But still, it is much better to know what the problems are going to be in
advance. If it is still worth getting done, at least the warnings about
the problems will have already been given. Even with the most powerful
cell phone, there are major problems. The worst of which is the ~1% duty
cycle required to get decent battery life.
jm7
Martin
<[email protected]
o.uk> To
Sent by: [email protected]
<boinc_dev-bounce cc
[email protected]
u> Subject
Re: [boinc_dev] BOINC for Mobile
Phones - please test on your Java
11/02/2009 06:53 phone
AM
Nick Simpson wrote:
> I just have to throw my two cents into this:
>
> The guy is willing to put time into it to make the platform available..
why
> all the negativity? If its available, you'd be surprised what someone
might
> come up with. To say that nothing exists right now that would use it is
not
> a reason to not develop it. John mentioned how CUDA wasn't very good in
the
> first generation and likened phones today to that. It takes that first
> generation to get it out of theory and into practice so that a
> person/community can see what works and what doesn't. Until it is tried,
it
> is all hypothetical.
I agree. Some of the "peer review" here does get to be a little
"resistant" to changes/developments sometimes.
My view is that Boinc is an awful lot more than just high power
super-computing number crunching.
Also, there's an awful lot of very capable mobile phones that are
quickly becoming ever more capable... That's a huge and often idle
resource that can be tapped.
As always in computing, it is important to appropriately match together
the application and the hardware. Perhaps that is something that Boinc
can be developed further to support better, so that Boinc tasks are
preferentially offered to the best suited hardware for the task.
Regards,
Martin
--
--------------------
Martin Lomas
m_boincdev ml1 co uk.ddSPAM.dd
--------------------
_______________________________________________
boinc_dev mailing list
[email protected]
http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev
To unsubscribe, visit the above URL and
(near bottom of page) enter your email address.
_______________________________________________
boinc_dev mailing list
[email protected]
http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev
To unsubscribe, visit the above URL and
(near bottom of page) enter your email address.
