The Scottish Boinc Team

Project details can be found here http://burp.renderfarming.net/intro.php

From AMD User NeoGen:

"Yea, BURP can take up the whole amount of RAM or just a few hundred megs... it depends on the job that its processing and the complexity of the specific frame that your computer is rendering. I guess BURP decides 12Gb is a safe value for all its rendering projects, instead of having to guess and tweak the amount with every single project that comes in.

I haven't run Lattice much but I'd like to see GARLI taking that much RAM, I may just fire up the project to take a look at that as my main desktop has 32Gb of RAM."

You need at least 12 GB in the system you run BURP with, 16 GB to be more sure and 32 GB to be absolutely sure...

I'm going to give it a go at some point although it sounds like it might be one wu at a time!

How much GB per core ? I have a few 32 GB systems but they are 8 core xeons or 4 core i7's ( HT turned off ) never tried running this project before.

One should think that 32 GB over 4 cores should be more than enough, that's 8 GB per thread.
You might even run more than one at a time.

Cheers Dirk, might try it on one of the servers if I get the RAID rebuild by the end of tonight......

And surprise! BURP has an Nvidia app. Just started on a 780. It's still using 280MB RAM. Will report back when it's done.

I've got one running on a tiny GT610

Here's some avg run times for different GPUs. Each WU pays 100 points.

GTX 770_____20 minutes
GTX 780_____22 minutes
GTX 960_____26 minutes
GTX 660Ti___24 minutes
GTX 750Ti___27 minutes
GTX 750Ti___29 minutes
GTX 460_____35 minutes

GPU utilization stays 90% or higher so I'm not sure it's worth trying run more than one at a time.

I'll stick some 660ti's on it and help you along as well. BURP languishing in 89th place anyway so a good boost will do no harm

Be aware (or beware) that some CPU WUs can take a long time. I read one post from a cruncher where it took 80 hours running on 15 cores. Not sure if it uses all available cores by default or how it's controlled. Maybe with an app_config.xml similar to yafu, which looks like this...
<app_config>
<project_max_concurrent>3</project_max_concurrent>
<app>
<name>yafu</name>
<avg_ncpus>2</avg_ncpus>
</app>
<app_version>
<app_name>yafu</app_name>
<plan_class>mt</plan_class>
<avg_ncpus>2</avg_ncpus>
<cmdline>--nthreads 2</cmdline>
</app_version>
</app_config>

Haven't got any cpu units, but on Linux the times are quite similar for the nVidia app.