WCG & Rosetta

[Exascaletech]

With the problems at WCG and the over all negative vibe on their site, I'm thinking of adding Rosetta. Looks easy to do since they're both BOINC. Also, what a great site Rosetta has. All the things you look for and can't find on WCG is front and center. I'm sure $$$ and staffing have a lot to do with it. I want to stay in the medical area of DC. I guess you just assign the % of each project. BOINC made it easy from what I've seen. Wasting time with no work from WCG. Any comments?

 

[Markfw]

Rosetta has been almost as bad as WCG, on and off for getting units. Right now its on, but don't be surprised if one day there are no units, and this continues for weeks. Thats the pattern. And right now its almost impossible to keep a 7950x filled , let alone 5 Genoa boxes, including the 96 core one ! Right now I am starving for WCG work. But I have 350 Rosetta right now, and about 50-75 WCG. Also at about 50% of total CPU work capability. (864 total turned on right now)

For GPU I use F@H, and with just 10 cards running I am at about 200 mil ppd. F@H has been pretty good for over 20 years.

 

[mmonnin03]

Rosetta has had work for weeks and literally just ran out today. They had over 5mil jobs with 200-300k done per day. See their front page server status that now says Total Queued Jobs = 0

Rosetta has some things that are better and some that are not vs WCG or other BOINC projects. You can select the run time but its one of the very few where you cannot select one app or another. Their apps are fat and bloated as devs just keep adding to them over the years, at least in the case of 4.x version.

 

[Markfw]

Rosetta has had work for weeks and literally just ran out today. They had over 5mil jobs with 200-300k done per day. See their front page server status that now says Total Queued Jobs = 0

Rosetta has some things that are better and some that are not vs WCG or other BOINC projects. You can select the run time but its one of the very few where you cannot select one app or another. Their apps are fat and bloated as devs just keep adding to them over the years, at least in the case of 4.x version.

exactly my point. for weeks, its OK, now they are out.

 

[Skillz]

Tngrid, WCG, Rosetta, denis@home and Sidock all are medical related BOINC projects.

You can attach to all of them simultaneously and let the BOINC client decide on what to run. While one project or projects might be out of work or difficult to manage one of the others should fill in the idle time.

All of those projects provide CPU work, while some only run in batches every so often and therefore can go days or weeks without any work.

GPUGrid is also a medical project, but these tasks utilize your GPU. There is no CPU only tasks, though some of them in the past have used both simultaneously. WCG also has a sub project that runs on GPUs only now, but those tasks also run out frequently.

 

[Exascaletech]

Yeah my quest at the start of the year to find a DC project I could fall in love with again has been a road of potholes. And yep, sure enough, just got it added and Rosetta is out also. I'm batting zero. Didn't use to be like this and I wonder what the real reason is. Maybe really no work, almost out of money, DC fatigue from the projects or a combination. Even spent a couple hours tonight doing yet another general search of DC, thought Rosetta had more success than most so I read all the news releases they highlight and most are pre-2017 and many from 2006-2012. The road to keep support in DC mainly seems to be the points chasers with big machines. Looks like they'll need it to make up for the ever declining interest.

 

[pututu]

Among the CPU bio-medical related projects (TNgrid, WCG, Rosetta, DENIS, and Sidock), Sidock has consistent supply of tasks. Other than maintenance downtime (doesn't happen often), I have not seen any issue on Sidock task availability, at least since the beginning of this year when I started running it. I just stopped running GPU WCG OPNG sub-project since they just ran out of gpu work yesterday. Maybe I'll move back over to gpugrid for my gpu crunching.

 

[StefanR5R]

Rosetta@Home and quite many other projects only intermittently supply work because scientists prepare specific numerical experiments, the admins put these on the BOINC server, volunteers complete them eventually, but the next experiments may not yet be rife for their computational stage at that point.

Folding@Home also executes experiments of limited size. But they always have so many experiments going on at once that there is never a period during which there was no work flowing to volunteers.

Other projects which practically never run out of work are mostly the mathematical projects which are conducted by private scientists rather than universities or corporations, and which have huge or even infinite "experiments".

For most BOINC projects, @Kiska is running a site which tracks uptime and tasks-ready-to-send and a lot of other server status:
https://grafana.kiska.pw/d/boinc/boinc?orgId=1&var-project=Rosetta@Home
At the top right, you can switch further back into the history of recorded data.