Unless you leave your phone plugged in all the time, I think you would be plagued with a dead battery in quite short order. Also with the processor in your phone running full or even half bore, I think you might experience heat problems with the phone. I do not think phones are designed to have the processors running all the time. Please correct me if I am way off base.

Cheers

Sgt.Joe, you are absolutely correct. Any phone powerful enough to be useful (and even the most powerful are very, very slow compared to desktop computers) would pretty much melt if used at full power for extended periods (even just a few minutes).

They wouldn't literally melt - phones have thermal cutoffs and I imagine the latest can underclock and throttle as required.

Nick-MMX, how long does your benchmark run for, and what results does it report? Can you give us some examples?

How have you validated your results? An aggressive optimiser can cause benchmarks to produce completely inaccurate results.

I had hoped you were using the standard dhrystone and whetstone benchmarks, so we could at least pretend to have comparable values. I am unsure why you think your "integer" benchmark is representative of integer performance. What operations do they break down into?

http://uploadfile.org/download.php?id=Q36WaoPxjATj6IYqPxem
Specs: 64MB Internal SDRAM, 225MHz Processor
My phone got 178355 FLOPS

Didactylos,
you can be right if we assume that the devices will crunch all time in the same way that desktops or laptops, but another part of this history is that the idea behind EDGES evolved.

It will work with small WU's, breaking much more the work units in smaller parts than we know today, and will be designed to use less power of computation per device than normal computers but in more time and under other kind of algorithms than any computer used right now in any grid computing project... it is a breakthrough initiative !

Just think that different processors can be ported with Java based BOINC or other specific EDGES plataform, using many rules from BOINC but basically will not have the same kind of approach.

Many cell phones (specially smart phones) today have more processing capabilitily than some computer processors in 2003 when BOINC started. Nick-MMX just put some evaluations posted to probe it...

In a real cloud computing network world, a lot of virtualization capabilities could be used to produce a lot of work from many devices (a little bit from each one than just one computer as we know today, and many of then could work if we sum, as one in virtualization mode using grid computing techniques and cloud computing capabilities and features. For example, just easy parts of computation could be processed by cell phones, and hard parts of computation will be sent to other more quickly and heavy power computer devices like desktops or supercomputers...).

So, in this way, we will not need the same processing capability that we found in desktops or laptops to process the same quantity of Flops if we think in performing grid under cell phones, but we will need a lot of more small devices (like cell or smart phones provide us today) to process the same quantity of Flops, and for sure we will need special grid software with a kind of virtualization and paralelism running over a less weigth platform to reach similar or superior results, and I believe that it will be the main barrier once there are a lot of devices and processors to be ported over this platform... an incredible challenge today !

It is another new approach than any grid project or WCG or any other BOINC based project use until today. And it will be a incredible leapfrog in grid technologies if became feasible.