Hello to the community.

Is there a way to limit the number of work units that received when PC connects to your server?
I would like to receive 4-5 work units when I request job and not 10 units.

Best
Stratos

Is there a way to limit the number of work units that received when PC connects to your server?
I would like to receive 4-5 work units when I request job and not 10 units.

The amount of work you get will depend on your cache setting, and your Resource Share settings between projects.
It is possible to limit the number of Tasks a particular project will run, but it will involve a huge amount of mucking around over time to get things to work the way you want them to.
The easiest & quickest way to get things to work the way you would like is to set your cache to zero (0 days and 0.01 additional days), and set your Resource Share so that Rosetta has less than your main project, but more than other projects.

With no cache, and depending on how much time BOINC has to process work while the systems are on, and how much time in a day they are on, your Resource share settings should be met within a few weeks. The larger your cache, the more you adjust per Project resources, the longer it will take (as in months).
And depending on the time your system has to process work each day, there may be times where you will have more (or considerably less) Rosetta Tasks running in order to meet your Resource Share settings. But over time they should settle down to what you are aiming for.
Of course if one project runs out of work for a while, then the other projects will process more. When the one that was out of work has work again, the other projects will get less work done & the recovered one will get extra done until the debt owed has been repaid.

---

Grant
Darwin NT

Hello to the community.

Is there a way to limit the number of work units that received when PC connects to your server?
I would like to receive 4-5 work units when I request job and not 10 units.

Best
Stratos

A follow up to this question are you asking because you only want to RUN 4 to 5 Rosetta workunits at one time? Or that you actually only want to RECEIVE 4 to 5 workunits at a time? Because if it's receive as you stated Grant gave you the correct answer.

Slightly related, I too am not happy with Rosetta scheduling. I've got my Rosetta resource share set to 1 and WCG set to 99 on an eight core laptop.

My PC will be happily running WCG for days at a time when WHAM. Rosetta decides it needs five, 8-hours tasks done in three days!

I know the standard answer is "Rosetta will balance itself out over time", but sheesh, the algorithm used is whacked. I don't run my laptop 24x7 and can anticipate that there's no way to complete the tasks in the requested time, so I abort them to make some use of my extra CPU cycles...

Rant over, no need to reply as I know the standard answer, but the Rosetta scheduler has problems...

I know the standard answer is "Rosetta will balance itself out over time", but sheesh, the algorithm used is whacked. I don't run my laptop 24x7 and can anticipate that there's no way to complete the tasks in the requested time, so I abort them to make some use of my extra CPU cycles...

Rant over, no need to reply as I know the standard answer, but the Rosetta scheduler has problems...

Of course the Scheduler will have issues- because you keep fiddling with things and don't let them run their course.
Set your cache to 0 days + 0.1S additional days and don't abort the Tasks. Let them miss the deadlines & it will figure things out. As long as you micro manage things, the Scheduler has no chance of sorting things out.

If you just want Rosetta as a backup project, set it's Resource share to 0. That way it should only run if there is no work for your other project (i'm not sure if Rosetta's server recognises the zero setting. Some Projects do, some don't).

---

Grant
Darwin NT

Just so it's clear, BOINC does the scheduling, rather than Rosetta.

---

Slightly related, I too am not happy with Rosetta scheduling. I've got my Rosetta resource share set to 1 and WCG set to 99 on an eight core laptop.

My PC will be happily running WCG for days at a time when WHAM. Rosetta decides it needs five, 8-hours tasks done in three days!

I know the standard answer is "Rosetta will balance itself out over time", but sheesh, the algorithm used is whacked. I don't run my laptop 24x7 and can anticipate that there's no way to complete the tasks in the requested time, so I abort them to make some use of my extra CPU cycles...

Rant over, no need to reply as I know the standard answer, but the Rosetta scheduler has problems...

Add this to a file called app_config.xml in the Rosetta directory under c:...Boincboinc.bakerlab.org_rosetta

<app_config>
<project_max_concurrent>1</project_max_concurrent>
</app_config>

do the same thing in the c:...boinc/projects/www.worldcommunitygrid.org directory except limit it to 7 tasks instead of 1 and you will ALWAYS run 1 Rosetta task and 7 WCG tasks as long as each has tasks available. The problem is if you forget that you have this in place you will be asking for help here in the forums about WHY you are only getting 1 or 7 tasks instead of the 8 you want to run.

In order to clarify what t I see happening in the Event Log, could I please have some definitions:
!. What is a task,?
and,
2, What is a work unit?
And, in order to predict when Rosetta (BOINC) will finish a '"task", how can I calculate, from the times shown for receipt and delivery, how much time remains for the tasks in progress?
This may assist in avoiding further rejection of timed out units/tasks. The rejection appears to be caused by excess fetching of "tasks",
Thank you.

---

!. What is a task,?
and,
2, What is a work unit?

They are the same thing- the work you process for the project, If you select the Advanced view in the BOINC Manager, they are the Tasks in the Tasks tab. There it will show how much work has been done, and how much time is left to go.

The default Target CPU time for a Task is 8 hours (some may finish sooner, others may run for longer- but 90% or more will run for the Target CPU time set).
You have set yours to run for over a day- i strongly recommend you set it back to 8 hours. Also your system looks as though it is busy trying to do other things at the same time as is processes work, making the Tasks take longer to complete- it took your system 39.5 hours to do 36 hours worth of processing.
Set your cache to 0 days & 0.01 additional days.

That will limit the amount of work the system downloads, reduce the time it takes to complete a Task, and will let the BOINC Manager work out much sooner just how much time the system is on, and when it is on how much time is available to process BOINC work & so no more missed deadlines.
There's no need for the app_config.xml settings to sort things out at this stage.

---

Grant
Darwin NT

Grant
Thank you for the clarification.
I am worried about wasted cpu time and lost projects, and tried disallowing work fetch for 36 hours, which seemed to be the critical period for getting current tasks completed on time. Here's what happened, saving me a lot of readjustment of the rules:

2021-08-31 3:12:45 PM | Rosetta@home | Sending scheduler request: To fetch work.
2021-08-31 3:12:45 PM | Rosetta@home | Requesting new tasks for CPU
2021-08-31 3:12:46 PM | Rosetta@home | Scheduler request completed: got 2 new tasks
2021-08-31 3:12:46 PM | Rosetta@home | Project requested delay of 31 seconds
2021-08-31 3:12:49 PM | Rosetta@home | Started download of degrader_site_3mup_jhr_bcov_flags2
2021-08-31 3:12:49 PM | Rosetta@home | Started download of degrader_site_3mup_jhr_bcov3_SAVE_ALL_OUT_IGNORE_THE_REST_9al1zp1r.zip
2021-08-31 3:12:50 PM | Rosetta@home | Finished download of degrader_site_3mup_jhr_bcov_flags2
2021-08-31 3:12:50 PM | Rosetta@home | Finished download of degrader_site_3mup_jhr_bcov3_SAVE_ALL_OUT_IGNORE_THE_REST_9al1zp1r.zip
2021-08-31 3:12:50 PM | Rosetta@home | Started download of degrader_site_3mup_jhr_bcov3_SAVE_ALL_OUT_IGNORE_THE_REST_9al1zp1r.flags
2021-08-31 3:12:50 PM | Rosetta@home | Started download of degrader_site_5nvx_jhr_bcov3_SAVE_ALL_OUT_IGNORE_THE_REST_3vd1ve6t.zip
2021-08-31 3:12:51 PM | Rosetta@home | Finished download of degrader_site_3mup_jhr_bcov3_SAVE_ALL_OUT_IGNORE_THE_REST_9al1zp1r.flags
2021-08-31 3:12:51 PM | Rosetta@home | Finished download of degrader_site_5nvx_jhr_bcov3_SAVE_ALL_OUT_IGNORE_THE_REST_3vd1ve6t.zip
2021-08-31 3:12:51 PM | Rosetta@home | Started download of degrader_site_5nvx_jhr_bcov3_SAVE_ALL_OUT_IGNORE_THE_REST_3vd1ve6t.flags
2021-08-31 3:12:53 PM | Rosetta@home | Finished download of degrader_site_5nvx_jhr_bcov3_SAVE_ALL_OUT_IGNORE_THE_REST_3vd1ve6t.flags

Thanks for the prompt reply and asistance. I am not in the geek class here, and always appreciate some helpfrom the experts.
jehop

---

Grant:
Here's what happened after "no new tasks" request,
Too many tasks downloading?


Task
click for details
Show names Work unit
click for details Computer Sent Time reported
or deadline
explain Status Run time
(sec) CPU time
(sec) Credit Application
1420883490 1268449227 4020382 31 Aug 2021, 18:20:40 UTC 3 Sep 2021, 18:20:40 UTC In progress --- --- --- Rosetta v4.20
windows_x86_64
1420883604 1268596264 4020382 31 Aug 2021, 18:20:40 UTC 3 Sep 2021, 18:20:40 UTC In progress --- --- --- Rosetta v4.20
windows_x86_64
1420843893 1268597029 4020382 31 Aug 2021, 18:12:44 UTC 3 Sep 2021, 18:12:44 UTC In progress --- --- --- Rosetta v4.20
windows_x86_64
1420843966 1268597054 4020382 31 Aug 2021, 18:12:44 UTC 3 Sep 2021, 18:12:44 UTC In progress --- --- --- Rosetta v4.20
windows_x86_64
1420843514 1268596763 4020382 31 Aug 2021, 18:12:07 UTC 31 Aug 2021, 18:20:40 UTC Error while computing 253.22 182.31 --- Rosetta v4.20
windows_x86_64
1420843517 1268596765 4020382 31 Aug 2021, 18:12:07 UTC 31 Aug 2021, 18:20:40 UTC Error while computing 260.17 183.45 --- Rosetta v4.20
windows_x86_64
1419620759 1267664636 4020382 29 Aug 2021, 5:15:59 UTC 1 Sep 2021, 5:15:59 UTC In progress --- --- --- Rosetta v4.20
windows_x86_64
1419569358 1267617838 4020382 29 Aug 2021, 2:04:04 UTC 1 Sep 2021, 2:04:04 UTC In progress --- --- --- Rosetta v4.20
windows_x86_64
1418885568 1267027134 4020382 27 Aug 2021, 9:27:55 UTC 30 Aug 2021, 9:28:46 UTC Not started by deadline - canceled 0.00 0.00 --- Rosetta v4.20
windows_x86_64
1418760363 1266910856 4020382 27 Aug 2021, 0:39:25 UTC 30 Aug 2021, 21:50:00 UTC Completed and validated 134,060.80 127,305.80 879.20 Rosetta v4.20
windows_x86_64
1418759367 1266909953 4020382 27 Aug 2021, 0:38:50 UTC 30 Aug 2021, 21:51:28 UTC Completed, too late to validate 134,135.55 127,384.80 0.00 Rosetta v4.20
windows_x86_64

---

In your Rosetta account page-
Preferences
When and how BOINC uses your computer, click on "Computing preferences"
Under Computing, Other

Store at least            0.01  days of work
Store up to an additional 0.01 days of work

Click on Edit preferences to change the values, then on the Update preferences button to save the changes.

Back on the main account page
Preferences
Preferences for this project, click on "Rosetta@home preferences"
Target CPU run time- set to "Not selected" (or select 8 hours). Click on Edit preferences to change the values, then on the Update preferences button to save the changes.

Once you've changed both of those settings, on the BOINC manager on you computer you then need to select Rosetta in the Projects tab, and click on Update for it to get those new settings & apply them (or wait till it next contacts the Scheduler- in this case it's best to update now & not wait).
Changing the cache settings will stop lots of work from down loading. Setting the Target CPU time to the project default will stop you from missing deadlines for any new work it downloads (i'm pretty sure the changed Target CPU time won't be be applied to already running Tasks, but it should be applied to any already downloaded but not running Tasks).

---

Grant
Darwin NT