N-body sim processing time issues

I have several N-Body sim tasks on one of my PCs (core 2 quad). All require 3 CPUs, when it comes time for them to process, the roughly 3 hour task turns into "elapsed - ~8:30:00 ; remaining (est) - 23 days x hours, etc" Which would put the task (well) past the deadline.
If I suspend all processing and restart BOINC, the Elapsed time drops (usually significantly) and the remaining estimate goes back to a reasonable duration, only to hyper-inflate once the task start running again.

Is there anything I can do or should I just bail on these tasks? And if no other option; how can I prevent this one PC from getting them again but leave my other PC able to get them?

I have several N-Body sim tasks on one of my PCs (core 2 quad). All require 3 CPUs, when it comes time for them to process, the roughly 3 hour task turns into "elapsed - ~8:30:00 ; remaining (est) - 23 days x hours, etc" Which would put the task (well) past the deadline.
If I suspend all processing and restart BOINC, the Elapsed time drops (usually significantly) and the remaining estimate goes back to a reasonable duration, only to hyper-inflate once the task start running again.

Is there anything I can do or should I just bail on these tasks? And if no other option; how can I prevent this one PC from getting them again but leave my other PC able to get them?

!!!There is a way to stop N-Body tasks but it will stop them on every computer listed with your account. You can make a new account for restricting just to that computer, BUT, probably there are too many background programs running if you use Windows 10. By terminating all the junk and unused apps in your app page that are not nessessary for proper running and for BOINC you may find your core 2 quad has new life. Be careful not to terminate an app you don't know what it does, it may be needed.


I go to Windows settings and turn off all apps in privacy. Select each windows permission and app permission one at a time in the left menu and turn them all off.

Then select APPS in windows settings and pick every app you know you don't use and after clicking the app select advanced options and click "terminate". One in particular that helps the most is terminating Microsoft Photos. Terminate all the games, 3-d viewer, camera, feedback hub, get help, groove music, mail and calendar, maps, messaging, Microsoft photos, store, mixed reality portal, mobile plans, movies and tv, one note, Paint 3-d, people, print 3-d, snip n sketch, sticky notes, all three x-boxes, phone.

This will clear up a huge amount of background stuff and give N-BODY a chance to run. Don't bother uninstalling them because Microsoft will just re-download them. Terminating them solves the background apps hogging compute power.

Hope this helps you.

---

100% POWERED BY SOLAR, SUNNY REGARDS

I've already pruned all of that stuff. right now the PC is set to have 75% of processors to handle BOINC tasks (for 90% of time), leaving 1 core to handle the Windows processes. As it is, it's currently running at ~72% of CPU utilization working on one of the 3-CPU tasks.

However the ~3 hour tasks inflate to days worth of processing.

I looked at your computers and tasks. It looks like you just joined this computer?

Sounds like you are familiar with the resource monitor. Are there any tasks besides milkyway that are constantly running in the CPU column? The next to last column on the right? And in the CPU graph, is there a steady CPU usage graph or does it look like a heartbeat?

Are you using any temperature apps?

I would try using only 52% cpu just to see if itl run 2 processor tasks better, for a test. Have you tried already?

Do you use anything like AVG Tuneup? And have you updated all your drivers with something like AVG Driver Updater?

You could run a free hardware test like UserBenchMark if not already tried. With BOINC Manager suspended.

Just thinking outloud;-)

---

100% POWERED BY SOLAR, SUNNY REGARDS

It's a computer that has been on previously. It's just been thru a move and a couple reformats (and renames). It runs single core tasks fine, it's been steadily crunching away at tasks for Milkyway, Rosetta, and Asteroids @home since I reinstalled BOINC. It just seems hit or miss on the 3-CPU processes.

Since the deadline was looming and advice was rolling in slow, I abandoned the tasks so that they go back into the pool for crunching/validation.

Perfect time for trying 52%. I only open up five processors for N-body tasks with 1 to 3 GPU's running too, with an i7 four core 8 processor. the tasks get done from 55 minutes to 2 hours 45 minutes. Some are bigger than others.

---

100% POWERED BY SOLAR, SUNNY REGARDS

Then if the GPU tasks run out I sometimes get a few 6 core N-body tasks and they run OK but a little slower.

---

100% POWERED BY SOLAR, SUNNY REGARDS

Hoolabis you could upgrade to an X3370 CPU if all else fails. $42. Several used for sale on ebay.

---

100% POWERED BY SOLAR, SUNNY REGARDS

I have Ryzen 3900x with 64 GB ram, I just had n-body use 16 cores for 1 day 18 hrs with 10 hrs remaining. After reading through some posts, I shut down bionic, restarted it. it now says that it has been running 58 mins with 11 mins left to finish but it is waiting to run. if this keeps up I will remove this program as I am the one paying the electric bill, I feel that if you cant keep this corrected, I will not spend my money on your messed up WU's.

I have Ryzen 3900x with 64 GB ram, I just had n-body use 16 cores for 1 day 18 hrs with 10 hrs remaining. After reading through some posts, I shut down bionic, restarted it. it now says that it has been running 58 mins with 11 mins left to finish but it is waiting to run. if this keeps up I will remove this program as I am the one paying the electric bill, I feel that if you cant keep this corrected, I will not spend my money on your messed up WU's.

Just go into the Boinc Manager and click Suspend on the MilkyWay Project, that way it will stop what it's doing and not get any new tasks. You can then abort all the workunits, click update on the MilkyWay Project and they will be gone and your pc is yours again.

The n-body workunits use every cpu core you allow Boinc to use and will wait for them to be free if they aren't all free. So apparently you have allowed Boinc to use 16 cpu cores so that's what the n-body tasks are trying to do. If you want it to use less cpu cores you can use an app_config file but if you don't know how it's not exactly a plug and play thing but here's an app_config file that tells MilkyWay to only use 2 cpu cores on the n-body workunits.

<app_config>

<app>
<name>milkyway</name>
<gpu_versions>
<gpu_usage>0.5</gpu_usage>
<cpu_usage>0.5</cpu_usage>
</gpu_versions>
</app>
<app_version>
<app_name>milkyway_nbody</app_name>
<max_concurrent>1</max_concurrent>
<plan_class>mt</plan_class>
<avg_ncpus>2</avg_ncpus>
<cmdline>--nthreads 2</cmdline>
</app_version>
</app_config>

Fewer cpu cores means longer tasks so you will have to play with it to find a sweet spot, but any tasks you have on your pc will be crunched using the old settings and only NEW tasks will use any changes.

I have Ryzen 3900x with 64 GB ram, I just had n-body use 16 cores for 1 day 18 hrs with 10 hrs remaining. After reading through some posts, I shut down bionic, restarted it. it now says that it has been running 58 mins with 11 mins left to finish but it is waiting to run. if this keeps up I will remove this program as I am the one paying the electric bill, I feel that if you cant keep this corrected, I will not spend my money on your messed up WU's.

Just go into the Boinc Manager and click Suspend on the MilkyWay Project, that way it will stop what it's doing and not get any new tasks. You can then abort all the workunits, click update on the MilkyWay Project and they will be gone and your pc is yours again.

The n-body workunits use every cpu core you allow Boinc to use and will wait for them to be free if they aren't all free. So apparently you have allowed Boinc to use 16 cpu cores so that's what the n-body tasks are trying to do. If you want it to use less cpu cores you can use an app_config file but if you don't know how it's not exactly a plug and play thing but here's an app_config file that tells MilkyWay to only use 2 cpu cores on the n-body workunits.

<app_config>

<app>
<name>milkyway</name>
<gpu_versions>
<gpu_usage>0.5</gpu_usage>
<cpu_usage>0.5</cpu_usage>
</gpu_versions>
</app>
<app_version>
<app_name>milkyway_nbody</app_name>
<max_concurrent>1</max_concurrent>
<plan_class>mt</plan_class>
<avg_ncpus>2</avg_ncpus>
<cmdline>--nthreads 2</cmdline>
</app_version>
</app_config>

Fewer cpu cores means longer tasks so you will have to play with it to find a sweet spot, but any tasks you have on your pc will be crunched using the old settings and only NEW tasks will use any changes.

Mikey, I appreciate that you are offering to help, but why are you posting my app_config from this thread that is having problems? At least trim out the section for the Milkyway Separation tasks that are using 1/2 a CPU and GPU, it isn't relevant to this particular subject.

I have Ryzen 3900x with 64 GB ram, I just had n-body use 16 cores for 1 day 18 hrs with 10 hrs remaining. After reading through some posts, I shut down bionic, restarted it. it now says that it has been running 58 mins with 11 mins left to finish but it is waiting to run. if this keeps up I will remove this program as I am the one paying the electric bill, I feel that if you cant keep this corrected, I will not spend my money on your messed up WU's.

Just go into the Boinc Manager and click Suspend on the MilkyWay Project, that way it will stop what it's doing and not get any new tasks. You can then abort all the workunits, click update on the MilkyWay Project and they will be gone and your pc is yours again.

The n-body workunits use every cpu core you allow Boinc to use and will wait for them to be free if they aren't all free. So apparently you have allowed Boinc to use 16 cpu cores so that's what the n-body tasks are trying to do. If you want it to use less cpu cores you can use an app_config file but if you don't know how it's not exactly a plug and play thing but here's an app_config file that tells MilkyWay to only use 2 cpu cores on the n-body workunits.

<app_config>

<app>
<name>milkyway</name>
<gpu_versions>
<gpu_usage>0.5</gpu_usage>
<cpu_usage>0.5</cpu_usage>
</gpu_versions>
</app>
<app_version>
<app_name>milkyway_nbody</app_name>
<max_concurrent>1</max_concurrent>
<plan_class>mt</plan_class>
<avg_ncpus>2</avg_ncpus>
<cmdline>--nthreads 2</cmdline>
</app_version>
</app_config>

Fewer cpu cores means longer tasks so you will have to play with it to find a sweet spot, but any tasks you have on your pc will be crunched using the old settings and only NEW tasks will use any changes.

Mikey, I appreciate that you are offering to help, but why are you posting my app_config from this thread that is having problems? At least trim out the section for the Milkyway Separation tasks that are using 1/2 a CPU and GPU, it isn't relevant to this particular subject.

I didn't realize I did that, I copied it from a friend who said it works fine.

I have some WU units that seem abnormal.
Luckily, I was keeping any eye on these tasks that were running, and spotted these fairly early.

I don't want to waste my CPU time running these tasks, unless I know that are OK, and that I get the appropriate credit for it.
So for the moment I will Suspend these particular tasks an monitor the remaining one in my queue.

If I see this happening more frequently, I will have to reluctantly consider abandoning this project.


Here is a real WU example (1)

Before starting
Remaining (estimate): was 00:03:23

After it started
Elapsed: 00:12:09
Remaining (estimate): 32d 08:30:28

Application
Milkyway@home N-Body Simulation 1.76 (mt)
Name
de_nbody_07_29_2020_v176_40k__data__5_1603804502_2740847
State
Task suspended by user
Received
1/12/2020 20:50:42
Report deadline
13/12/2020 20:50:42
Resources
12 CPUs
Estimated computation size
28,595 GFLOPs
CPU time
00:00:50
CPU time since checkpoint
00:00:13
Elapsed time
00:12:09
Estimated time remaining
32d 08:30:28
Fraction done
0.026%
Virtual memory size
13.39 MB
Working set size
16.68 MB
Directory
slots/2
Process ID
10424
Executable
milkyway_nbody_1.76_windows_x86_64__mt.exe

Here is a real WU example (2)

Before starting
Remaining (estimate): was 00:03:23

After it started
Elapsed: 00:04:53
Remaining (estimate): 13d 00:02:55

Application
Milkyway@home N-Body Simulation 1.76 (mt)
Name
de_nbody_07_29_2020_v176_40k__data__5_1603804502_2637435
State
Task suspended by user
Received
1/12/2020 20:50:42
Report deadline
13/12/2020 20:50:42
Resources
12 CPUs
Estimated computation size
28,595 GFLOPs
CPU time
00:26:42
CPU time since checkpoint
00:00:01
Elapsed time
00:04:53
Estimated time remaining
13d 00:02:55
Fraction done
0.026%
Virtual memory size
13.43 MB
Working set size
16.73 MB
Directory
slots/1
Process ID
16876
Progress rate
-3961.080% per hour
Executable
milkyway_nbody_1.76_windows_x86_64__mt.exe