Posts by Sabrina Tarson

1) Message boards : Number crunching : Default workunit preferred runtime increases to 16 hours. (Message 93127)
Posted 2 Apr 2020 by Sabrina Tarson
Post:
So I'm confused, should we just disregard the first post in this thread, seeing as it's been rolled back to 8 hours?
2) Message boards : Number crunching : COVID 19 WU Errors (Message 92586)
Posted 30 Mar 2020 by Sabrina Tarson
Post:
Is there any possibility this could be from the fact that Rosetta is a 32-bit application, but some COVID-19 WUs (the 2GB ones) could be trying to request even more memory than normal (over 4GB), resulting in the Out Of Memory error?

I ask this because I have a 32 core Threadripper that has 64 GB of ram, and I share these resources between World Community Grid, LHC@Home, and Rosetta@home, meaning it's rare that all 32 cores are crunching Rosetta@home at the same time. However, even in the worse case scenario of running 32 Rosetta tasks at once, with the 2GB per WU rule I would still meet the RAM requirement, not to mention that worse case there's around 10GB of Swap space.

This thought comes up as all the errored COVID-19 work units on my machines all errored out on other machines as well.

Just a hunch.
3) Message boards : Number crunching : Max # Jobs (Message 91627)
Posted 29 Jan 2020 by Sabrina Tarson
Post:
Is there any word from the Rosetta@Home staff on when they will be upgrading the BOINC Server backend to allow for the Max # Jobs parameter that many other projects have? I would love to be able to set certain computers to only download 1 or 2 WUs as performance grinds to a halt when many Rosetta tasks are L3 cache constrained.

This missing feature is preventing me from adding Rosetta@Home to my grid.

I'm aware you can achieve a similar outcome using "project_max_concurrent" but this messes with the work fetch algorithm and generally causes issues so this isn't an option for me.
4) Message boards : Number crunching : MAC OS X 6.12.35 Errors (Message 73751)
Posted 3 Sep 2012 by Sabrina Tarson
Post:
Im not talking about tasks that took 16 days, I mean after the computer running 24/7 for 16 Days crunching Rosetta@home WU, the problem came back.

A re-boot fixed it, but I don't want to restart every 16 days, its suppose to be headless and require no more input from me when running.

Those logs were from the Console for Mac OS X Snow Leopard.

BOINC is the process running, and so far those are the only things that are outputted in the Console. I have since restarted, and the problem has not come back, however I susepct, I will be posting here again in another 16 days.
5) Message boards : Number crunching : MAC OS X 6.12.35 Errors (Message 73745)
Posted 3 Sep 2012 by Sabrina Tarson
Post:
After a restart, the problem went away. However, after around 16 days, the problem came back. Any news on this so far?
6) Message boards : Number crunching : MAC OS X 6.12.35 Errors (Message 73647)
Posted 18 Aug 2012 by Sabrina Tarson
Post:
I woke up today, and on the Mac Mini that I crunch Rosetta@home on 24/7, I noticed that only one of the two processors were actually doing work. The other processor was stuck doing a syslogd operation and mDNSResponder operation.

I took a look at the console, and I keep getting these errors over and over back to back.

Console:
new-host3 /Applications/BOINCManager.app/Contents/Resources/boinc[7060]: *** process 7060 exceeded 500 log message per second limit - remaining messages per second discarded ***

new-host3 /Applications/BOINCManager.app/Contents/Resources/boinc[7060]: dnssd_clientstub deliver_request: socketpair failed 24 (Too many open files)

and then repeat.

Any knowledge why this is happening?
Also, I am running BOINC 6 because it is the recommended version for Rosetta@home.
7) Message boards : Number crunching : Alot of Errors on Frizzled Workunits (Message 73568)
Posted 4 Aug 2012 by Sabrina Tarson
Post:
BOINC 6 is the most optimised version for Rosetta@home. Rosetta@home was not coded for BOINC 7, and can come across some error running it.
8) Message boards : Number crunching : Alot of Errors on Frizzled Workunits (Message 73565)
Posted 2 Aug 2012 by Sabrina Tarson
Post:
Looking at my tasks, almost all Workunits I get that start with Frizzled have errors. Its not just me either, as the other computer that has crunched the same workunit also got a client error.

Anyone else getting this, and is this a problem with the Workunit. Running the latest BOINC 6 for Mac BTW.
9) Message boards : Number crunching : Does using the Screensaver affect preformace of Rosetta@home? (Message 73523)
Posted 24 Jul 2012 by Sabrina Tarson
Post:
I have recently moved all my Rosetta@home work on to a old Intel Mac Mini I had laying around.

I have a monitor that will display the screensaver because I think it is really neat when you walk into your room, you can see your computer participating in science.

My question is, does using this screensaver take cycle away from the project, and if I want to get the most out of my computational time, should I turn off the screensaver?

Thanks in advance!
10) Message boards : Number crunching : No new Rosetta@home tasks? (Message 73498)
Posted 19 Jul 2012 by Sabrina Tarson
Post:
Alright, thanks Link. I'll leave that Update button alone then.

Thanks everyone for the help!
11) Message boards : Number crunching : No new Rosetta@home tasks? (Message 73495)
Posted 19 Jul 2012 by Sabrina Tarson
Post:
How do I ask it to not get GPU work for Rosetta@home?
12) Message boards : Number crunching : No new Rosetta@home tasks? (Message 73493)
Posted 19 Jul 2012 by Sabrina Tarson
Post:
So, will it just fix itself eventually then?
13) Message boards : Number crunching : No new Rosetta@home tasks? (Message 73491)
Posted 19 Jul 2012 by Sabrina Tarson
Post:
I don't think you guys are understanding me.

I have 3 projects. World Community Grid, Rosetta@home, and Milkyway@home.

World Community Grid and Rosetta@home use CPU, and Milkyway@home uses GPU.

From what you guys have posted, the cc_config.xml fix that Polian posted will not allow any of my BOINC projects to use the GPU, but Milkway@home does.

Link, has said that the cc_config.xml that Polian posted will not allow ANY BOINC projects to use the GPU.

So now I am asking, how do I fix the issue of Rosetta@home trying to use the GPU, without denying Milkyway@home the use of the GPU.
14) Message boards : Number crunching : No new Rosetta@home tasks? (Message 73489)
Posted 19 Jul 2012 by Sabrina Tarson
Post:
Then how do I fix the issue? Or do I just drop Rosetta@home?


There is no issue to be honest. Rosetta is happily churning away using your CPU. GPU work is for your video card. Rosetta doesn't have GPU work.

Yes, but I have another project that uses my GPU. And from what I have gathered here, the solution that you have posted to use for the cc_config.xml will not let my Milkway@Home project use my GPU.
15) Message boards : Number crunching : No new Rosetta@home tasks? (Message 73487)
Posted 19 Jul 2012 by Sabrina Tarson
Post:
Then how do I fix the issue? Or do I just drop Rosetta@home?
16) Message boards : Number crunching : No new Rosetta@home tasks? (Message 73485)
Posted 19 Jul 2012 by Sabrina Tarson
Post:
Now, will the cc_config.xml stop requesting GPU work for even my Milkyway@home program? Because Milkyway@home does do GPU work.
17) Message boards : Number crunching : No new Rosetta@home tasks? (Message 73483)
Posted 19 Jul 2012 by Sabrina Tarson
Post:
There is alot of this in the event log:

"7/18/2012 3:37:30 PM | rosetta@home | update requested by user
7/18/2012 3:37:34 PM | rosetta@home | Sending scheduler request: Requested by user.
7/18/2012 3:37:34 PM | rosetta@home | Requesting new tasks for ATI
7/18/2012 3:37:35 PM | rosetta@home | Scheduler request completed: got 0 new tasks
"

18) Message boards : Number crunching : No new Rosetta@home tasks? (Message 73476)
Posted 19 Jul 2012 by Sabrina Tarson
Post:
I recently added World Community Grid and Einstein@home to my BOINC manager, and now Rosetta@home does not download new tasks.

I tried flushing all of the Rosetta@home files and re-downloading them, but Rosetta@home will not even download the files again.

Can someone please help me out on this issue?
19) Message boards : Number crunching : Am I even contributing? (Message 73427)
Posted 11 Jul 2012 by Sabrina Tarson
Post:
I did the math.

I have a Quad Core Stock Clocked i5 3570K. It takes one core 3 hours to do one Rosetta task. On the weekends, the computer runs from 10:05 PM to 6:00 AM. Doing the math, during those hours, I only do 10 tasks. On the weekdays, It will be doing tasks from 9:05 PM to 2:00 PM the next day, and with that math I will only be doing 21 tasks.

With the above math, am I even contributing to the project?

Edit: Sorry wrong category.
20) Message boards : Number crunching : Workunit only running at 25% CPU (Message 73352)
Posted 27 Jun 2012 by Sabrina Tarson
Post:
I am currently running only one Workunit as I am about to go on a trip and don't want to fill up on Rosetta as I won't be here nor will the computer be crunching it.

However, for this workunit, using task manager, I am seeing that I am only using 25% of my CPU. Why am I not running at a full 100% CPU load?






©2024 University of Washington
https://www.bakerlab.org