Computing tasks

Questions and Answers : Preferences : Computing tasks
Message board moderation

To post messages, you must log in.

AuthorMessage
ScottieD369

Send message
Joined: 8 Jun 15
Posts: 26
Credit: 264,597
RAC: 2
United States
Message 2028769 - Posted: 23 Jan 2020, 16:47:27 UTC

Hi there I want to know if there is a way that I can get more tasks to compute? I can click update but only "2" run on my pc. Then on my Raspberry Pi cluster I have 4 running on the at a time. I see other users on here that have average million?! How is this possible? I can only get with my pc and pi's up too 2k avg. How can I get this number higher? On my pc and pi's?

Thanks
Scott

Keep looking UP!
ID: 2028769 · Report as offensive
Profile Bernie Vine
Volunteer moderator
Volunteer tester
Avatar

Send message
Joined: 26 May 99
Posts: 9941
Credit: 103,452,613
RAC: 328
United Kingdom
Message 2028791 - Posted: 23 Jan 2020, 18:15:45 UTC
Last modified: 23 Jan 2020, 18:19:48 UTC

SETI@Home uses CPU cores, one for each work unit, it cannot do more, so your Pi's can do 4 at a time your PC 2 at a time

Most of the users with high RAC are using graphics cards (GPU's) to process work. GPU's can process work faster than CPU's. Also on Linux there is a "special app " that can boost the speed even more.

Many users have multiple GPU's that can easily get large RAC's

I run one Linux machine with just 2 Nvidia GTX 1060's and it maxes out at around 83,000 RAC
ID: 2028791 · Report as offensive
ScottieD369

Send message
Joined: 8 Jun 15
Posts: 26
Credit: 264,597
RAC: 2
United States
Message 2028795 - Posted: 23 Jan 2020, 18:27:33 UTC - in response to Message 2028791.  
Last modified: 23 Jan 2020, 18:29:13 UTC

OK so how can I speed up the process on the Pi's?
I know the Pi's have GPU.
Yes I guess that is my question how can I get a high RAC #?
ID: 2028795 · Report as offensive
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 12369
Credit: 1,160,866,277
RAC: 1,873
United States
Message 2028797 - Posted: 23 Jan 2020, 18:32:21 UTC - in response to Message 2028795.  
Last modified: 23 Jan 2020, 18:33:22 UTC

Nobody has written a science application for the Pi's igpu. You are most welcome to do so.
To get high RAC, use hardware supported by the projects, preferably mainstream Nvidia and AMD.
Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)
ID: 2028797 · Report as offensive
ScottieD369

Send message
Joined: 8 Jun 15
Posts: 26
Credit: 264,597
RAC: 2
United States
Message 2028800 - Posted: 23 Jan 2020, 18:46:38 UTC - in response to Message 2028797.  

ok I seen a few videos on youtube now about 13 GPU mining. basically if I ran that I would have those numbers. Looks expensive for me at the moment. Wish I could contribute like that!
ID: 2028800 · Report as offensive
ScottieD369

Send message
Joined: 8 Jun 15
Posts: 26
Credit: 264,597
RAC: 2
United States
Message 2028804 - Posted: 23 Jan 2020, 18:58:53 UTC

I hope this down time is over soon! My work has been finally reported. No new tasks yet.


Always Looking UP!
ID: 2028804 · Report as offensive
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 12369
Credit: 1,160,866,277
RAC: 1,873
United States
Message 2028813 - Posted: 23 Jan 2020, 19:29:53 UTC - in response to Message 2028804.  

It usually takes an hour or so before the schedulers start sending out work once they start accepting work upload reports. Be patient. You are in a 100,000 large feeding frenzy pool of empty hosts.
The trick to reporting is to limit the number reported to a small value of 50 or so. And then set NNT to not ask for work. The scheduler is more likely to recognize your report when you don't ask for work.
I have 15,000 tasks or so to report. Will take most of the day.
Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)
ID: 2028813 · Report as offensive
Profile Bernie Vine
Volunteer moderator
Volunteer tester
Avatar

Send message
Joined: 26 May 99
Posts: 9941
Credit: 103,452,613
RAC: 328
United Kingdom
Message 2028951 - Posted: 24 Jan 2020, 8:49:33 UTC
Last modified: 24 Jan 2020, 8:53:32 UTC

Just so you can see the difference in various processing speeds:

Note these are very "average" values based on a few samples from validated tasks

Raspberry Pi takes around 8 hours per task therefore 3 tasks per day

Your Intel(R) Celeron(R) CPU B820 @ 1.70GHz tasks around 7 hours per task therefore 3.5 tasks a day

My Intel(R) Core(TM) i5-7400 CPU @ 3.00GHz running windows 10 and using the Lunatics optimised app takes around 1 hour per task 24 tasks a day per core ( I only run SETI@Home on 2 cores so I don't slow the machine down.)

My Nvidia GTX 1660Ti running on the above machine takes around 10 minutes a task 144 tasks a day

My Linux machine does not run any CPU tasks but has 2x Nvidia GTX1060 running the optimised app, each card does a task in around 3 minutes so 400 tasks a day per GPU

As I have said these are very rough figures and different types of task take different amounts of time to process, but I hope give you an idea of how to process more tasks.

At the moment the fastest way is a modern Nvidia GPU running Linux with the oprimised app
ID: 2028951 · Report as offensive

Questions and Answers : Preferences : Computing tasks


 
©2021 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.