vinas1,

The basic idea of all (or most at least) distributed computing is that it works during the spare cycles. One might even argue that "investing in 50+ cores" exclusively for DC is even cheating because you create on purpose a pool of spare cycles instead of just recycling those cycles that are idle ayway.

Then again, if and when I decide to buy a new box for private use, I might also check the GPU for compatability. And maybe spend a few € extra...



Not really jealous of the GPU crowd concerning crunching. Only unwilling at this time to spend money solely for WCG credits.

I too have a BUNCH of cpu cores and gpu's too...and beg to differ with your characterization about 'spare cycles'. I have only bought new cpu's and gpu's as a result of my working on friends pc's which has created a small amount of disposable income. But MOST of my pc's come from those friends pc's that I have worked on upgrading them to newer versions of hardware, keeping their older ones as my 'spare' ones. YES I have bought SOME new parts, but after spending only $200US on parts and using those friends pc's 'spare' parts I can make an 8 core AMD pc with two AMD 5770 gpu's capable of returning 2 gpu units in under 5 minutes while at the same time running 8 cpu units! Now to be fair I build my own pc's so the costs are MUCH lower than if I had someone else do it. But I DO use 'spare' parts, I have DRAWERS full, and have 8 pc's sitting on the floor that COULD be running by this afternoon IF I had the electricity available here at my home. I currently have 13 pc's up and running and have to be careful as the circuit breaker will pop if too many pc's are plugged into the same circuit! ALSO outlet plates getting HOT to the touch is NOT a good thing either!! I DO agree that BUYING pc's is NOT what was originally how Boinc was conceived but 'spare' is in the mind of the beholder, so my 'spares' could be your primary pc's.

I have a 3 core that has been running at 90% for the last 3 years and never needed to worry about impacting my daily work. My new Quad, also running at 90%, comes to a violent stall with the first GPU WU.

The first rhetorical question is: Why run a CPU job with zero user impact when you can run GPU jobs that brings a Quad to the ground?

The second rhetorical question is: Why would you want to ignore the available spare CPU cycles on a computer with no SUPPORTED GPU?

As It may be faster, and will be here soon anyways, lets all shut down and wait for ... (insert next best thing here).

Better yet why not take what we have and use it to the fullest and get it done before it arrives. Just becuase a GPU does xxxx wu a day and a cpu does only yy that is still xxxx+yy done (aka closer to the goal and that many more then GPU alone can do) Even a 1% increase in production is good

Considering how I joined a team of over 768 active members, and my single HD 7950 produces 60% of all the results and PPD from that alone, tells me that there is no reason to carry CPU apps for this project any more. It would be better for people to use their CPUs on other projects that do not have GPU acceleration. One person using a single GPU is worth many, many, many countless high end processors. In order to get equivalent results with processors alone, I would have to have upgraded my 24+ cores of highly overclocked processor cores to around 240+ cores worth of machines in order to equal the single graphics card I use today.