So people don't need to look back up the thread to find out what "team" we are talking about.

I created the team "SETI orphans" at Rosetta

Feel free to join. Much of their work at present is COVAD19.

We have 9 members already,

Let's see how big we can make it.

Count me in :)

BTW, there are few SETI groups on Rosetta already. Mostly national ones.

EDIT: netbook out of fleet for now: Rosetta needs 979.64MB more disk space. You currently have 736.98 MB available and it needs 1716.61 MB.

---

SETI apps news
We're not gonna fight them. We're gonna transcend them.

Does Rosetta support Android devices?

Yes, but it needs 1.5GB min RAM per Task (Many are only 400MB, but there are some huge ones).

1.5GB? Hmmm, here all my devices tell me they need a minimum of 1.984GB free. Which even on the device with 3GB of working memory isn't available.

My Xiaomi Redmi Note 8 (8 cores) does 7 SETI MB tasks and 1 Rosetta COVID19 task currently.

Good so far.

EDIT: and BOINC reports 3.56GB physical and 2.2GB virtual memory.

host on SETI : https://setiathome.berkeley.edu/show_host_detail.php?hostid=8926115

and on Rosetta: https://boinc.bakerlab.org/rosetta/show_host_detail.php?hostid=4069076

It seems they work with initial replication of 1.No quorum??

---

SETI apps news
We're not gonna fight them. We're gonna transcend them.

EDIT: netbook out of fleet for now: Rosetta needs 979.64MB more disk space. You currently have 736.98 MB available and it needs 1716.61 MB.

Advise, please, most painless way to move BOINCdata folder with projects to another drive.

---

SETI apps news
We're not gonna fight them. We're gonna transcend them.

Advise, please, most painless way to move BOINCdata folder with projects to another drive.

Easiest way I found is to use symbolic links.

1. Move the entire directory or only certain files and sub-directories to another drive (if you want or need to spread it over multiple disks)

2. Place a symbolic link to the new location(s) at the original location(s)

3. BOINC will operate as if it is still at the original location

If you've never done this: https://www.howtogeek.com/howto/16226/complete-guide-to-symbolic-links-symlinks-on-windows-or-linux/

---

I'm getting low on seti tasks (20 left) and will be done in about 24 hours. I finally got the rosetta stuff set up, and joined the orphans. I'm running my first task (9 hours??). no wingman? no match to test for accuracy? interesting...

I'm getting low on seti tasks (20 left) and will be done in about 24 hours. I finally got the rosetta stuff set up, and joined the orphans. I'm running my first task (9 hours??). no wingman? no match to test for accuracy? interesting...

Yep, same here. No wingman, no quorum, no trust to results :)

(I thinks those minor projects hide a few things that would surprise us, SETI veterans :D )

---

SETI apps news
We're not gonna fight them. We're gonna transcend them.

Thanks Raistmer and Wezh for joining.

Only 1 and 1/2 days and we already have earned a 25% badge as a team.

Let's keep it going.

---

Dave

I'm getting low on seti tasks (20 left) and will be done in about 24 hours. I finally got the rosetta stuff set up, and joined the orphans. I'm running my first task (9 hours??). no wingman? no match to test for accuracy? interesting...

The default target CPU runtime is 8 hours. If after 12 hours (4 hr extension) it's still running a watchdog timer will kill it off.
And yes, no comparison of results to validate the result.

And it looks like they are using Credit New, given the variability of Credit granted. Not as bad as here at Seti (no Anonymous platform support there), but it does still vary significantly.
But we're used to that here anyway.

---

Grant
Darwin NT

Gecko and Cygnus X-1 have joined us.

Welcome

In less than 2 days, team has passed 100,000 credit and almost 10,000 RAC.

Thanks all and pass the word.

---

Dave

I'm getting low on seti tasks (20 left) and will be done in about 24 hours. I finally got the rosetta stuff set up, and joined the orphans. I'm running my first task (9 hours??). no wingman? no match to test for accuracy? interesting...

Don't freak out too fast if the memory per task is very high. It recently declined but it started out at upto 2GB per task on the 4.1x version.

I am now getting tasks in the 500MB to 750MB range which is a vast improvement on the "ram crammed and jammed" problem.

Tom M

---

A proud member of the OFA (Old Farts Association).

with that i can run 2 tasks together with 2GB of ram ^^

---

My 4 core with 4 gig of memory runs 4 at a time.
Occasionally, one of the tasks goes to "waiting for memory" but, recovers as other tasks demand less.

My 8 core with 16 gig of memory, runs 8 all the time.

YMMV

---

Dave

Help with rosetta please.
It loads a ton of tasks with a short deadline that stops seti tasks from running. even without seti, it gives me more tasks than I can finish by the deadline.

It's wrong to think of it as "the project sends me...".

Projects never send you work unsolicited. You request the work, and the project responds - usually much better than SETI ever did. If you ask for a lot of work - you get a lot of work. Always start small (reduce your cache setting), and build up gradually as you get used to how each project works.

Having said that, there's often a problem when joining a new project for the first time: project administrators are not good at filling in the right numbers to set the initial runtime estimates correctly. The first runtime figures you see for a new project are likely to be wrong, and BOINC doesn't help by being slow to adjust them.

And finally, BOINC tries to even out the amount of time you spend on each project, in accordance with the various 'Resource Share' figures you assign to each project. If you've been running SETI before, and you add a new project - that new project will be out of balance with your existing projects. BOINC tends to overwork the new project, in the hope it will catch up. Broadly speaking, I'd simply suggest letting BOINC do it's thing: once you reach a state of equilibrium, it actually works quite well.

I had the same issue when I started.

Because of the project sharing nature of Boinc, Rosetta is in the hole to start with and takes over.

I suspended the Rosetta tasks and allowed the Seti to complete.

After that I allowed Rosetta to run and it straightened itself out quickly.

I ran with 10 days additional work on Seti.

I reduced it to 1 day with Rosetta

---

Dave

I've found a 1 day cache with 0.02 days extra has been working well for me.
More than one project, use a small cache. More than a couple of projects, use an even smaller cache.

---

Grant
Darwin NT

All well and good if you have had a small cache always. But if your last cache setting update on any project was for greater than a fraction of a day, the very first connect/join of the project nets you way too much work since cache settings are global and propagated across all projects for whichever was the last cache setting to be updated.

That is what got me in trouble on my first connection to Rosetta. It picked up my 1 day cache setting from Einstein since it was the last edited. You inherit that setting on the very first connect and download of the project files BEFORE you can do anything about it. Rosetta sent me 260 tasks on the first connect. Never could have finished all of them before the 5 day deadline.

The joys of BOINC. Ha ha.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

All well and good if you have had a small cache always. But if your last cache setting update on any project was for greater than a fraction of a day, the very first connect/join of the project nets you way too much work since cache settings are global and propagated across all projects for whichever was the last cache setting to be updated.

Yeah, i remember that joy when running Seti Beta. Smallest possible cache there to get the work back ASAP, but a large cache here to get through outages.
Took a bit of fiddling to get a big cache here, set NNT, then set a small cache for Beta. So i'd do Beta most of the time, then Main during the weekly outage, then back to Beta. Fiddling the cache settings each time.

---

Grant
Darwin NT

thanks for the setting suggestions. It helped. I thought setting resource share would do more to throttle and prioritize than it did. Now I'm playing with the right values. Thanks.

edit: my backoff for communicating with seti went from 10:06 mins to 30:18 mins