I have a minor mystery on my hands. One of my computers is running S@H, E@H, and Milkyway@H. For some reason I am only getting CPU WUs for E@H, the others are only getting GPU WUs. Now, I did update my storage preferences recently (all store min 0.5 days, and no more than 5 additional days, and 1.0 GB of total HD storage). I don't see how that could be causing the issue. My only guess is that E@H has WUs that have a much greater priority. Any help?

---

Seti@home classic: 1,456 results, 1.613 years CPU time

Einstein has MUCH shorter deadlines than Seti. So it will always get priority over Seti. It will balance debt out eventually so you will have to have patience. Or greatly reduce your project percentage for Einstein.

---

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

A proud member of the OFA (Old Farts Association)

I knew of the deadline differences in the project, I just find it odd that I've been running the same setup for a long time, and the only thing I changed was WUs stored and HD usage. Perhaps it is what it is for now. At least the computer is working to its potential, albeit in portions I was not expecting.

---

Seti@home classic: 1,456 results, 1.613 years CPU time

Also everyone always misinterprets the "Store up to an additional" days of work setting. That sets the length of time between project network connections, not any amount of work assigned your cache size. The only think limiting your cache size is the primary "Store at least" days of work, the server limit of 100 cpu tasks per cpu and the 100 tasks per gpu and the global limit of hard drive size allotted to BOINC.

I found I could limit the amount of work that Einstein sent me by reducing the "Use no more than" hard drive allotment since Einstein always sends more work than can be finished by deadline. But the side effect of doing that was it limited the amount of work that Seti would send on requests after the Tuesday outages when trying to refill my cache. Once I increased the hard drive limit to 10GB, I got the full allotment of the RTS buffer at each Seti server connection. But the side effect of that was that Einstein would send gigabytes of work and would totally commandeer the hosts to exclusive Einstein use trying to fulfill the short deadlines of Einstein work. Since Seti is my primary project that was unacceptable. So now I just get a bit of Einstein work and then set NNT until the Einstein work is finished or getting low and then I refill again with another batch of 100 tasks or so and set NNT again.

If you only run single projects on each host, you can use venues assigned to each host and the project use percentages can be better managed. But if you run multiple projects per host, the global hard drive size bites you in the ass every time.

---

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

A proud member of the OFA (Old Farts Association)

First, as Keith has said you are not using the cache settings correctly. The first figure "store at least 0.5 days" means you will only ever collect 0.5 days of work, the second "store additional 5 days" means, as Keith has said your computers will only contact the servers every few days, or when they run out of work. A better set of settings given your mix of projects would be - "store at least 1 day" and "store an additional 0.01 days". This will ensure you keep contacting the server and longer deadline projects will get a shout at getting work, rather than the very short deadline projects dominating.
I think your 1Gb of disk space is a bit on the "short" side as some projects do send very large tasks and require a lot of working space on disk, so I wouldn't go bellow 5Gb.

---

Bob Smith
Member of Seti PIPPS (Pluto is a Planet Protest Society)
Somewhere in the (un)known Universe?

First, as Keith has said you are not using the cache settings correctly. The first figure "store at least 0.5 days" means you will only ever collect 0.5 days of work, the second "store additional 5 days" means, as Keith has said your computers will only contact the servers every few days, or when they run out of work. A better set of settings given your mix of projects would be - "store at least 1 day" and "store an additional 0.01 days". This will ensure you keep contacting the server and longer deadline projects will get a shout at getting work, rather than the very short deadline projects dominating.
I think your 1Gb of disk space is a bit on the "short" side as some projects do send very large tasks and require a lot of working space on disk, so I wouldn't go bellow 5Gb.

Funny enough, I increased that from 500 MB, so that is more! I think the 1 GB is fine for now, this computer has other uses than a cruncher and I don't want to store too much on it.

I believe changing the "store up to an additional" variable back down worked, I am getting more non-GPU WU. Thank you, Keith & Rob!

I did try looking for a similar discussion prior to starting this one and I didn't find one, but after a second look I found another related one here: https://setiathome.berkeley.edu/forum_thread.php?id=83515 I apologize for you having to repeat yourselves.

---

Seti@home classic: 1,456 results, 1.613 years CPU time