On Wed, Oct 05, 2011 at 09:06:40AM -0700, James Perkins wrote:
> On 10/05/2011 08:22 AM, Andrew Greenberg wrote:
> >Hi everyone,
> >
> >We got the new PCI-104 Atom-based flight computer stack up and running
> >tonight! See:
> >
> >http://psas.pdx.edu/avionics/av3_atom_fc_stack/
> >Although a few warranties were voided, everything went very well and
> >we're ready to start building Linux images and integrating the stack
> >into the existing avionics system.
> Congratulations to the team!
> This seems to validate my long-held opinion... PSAS has had another
> long journey through PowerPC land to eventually end up with another
> X86 architecture. For unique custom systems (given a plethora of
> alternatives) hardware cost, weight or performance are rarely gating
> factors to success. The ability to have engineers quickly dispense
> with platform and toolchain issues and get into application
> development quickly usually wins the deal. Intel is wise to this
> model and it maintains their relevance with all these embedded
> commercial and industrial computers, and the programming systems for
> them. Although it may not be fastest, smallest, least power hungry,
> cleverest, or best, Intel keeps being relevant to a lot of players.

I'll be interested in seeing how the atom box checks out.  I have heard
talk of some latency deltas between core2's and atoms. But have zero
data other than conversations that atoms have more latency jitter in
real time applications than core2 systems.

I do know there is more value in "tuning" code for max performance on
atom as compared to core2 as well.


psas-avionics mailing list

Reply via email to