try smaller values first: then, if you find one that works, clear the settings in BOINC Manager, and update the website with the figure you want to use.] If experimenting with the cache size doesn't sort it, please post again with a note of how long AP tasks usually take - both the CPU time reported by BOINC, and the actual beginning-to-end time on your computer, if you don't leave it running 24 hours a day. The other problem you might be running into is deadline pressure - that would also inhibit BOINC from fetching more work. |
Charles Anspaugh Send message Joined: 11 Aug 00 Posts: 48 Credit: 22,715,083 RAC: 0 |
I think this was brought up already,. the preference settings listed below should result in cuda downloads with ap and apv5. Run only the selected applications SETI@home Enhanced no Astropulse yes Astropulse v5 yes If no work for selected applications is available, accept work from other applications? no Use Graphics Processing Unit (GPU) if available yes the only way to get cuda to download is to change either the SETI@home Enhanced yes or If no work for selected applications is available, accept work from other applications? yes If I change the settings to: Run only the selected applications SETI@home Enhanced yes Astropulse yes Astropulse v5 yes If no work for selected applications is available, accept work from other applications? yes Use Graphics Processing Unit (GPU) if available yes I get ap, apv5, cuda. but no seti@home at all. |
Richard Haselgrove Send message Joined: 4 Jul 99 Posts: 14653 Credit: 200,643,578 RAC: 874 |
I think this was brought up already,. CUDA work is SETI@home Enhanced work. |
Charles Anspaugh Send message Joined: 11 Aug 00 Posts: 48 Credit: 22,715,083 RAC: 0 |
cuda work works only on video card, seti@home work is on cpu only. what should the settings be to get ap,apv5 and cuda with no seti@home cpu |
Richard Haselgrove Send message Joined: 4 Jul 99 Posts: 14653 Credit: 200,643,578 RAC: 874 |
cuda work works only on video card, I'm sorry, but strictly speaking seti@home work - setiathome_enhanced work, to be even more pedantic - is a single pool of work which is capable of being processed on either the CPU (application v6.03) or CUDA (application v6.08). I understand your request for settings like that: off the top of my head, I can't think of a way of achieving what seems to be a very rational choice. I wonder what the developers would have to say on the subject (there was a discussion in the bug_fix mailing list about this last night). Since you are posting in a thread about optimised applications, it seems pertinent to point out that the settings you desire could very easily be achieved with an optimised Astropulse application, with an additional app_info entry for the CUDA version of the setiathome_enhanced application. Again being strictly pedantic there are no 'optimised' versions of the CUDA application available yet, though there are modified versions which work round certain problems like the slow processing of VLAR work. Whether you choose to use one of these modified versions, or just accept that the VLAR work has to be processed by somebody and it may as well be you, is up to you. |
Charles Anspaugh Send message Joined: 11 Aug 00 Posts: 48 Credit: 22,715,083 RAC: 0 |
If this is true, then why on the preference page does it specify : SETI@home Enhanced Use Graphics Processing Unit (GPU) if available If they are the same then why two separate settings. I understand your request for settings like that: off the top of my head, I can't think of a way of achieving what seems to be a very rational choice. I wonder what the developers would have to say on the subject (there was a discussion in the bug_fix mailing list about this last night). I am just reporting a possible problem, It is not a complicated problem. It just may be that I have set something wrong. I posted in the astropulse errors-optimized version 5 thread because that is what I am running, I am not interested in optimized cuda or VLAR wich by the way I have no idea what VLAR is. If I posted in the wrong thread, please let me know and I can repost in the correct thread. |
©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.