Questions and Answers :
Windows :
How is the analysis program for a task selected?
Message board moderation
Author | Message |
---|---|
Bert Hyman Send message Joined: 18 May 99 Posts: 15 Credit: 1,236,995 RAC: 15 |
I'm running an i5-3470 CPU on Windows 10 Pro, ID: 8891814. It currently has two AstroPulse tasks listed, one running, one waiting. The one that's running is using AstroPulse v7 v7.00 windows_intelx86 (a 32-bit program). The one that's waiting will use AstroPulse v7 v7.03 (sse2) windows_x86_64 (probably a 64-bit program). The "SETI@home v8" tasks are processed by a small handful of programs with different versions, some apparently 64-bit, others 32-bit. I understand that there's a performance penalty in running 32-bit software on 64-bit processors, although I don't know how much. What factors go into choosing what software is applied against a given task? Simple curiosity here; I don't expect to do anything on my end to change anything. |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873 |
The underlying code for the science app is the same regardless whether the app is compile for a 32bit OS or a 64bit OS. The results will be the same if the hardware and software of the hosts tasked with crunching the task are working correctly. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
Bert Hyman Send message Joined: 18 May 99 Posts: 15 Credit: 1,236,995 RAC: 15 |
OK. So the 32- and 64-bit executables with the same version are the same. But in the AstroPulse tasks, I see versions 7.00, .03 and .04. Then, I see in SETI@home v8 executables, versions 00, 05, 08 being applied to seemingly similar tasks. There are probably other versions of those executables I haven't run across. It's what drives those choices that I'm curious about, as well as why use a 32-bit executable on a 64-bit machine, (assuming a matching 64-bit executable exists). There's certainly a good reason for all of this. Most likely, there's ->something different among the thousands of tasks that are generated, and there's automatic filtering going on to get the appropriate executable paired with each task. And, it's likely I'd never understand a detailed explanation. So, if I'm right about that, just say so and I'll go away, happy. Heck. I should probably just go away anyway. |
Grant (SSSF) Send message Joined: 19 Aug 99 Posts: 13842 Credit: 208,696,464 RAC: 304 |
But in the AstroPulse tasks, I see versions 7.00, .03 and .04.There are multiple applications that have been developed over the years for improved performance, ie reduced processing time. However some more recent applications aren't as good on older hardware, and others can be run on older hardware at all. So the BOINC Manager will try each Stock application out on a selection of work to determine which is the best to go with (however due to the different types of WUs (Work Units) processed it can & does sometimes select one that isn't best best, & on occasion can end up selecting the worst). as well as why use a 32-bit executable on a 64-bit machine, (assuming a matching 64-bit executable exists).Computers and operating systems going from 32bits to 64bits was never about performance, just about memory. 64bit systems can handle (without all sorts of weird tricks) a lot more than a 32bit system can. So there was no benefit in producing a 64bit application as it wouldn't have improved the accuracy of the science, nor would it have resulted in improved performance. Grant Darwin NT |
Harri Liljeroos Send message Joined: 29 May 99 Posts: 4741 Credit: 85,281,665 RAC: 126 |
You can see how each application is doing on your computer by checking it here on the web site from your computer details page and selecting there Application details. Note that it takes at least 10 validated tasks to be finished with application until you can "trust" those numbers. |
©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.