Message boards :
News :
New N-Body Runs
Message board moderation
Previous · 1 · 2 · 3 · Next
Author | Message |
---|---|
Send message Joined: 18 Nov 22 Posts: 84 Credit: 640,530,847 RAC: 0 |
easier way. Google: "BOINC Client Config" click the first link. |
Send message Joined: 13 Apr 17 Posts: 256 Credit: 604,411,638 RAC: 0 |
easier way. +1 Excellent - never thought about this possibility. But, as I said somewhere before, I (personally) sort of like/prefer the long way. It costs time, but I always enjoy finding some other interesting things along the way ... Have a nice Sunday S-F-V |
Send message Joined: 6 May 23 Posts: 2 Credit: 2,101,909 RAC: 0 |
Thank you. I improved the CPU usage using your method. |
Send message Joined: 2 Mar 20 Posts: 131 Credit: 320,573,149 RAC: 15,761 |
Seems to me that n body tasks are too big to run. Mine usually take about 2 days and often die before finishing; Is this the nature of the beast? |
Send message Joined: 19 Jul 10 Posts: 627 Credit: 19,303,095 RAC: 855 |
Seems to me that n body tasks are too big to run. Mine usually take about 2 days and often die before finishing; Is this the nature of the beast?Are you using less than 100% of CPU time? |
Send message Joined: 2 Mar 20 Posts: 131 Credit: 320,573,149 RAC: 15,761 |
Seems to me that n body tasks are too big to run. Mine usually take about 2 days and often die before finishing; Is this the nature of the beast?Are you using less than 100% of CPU time? I was using 75 on percent cpu time, but just changed to 100. I have a i3-8130u 2.2GHz with Turbo Boost to 3.4GHz. Boinc shows 1 cpu with 4 cores. I was concerned about temperature as it's a laptop. Using the Intel GPU for running Einstein. Btw, the wu's started at 8 min and 6 secs and 10 min and 10 secs and are now at 18 hours and over a day. Application Milkyway@home N-Body Simulation 1.82 (mt) Name de_nbody_02_27_2023_v182_pal5__data__6_1685648250_144011 State Running Received 6/23/2023 1:24:33 PM Report deadline 7/5/2023 1:24:33 PM Resources 2 CPUs Estimated computation size 5,916 GFLOPs CPU time 00:37:12 CPU time since checkpoint 00:00:45 Elapsed time 00:47:44 Estimated time remaining 1d 06:03:51 Fraction done 2.578% Virtual memory size 12.88 MB Working set size 16.86 MB Directory slots/1 Process ID 3528 Progress rate 3.240% per hour Executable milkyway_nbody_1.82_windows_x86_64__mt.exe |
Send message Joined: 8 May 09 Posts: 3339 Credit: 524,010,781 RAC: 0 |
Seems to me that n body tasks are too big to run. Mine usually take about 2 days and often die before finishing; Is this the nature of the beast?Are you using less than 100% of CPU time? You do know that the Intel gpu and Intel I3-8130u are the exact same thing...right? Personally I would turn off the Einstein gpu tasks and run Einstein cpu tasks instead and use an app_config.xml file to limit Milky Way and Einstein to 1 task at a time each. |
Send message Joined: 2 Mar 20 Posts: 131 Credit: 320,573,149 RAC: 15,761 |
Seems to me that n body tasks are too big to run. Mine usually take about 2 days and often die before finishing; Is this the nature of the beast?Are you using less than 100% of CPU time? Yes, I know that the same chip is the cpu and the gpu, but they are two components and should be able to run independently to a point. I've not found the Einstein cpu task to be much better time wise than the Milkyway, so I just wanted to see if I could get one of them to work with the cpu. Seems that they are both very demanding task wise. Thanks for the input. I will continue to see what I can get to work, one way or the other. Wish separation tasks wouldn't have ended, it's giving me a headache. Allen |
Send message Joined: 2 Mar 20 Posts: 131 Credit: 320,573,149 RAC: 15,761 |
Mikey, Can you show me how to make Einstein run only one unit with all 4 cores? So far, I've got 4 wu's running on 4 cores and each should take 21 hours. Unfortunately, I have been running them for over 10 minutes and it has only taken 10 minutes off the est. time. Thanks, Allen |
Send message Joined: 8 May 09 Posts: 3339 Credit: 524,010,781 RAC: 0 |
Mikey, No because they are not multi-threaded tasks, you might try running only 3 tasks at a time, 75% of your cpu's, and see if they run faster. |
Send message Joined: 2 Mar 20 Posts: 131 Credit: 320,573,149 RAC: 15,761 |
Thanks, I'll try that, but I kinda doubt it will work. Let you know if it does though. Allen |
Send message Joined: 2 Mar 20 Posts: 131 Credit: 320,573,149 RAC: 15,761 |
I can hardly believe this. I stopped two of the wu's and even though they show stopped, their time running and their estimated time continue keeping pace with the ones that are still running. |
Send message Joined: 8 May 09 Posts: 3339 Credit: 524,010,781 RAC: 0 |
I can hardly believe this. I stopped two of the wu's and even though they show stopped, their time running and their estimated time continue keeping pace with the ones that are still running. I don't think I've ever seen that before, is it still doing it? How did you stop them? When I stop a task I suspend it and it stops, some of my pc's are maxed out more than others so it can take more than a few seconds to recognize I've suspended them but never as long as a minute. |
Send message Joined: 2 Mar 20 Posts: 131 Credit: 320,573,149 RAC: 15,761 |
I can hardly believe this. I stopped two of the wu's and even though they show stopped, their time running and their estimated time continue keeping pace with the ones that are still running. I stopped them by setting the run preference to 50%. They are all still showing that they are running, even though 2 of them say they are not. Really weird. I'm pretty sure that if I stopped them with suspend that they would really stop. Shows they have been running for almost 20 hours, with about an hour to go. As expected, suspend does shift all computing to the still running cores. |
Send message Joined: 8 May 09 Posts: 3339 Credit: 524,010,781 RAC: 0 |
I can hardly believe this. I stopped two of the wu's and even though they show stopped, their time running and their estimated time continue keeping pace with the ones that are still running. Yes Suspend is the way to go to save the resources |
Send message Joined: 24 Apr 09 Posts: 13 Credit: 50,337,909 RAC: 0 |
This was posted in the OS-specific board but no answer showed: why is there no Mac version? Another option would be to Docker it, as I know BOINC supports that (Cosmology@home has used docker for years, though that project now seems kaput). Should be straightforward for a CPU-only application. -Marty Better is the enemy of the good. -Voltaire Never ascribe to malice that which is adequately explained by incompetence. - Napoleon |
Send message Joined: 3 Mar 13 Posts: 84 Credit: 779,527,712 RAC: 0 |
Cosmology@home has used docker for years, though that project now seems kaput Cosmology@home is not dead [yet] I still get work from it , though if you want to post on the forum or alter account settings you/we have to log in via console/commandline [press F12] workaround , other than that its a boincless zombie . |
Send message Joined: 8 May 09 Posts: 3339 Credit: 524,010,781 RAC: 0 |
Cosmology@home has used docker for years, though that project now seems kaput Lots of people still get work from Cosmology and yes the Admins seem to be having a very difficult time keeping the website up and running on a consistent basis. |
Send message Joined: 13 Mar 23 Posts: 1 Credit: 301,133 RAC: 0 |
hi, not receiving any work units from N-Body. Separation were last units received. how do I get the n-body units? |
Send message Joined: 8 May 09 Posts: 3339 Credit: 524,010,781 RAC: 0 |
hi, As long as you were not running gpu tasks nothing at all, this page https://milkyway.cs.rpi.edu/milkyway/prefs.php?subset=project shows that there is no selection for Separation tasks anymore just the NBody tasks. |
©2024 Astroinformatics Group