I have 3 projects set up in my BOINC: SETI@Home, World Community Grid, and Rosetta@Home. I wanted it to run mostly SETI@Home tasks though, but for some reason, the vast majority of SETI@Home tasks I'm getting are for my GPU. I'm getting very few CPU-bound SETI@Home tasks. Thus, most of the tasks my BOINC has been running are for my other projects.

I've even tried setting the resource share of SETI@Home way higher than the others. Currently I have the SETI@Home resource share set to 500 and the other two projects set at 75. Despite that, I seem to be getting mostly GPU tasks for SETI@Home.

Currently, I have 101 SETI@Home tasks in my queue, and 100 of them are for the GPU (only one is for the CPU). It seems a little imbalanced. I only have one GPU, so each of those tasks will run sequentially, whereas my CPU has 8 cores with hyperthreading (16 threads total).. Though, I realize a GPU has many more compute cores to work with.. My GPU is currently an Nvidia GeForce GTX 1070 TI 8GB (Asus ROG Strix GeForce GTX 1070 TI).

Is it just that there isn't enough SETI@Home work right now, or is there anything I can do to at least balance it out more?

---

Is it just that there isn't enough SETI@Home work right now, or is there anything I can do to at least balance it out more?

There is plenty of Seti work, it's just that your resource share is what is limiting the amount of Seti CPU work you will get. Unfortunately some projects (apparently WCG is one of those) that isn't so good at honouring share settings.
And the way the work share being done is actually determined it can take a very long time for things to balance out. And if you try to micro manage things, that can often make things very much worse.

Since Seti is your main project I'd suggest setting Seti as 1000, and the other projects as 1. See how things go over a week or so (it will take a while for your present work to be processed).
It is also worth running with a small cache when running more than one project- it allows the work share to balance out much, much, much sooner than if running with a large cache.
In your computing preferences, in the Other section, i'd suggest
Store at least 0.5 days of work (or even less)
Store up to an additional 0.01 days of work
for your cache settings.

---

Grant
Darwin NT

It is also worth running with a small cache when running more than one project- it allows the work share to balance out much, much, much sooner than if running with a large cache.

+1

---

~ Life's short; make fun of it.
User ID: 39750; SETI@home member since: 5 May 1999; Classic workunits: 5,851; Classic CPU time: 108,169 hours

Main issue is the other projects don't have any GPU tasks. So if the System requested CPU just from Seti, it would end up doing much more work for this project. So you end up with the CPU working on the other projects, and GPU for Seti to get the balance right.

You can change the work share , but then you end up using all the GPU time, AND most of the CPU as well. Good if you want to do more for Seti, but the other projects will suffer.

I have set World Community Grid and Rosetta to not get more tasks for now, and now I'm getting a lot more SETI tasks.

The GPU tasks seem to take a lot less time than the CPU tasks, so I wonder if the balance was good already..

---

I have set World Community Grid and Rosetta to not get more tasks for now, and now I'm getting a lot more SETI tasks.

Which will result in getting less Seti tasks when you enable the other projects again in order to balance out the work share according to your settings.

The GPU tasks seem to take a lot less time than the CPU tasks, so I wonder if the balance was good already..

WUs that can take 50min on the CPU can be done (with the right hardware and application) in around 30 seconds on a GPU.
For most of us it's more like 15 down to 4min depending on the GPU being used.

---

Grant
Darwin NT

Since I joined Science United on a Windows 8.1 PC i am getting Milkyway@home tasks, Aateroids@home tasks and Cosmology@home tasks. Milkyway GPU tasks are a lot faster than CPU tasks, Asteroids GPU tasks are much slower on the same GTX 1050 board.Cosmology tasks can use a Virtual Machine via VirtualBox but are very slow.
Tullio

---