Hi!

First of all, thanks for supporting Raspberry Pi. I love running asteroids@home on my Pis!

I do have some suggestion though. The deadlines are too tight for the Pis. My Pis are running asteroids@home 24/7 at full CPU load (at 700 MHz, not over clocked) and they can barely meet the deadlines of the tasks. It takes about 2 weeks for a Raspberry Pi to complete a task. It even seems that the recently downloaded tasks won't finish in time. If anything else runs on the Pis, there's no chance to complete the tasks. Pis need either smaller tasks or longer deadlines.

Any chance to make use of the Mali400 GPU in the Pis?

Thank you!

---

http://iqjar.com

The deadline will be adjusted.

Much appreciated! Thank you!

I know that you guys are really busy and I respect that.
Still, when you can spare a few minutes, could you please adjust the deadlines for the Raspberry Pi tasks? One of my Pis has just completed a task 30 minutes before the deadline, after nearly 10 days of computing... scary. I would hate to see 10 days of RPI work go to waste just because the Pi is a few minutes late after the deadline.

If the deadlines cannot be adjusted separately just for RPI tasks, but globally, then it's probably not a good idea to increase them by much (it would increase overall task validation time). A 25% increase would be enough from my point of view.

Thank you!

One of my Pis has just completed a task 30 minutes before the deadline, after nearly 10 days of computing... scary. I would hate to see 10 days of RPI work go to waste just because the Pi is a few minutes late after the deadline.

Are you aware that Pi's can be overclocked a liitle bit? This could give you the necessary headroom.

Yes, I know :) But I don't want to overclock them. Those Pis have important missions in my house. They command automated systems that control the heating and the lighting, so I would not risk destabilizing them.

Hi, I have a small rack of Raspberry Pi's dedicated to crunching (Asteroids and WUProp@Home).

Is there any chance to expand the software to enable the GPU to be used as well? I know there is a lot more complexity involved. Will the GPU give a good speed-up on crunching and make it worth it?

The current app is rather dated and does not utilise the neon extension that the ARM v7 and later have. In theory all it would take to use them would be a recompile with an up to date gcc with the appropriate options. Unfortunately Kyong is rather busy just keeping the project running.

As for using the Videocore IV that would require coding changes and there is no guarantee that it would be faster. You also have to remember there is only one Videocore but there are 4 CPU cores so it would probably be more beneficial to optimise the CPU part.

---

BOINC blog