Rosetta@home BOINCstats Challenge - Rosetta@home Against Corona!

Nick Name

Administrator
USA team member
Just downloaded 1000 WUs from Ralph@Home. Testing new Rosetta 4.14 program. New work doesn't prempt WCG. As the SCC work ends, the Ralph work starts
I just saw on the Twitter that Rosetta is requesting help on Ralph to test this so I'm rolling over too. If there will still be work when I run out this WCG queue.
 

DrBob

Administrator
USA team member
I haven't had any problems keeping 2-3 days Rosetta on all my boxen. :unsure:
Guess I'm just lucky. :whistle:

CRUNCH ON! :USA:
 

BeauZaux

Active Member
USA team member
Got some Rosetta tasks on one my 2 core boxes and it did set aside WCG tasks. Maybe it was the swap setting, though I'm not totally clear what that does. So in the mean time, i've suspended WCG projects...for the next 7 hours, argh. As I type a 4 core box got 2 Rosetta (16hr) tasks and pushed 2 WCG tasks aside. Now running 2 cores on Rosetta and 2 on WCG. Third 2 core box got 2 tasks but sent them back completed without ever starting. argh
 

doneske

Well-Known Member
USA team member
I haven't received any additional Ralph units in over an hour but I'm filling up with work on R@H. Was down to about 500 WUs and now back up over 1200. Now I'm trying to drain the last of the WCG work and work through the last 800 of the Ralph work. Should have enough for a least a day or 2. SCC (WCG) on one box was suspended because the Ralph work had a shorter deadline but FAH kept running. Almost all of the WCG is gone now.
 

Nick Name

Administrator
USA team member
Yep, it looks like they added more work. Currently 14,970 unsent tasks for Rosetta main app.
Some of that is probably for Android, nice to see that's available after a lot of folks asking for it in recent months.

I got a fair amount of work sometime this afternoon, all machines are full now.
 

NorthAlabamaCharitable

Member
USA team member
Some of that is probably for Android, nice to see that's available after a lot of folks asking for it in recent months.

I got a fair amount of work sometime this afternoon, all machines are full now.
Below is the current server status (from here):

Tasks by application

ApplicationUnsentIn progressRuntime of last 100 tasks in hours: average, min, maxUsers in last 24 hours
Rosetta149077771687.94 (0.12 - 148)15337
Rosetta Mini0349399.86 (0.16 - 98.64)5305
Rosetta for Portable Devices14998255797.34 (0.13 - 50.41)1215
 

Nick Name

Administrator
USA team member
Looks like an interesting battle between Doneske and NorthAlabamaCharitable for top team rank in this one. :woot:

I don't have hard numbers but my impression is, longer run times are penalizing credit. I changed my selection from 4 hours to Not Selected, run times seem to have settled at about 8 hours, which is what they said they were going back to. I'll have a better idea tomorrow if this is true or not since I shouldn't have any work shortage.
 

BeauZaux

Active Member
USA team member
Two questions:
1. Not Selected defaults to 8 hours. I'm getting 15 hour tasks. Won't selecting 4 hour tasks be better. Of course I take the chance of getting no tasks at all.
2. If 1 above is correct, would it be a good idea to abort my 15 hour ready to run tasks. I may first suspend them to see if i get something better.
Thanks
 

Nick Name

Administrator
USA team member
Two questions:
1. Not Selected defaults to 8 hours. I'm getting 15 hour tasks. Won't selecting 4 hour tasks be better. Of course I take the chance of getting no tasks at all.
2. If 1 above is correct, would it be a good idea to abort my 15 hour ready to run tasks. I may first suspend them to see if i get something better.
Thanks
Any work that's already in your cache will have the settings from when it was downloaded. If you had no time selected and the project was using 16 hours, that might explain what's happening.

I wouldn't abort unless you know for sure you won't be able to complete it before the deadline. If your tasks are running that long and they should have the current time setting, then yes I'd specify 4 hours or maybe even 2. Running that far past sounds like your machine might be a bit overloaded.
 

doneske

Well-Known Member
USA team member
Witha little bit of a push, we could get into 4th place again. A little bigger push and we could take 3rd again. There seems to be enough work to keep the machines busy.
 

doneske

Well-Known Member
USA team member
I have been noticing that it takes a bit of time for work to validate. There are times when I have a constant queue of 5 to 20 tasks waiting for validation. The server page shows a consistent 17000 to 20000 tasks waiting for validation. They also have a consistent 600,000 to 700,000 waiting for assimilation. I don't know what they are doing during assimilation but it's consistently backed up. They should look into starting a few more assimilation tasks.
 

BeauZaux

Active Member
USA team member
Any work that's already in your cache will have the settings from when it was downloaded. If you had no time selected and the project was using 16 hours, that might explain what's happening.

I wouldn't abort unless you know for sure you won't be able to complete it before the deadline. If your tasks are running that long and they should have the current time setting, then yes I'd specify 4 hours or maybe even 2. Running that far past sounds like your machine might be a bit overloaded.
Yeah, this is running on some pretty slow machines. Set to 4 hour and then to 2. Now got 4 to 10 hour tasks. I'll save that 15 hour task for when a dry spell appears again.
Thanks
 

Nick Name

Administrator
USA team member
Stats have been a little wonky, probably just the overall load on the server is slowing things down. zombie67 from SUSA mentioned a similar problem a couple days ago.
 
Top