Problems and Technical Issues with Rosetta@home

Message boards : Number crunching : Problems and Technical Issues with Rosetta@home

To post messages, you must log in.

Previous · 1 . . . 18 · 19 · 20 · 21 · 22 · 23 · 24 . . . 28 · Next

AuthorMessage
Profile Timo
Avatar

Send message
Joined: 9 Jan 12
Posts: 178
Credit: 34,158,374
RAC: 27,156
Message 87913 - Posted: 16 Dec 2017, 0:59:19 UTC - in response to Message 87912.  

Seems like it's now fixed (only about 1 minute after I posted this!) Nice work :)
**38 cores crunching for R@H on behalf of cancercomputer.org - a non-profit supporting High Performance Computing in Cancer Research
ID: 87913 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
xaminmo
Avatar

Send message
Joined: 14 Nov 17
Posts: 2
Credit: 1,553,363
RAC: 0
Message 88212 - Posted: 2 Feb 2018, 1:37:37 UTC

It seems when BOINC requests work, Rosetta sends more than requested.
I do not have this problem with any other project.
This is worse as of the last week or so, but it does not always happen this way.

The main system I notice this on is here:
https://boinc.bakerlab.org/rosetta/show_host_detail.php?hostid=3290081

My runtimes closely match what I've selected in project properties, so it's not that.

If I set resource share to 10, I still pull down about 5x the work of any of the other projects I crunch on.

Deadline is around a week out, so this forces other projects to not run, in favor of Rosetta, because eventually it wants to make sure the jobs are not at risk.

When it uploads the work, it grabs new replacement work, and still favors Rosetta.

I've opened a similar thread in the BOINC forums, since I think they should force the scheduler to comply with user wishes, not with what the project sends.

But I see it as a problem with both sides.

I'm hoping to get a workaround, or some sort of dev committment to help improve this situatoin.
ID: 88212 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Jim1348

Send message
Joined: 19 Jan 06
Posts: 300
Credit: 9,167,631
RAC: 12,049
Message 88213 - Posted: 2 Feb 2018, 2:35:29 UTC - in response to Message 88212.  

If I set resource share to 10, I still pull down about 5x the work of any of the other projects I crunch on.

It takes a while for the work units to adjust to a new resource share setting (often several weeks).

In order to speed up this process, you can use a cc_config.xml file (placed in the BOINC Program Data folder) with the following,
which will speed up the convergence to a day or two:

<cc_config>
  <options>	
  	<rec_half_life_days>1.000000</rec_half_life_days>
  </options>
</cc_config>


Or you can chose any other value. I routinely use it when setting up a new machine.
ID: 88213 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
xaminmo
Avatar

Send message
Joined: 14 Nov 17
Posts: 2
Credit: 1,553,363
RAC: 0
Message 88241 - Posted: 7 Feb 2018, 6:00:21 UTC - in response to Message 88213.  
Last modified: 7 Feb 2018, 6:02:19 UTC

If I set resource share to 10, I still pull down about 5x the work of any of the other projects I crunch on.
It takes a while for the work units to adjust to a new resource share setting (often several weeks).
...
<rec_half_life_days>1.000000</rec_half_life_days>

...
Or you can chose any other value. I routinely use it when setting up a new machine.
That's awesome. I completely missed that, and it looks like exactly what I needed. A++ Kudos, and thank you very much.

-JD
ID: 88241 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Jim1348

Send message
Joined: 19 Jan 06
Posts: 300
Credit: 9,167,631
RAC: 12,049
Message 88242 - Posted: 7 Feb 2018, 12:22:24 UTC - in response to Message 88241.  

You are welcome. It is from Richard Haselgrove, the great BOINC guru.
ID: 88242 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
PFLIEGER Guy

Send message
Joined: 20 Dec 15
Posts: 3
Credit: 1,230,645
RAC: 0
Message 88248 - Posted: 8 Feb 2018, 9:47:33 UTC

Yesterday evening in europe took place a little upgrade of rosetta server because he was one hour along down.

Guy PFLIEGER MASEVAUX ALSACE France
phone: 0033973514697
ID: 88248 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Carpathian

Send message
Joined: 29 Aug 11
Posts: 1
Credit: 49,289
RAC: 1
Message 88253 - Posted: 9 Feb 2018, 13:33:49 UTC

I have in my preferences CPU time of 1 hour. Seems to ignore it and I have just wasted over 18,000 seconds on a job that got a validate error.

Just finishing the rest then I am not processing any more tasks.
ID: 88253 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
sinspin

Send message
Joined: 30 Jan 06
Posts: 26
Credit: 6,355,856
RAC: 0
Message 88254 - Posted: 9 Feb 2018, 17:34:44 UTC

It seems with the calculation of the amount of workunits is someting really wrong.
My default runtime is set to 24 hours. So, Rosetta runs each WU for 24 hours. That works very well. But: Boinc is still to dumb to realize that.
The calculated runtime for new WU's is always around 4 hours. So, i receive tooo much WU's which i have to cancel all the time. Otherwise crunch i only WU's which are ready after deadline ends.
I have no other project running.
Interesting is, i crunch 6 WU's at the same time (6 of 8 CPU cores), 4 hours sets Boinc as runtime for new WU's -> 6*4 = 24
On my other PC is it 4 WU's at the same time and Boinc sets 6 hours for new WU's.
Both Windows, differend Boinc versions. Android is still woking fine.
ID: 88254 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
001ar

Send message
Joined: 8 Feb 18
Posts: 2
Credit: 10,544,319
RAC: 0
Message 88281 - Posted: 13 Feb 2018, 21:07:49 UTC

Hi,

is there any chance for adding support for ARM platforms + Linux to rosetta project? Would like to run it on my Raspberry pi 3, or other ARMs.

This is the error log from the Raspberry running the Ubuntu Linux:

13-Feb-2018 20:42:58 [https://boinc.bakerlab.org/rosetta/] Requesting new tasks for CPU
13-Feb-2018 20:43:01 [Rosetta@home] Scheduler request completed: got 0 new tasks
13-Feb-2018 20:43:01 [Rosetta@home] This project is using an old URL.  When convenient, remove the project, then add http://boinc.bakerlab.org/rosetta/
13-Feb-2018 20:43:01 [Rosetta@home] This project doesn't support computers of type arm-unknown-linux-gnueabihf


Note that this works e.g. for Einstein@Home.
ID: 88281 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Jon Strickland

Send message
Joined: 7 Oct 15
Posts: 1
Credit: 7,593,705
RAC: 1,228
Message 88283 - Posted: 13 Feb 2018, 21:43:09 UTC

Is there a known communication issue with 7.8.6 for Mac? I've got 7.8.4 and 7.83 clients that have contacted the project as recently as 2 hours ago and my two 7.8.6 clients on Mac keep deferring communication.
ID: 88283 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Jim1348

Send message
Joined: 19 Jan 06
Posts: 300
Credit: 9,167,631
RAC: 12,049
Message 88284 - Posted: 13 Feb 2018, 23:39:10 UTC - in response to Message 88283.  

Is there a known communication issue with 7.8.6 for Mac?

There is a new 7.9.1 experimental version out for Mac, but I don't know what it fixes.
http://boinc.berkeley.edu/dl/?C=M;O=D
ID: 88284 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile robertmiles

Send message
Joined: 16 Jun 08
Posts: 689
Credit: 9,340,157
RAC: 4,557
Message 88285 - Posted: 14 Feb 2018, 0:01:58 UTC - in response to Message 88253.  

Are you sure that 1 hour is even an allowed value for CPU time? I haven't checked lately, bu 3 hours used to be the lowest allowed value.
ID: 88285 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
markj

Send message
Joined: 21 Jun 08
Posts: 6
Credit: 15,340,138
RAC: 8,678
Message 88292 - Posted: 14 Feb 2018, 7:50:21 UTC - in response to Message 88284.  
Last modified: 14 Feb 2018, 7:51:09 UTC

Re re: Is there a known communication issue with 7.8.6 for Mac?

Re: There is a new 7.9.1 experimental version out for Mac, but I don't know what it fixes.
http://boinc.berkeley.edu/dl/?C=M;O=D

Have the same problem. 7.9.1 does not fix it.
ID: 88292 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
markj

Send message
Joined: 21 Jun 08
Posts: 6
Credit: 15,340,138
RAC: 8,678
Message 88293 - Posted: 14 Feb 2018, 8:04:44 UTC - in response to Message 88292.  

Re re: Is there a known communication issue with 7.8.6 for Mac?

Re: There is a new 7.9.1 experimental version out for Mac, but I don't know what it fixes.
http://boinc.berkeley.edu/dl/?C=M;O=D

Have the same problem. 7.9.1 does not fix it.


Downgrading to 7.8.4 or to 7.6.22 also does not solve it.
I only run Rosetta, it would be nice if someone could comment about whether other projects are also affected or not.
I suspect it might be a Rosetta server problem, not a Boinc problem.
ID: 88293 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
markj

Send message
Joined: 21 Jun 08
Posts: 6
Credit: 15,340,138
RAC: 8,678
Message 88296 - Posted: 14 Feb 2018, 15:42:54 UTC - in response to Message 88293.  

Whatever the problem was, it seems to have been corrected / disappeared.

Re re: Is there a known communication issue with 7.8.6 for Mac?

Re: There is a new 7.9.1 experimental version out for Mac, but I don't know what it fixes.
http://boinc.berkeley.edu/dl/?C=M;O=D

Have the same problem. 7.9.1 does not fix it.


Downgrading to 7.8.4 or to 7.6.22 also does not solve it.
I only run Rosetta, it would be nice if someone could comment about whether other projects are also affected or not.
I suspect it might be a Rosetta server problem, not a Boinc problem.
ID: 88296 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
oracledude

Send message
Joined: 12 Feb 18
Posts: 5
Credit: 0
RAC: 0
Message 88308 - Posted: 17 Feb 2018, 15:47:45 UTC

Have been mining for over a week but computer # not syncing with profile. My profile show no computers attached, but grcpool is linking to a valid computer...http://boinc.bakerlab.org/rosetta/show_host_detail.php?hostid=3351394

please assist in syncing processing computer with rosetta account, so I can get credit for processing.
ID: 88308 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Sid Celery

Send message
Joined: 11 Feb 08
Posts: 979
Credit: 21,625,433
RAC: 14,372
Message 88314 - Posted: 19 Feb 2018, 3:57:02 UTC - in response to Message 88212.  

It seems when BOINC requests work, Rosetta sends more than requested.
I do not have this problem with any other project.
This is worse as of the last week or so, but it does not always happen this way.

The main system I notice this on is here:
https://boinc.bakerlab.org/rosetta/show_host_detail.php?hostid=3290081

My runtimes closely match what I've selected in project properties, so it's not that.

If I set resource share to 10, I still pull down about 5x the work of any of the other projects I crunch on.

Deadline is around a week out, so this forces other projects to not run, in favor of Rosetta, because eventually it wants to make sure the jobs are not at risk.

When it uploads the work, it grabs new replacement work, and still favors Rosetta.

I've opened a similar thread in the BOINC forums, since I think they should force the scheduler to comply with user wishes, not with what the project sends.

But I see it as a problem with both sides.

I'm hoping to get a workaround, or some sort of dev committment to help improve this situation.

I've just responded to another user who's having problems who has a 1hr runtime set. I have no idea why this is happening, but the default runtime should be 8hrs and Boinc should only be sending one-eighth of the tasks than the 277 that you're showing.

Is this a consequence of some auto setup coming from mining sites or something? Because it's massively wrong and could lead to problems with the website as well as individual users with bad settings.

See my reply here:
ID: 88314 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Sid Celery

Send message
Joined: 11 Feb 08
Posts: 979
Credit: 21,625,433
RAC: 14,372
Message 88315 - Posted: 19 Feb 2018, 4:05:09 UTC - in response to Message 88285.  

Are you sure that 1 hour is even an allowed value for CPU time? I haven't checked lately, but 3 hours used to be the lowest allowed value.

3hrs used to be the default, but 1hr was (and still is) the minimum allowed.

I agree the 1hr option should be removed. And with so many multi-core processors out there, the minimum should probably be 3hrs. 2hrs is also a current option.
ID: 88315 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Sid Celery

Send message
Joined: 11 Feb 08
Posts: 979
Credit: 21,625,433
RAC: 14,372
Message 88316 - Posted: 19 Feb 2018, 4:18:31 UTC - in response to Message 88308.  

Have been mining for over a week but computer # not syncing with profile. My profile show no computers attached, but grcpool is linking to a valid computer... http://boinc.bakerlab.org/rosetta/show_host_detail.php?hostid=3351394

please assist in syncing processing computer with rosetta account, so I can get credit for processing.

The link to the PC you provided is showing credits, but that's not the account you're posting to here which, as you say, shows no credit or attached PC.

I'm probably very naïve or stupid, but I don't understand what you're asking.

You mention "mining" and my eyes roll over, mainly because I don't understand a thing about it. But it seems to me you should be asking whoever you're mining for, because they're getting all the credits right now istm. Is there some other website somewhere which subsequently grants you some share (they claim to be) attributable to you? Ask there. No-one here will ever know a thing about it afaik.

Maybe someone can PM me to tell me what all this mining stuff is all about - before I embarrass myself and others by saying something I may regret #CanYouSayRipOff
ID: 88316 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
mmonnin

Send message
Joined: 2 Jun 16
Posts: 41
Credit: 5,881,525
RAC: 1,075
Message 88319 - Posted: 19 Feb 2018, 19:17:43 UTC - in response to Message 88315.  

Are you sure that 1 hour is even an allowed value for CPU time? I haven't checked lately, but 3 hours used to be the lowest allowed value.

3hrs used to be the default, but 1hr was (and still is) the minimum allowed.

I agree the 1hr option should be removed. And with so many multi-core processors out there, the minimum should probably be 3hrs. 2hrs is also a current option.


Due to the tasks running for 6 hours then having computation errors its better to have the 1 hour task. Then once it reaches 2-3 hours its know to be bad and can be manually aborted instead of wasting a full 6 hours.
ID: 88319 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Previous · 1 . . . 18 · 19 · 20 · 21 · 22 · 23 · 24 . . . 28 · Next

Message boards : Number crunching : Problems and Technical Issues with Rosetta@home



©2019 University of Washington
http://www.bakerlab.org