[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Freescale iMX53 Quick Start board



On Fri, Aug 05, 2011 at 04:07:06PM +0000, Phil Endecott wrote:
>Hi Steve,
>
>Steve McIntyre <steve.mcintyre <at> linaro.org> writes:
>> Hi folks,
>> 
>> We've got a number of discount vouchers for these
>
>Too late for me...
>
>> We're about to purchase a number
>> of these for buildds for the Debian armhf port
>
>Hmm.  Are you sure it's the best choice?  It has a single Cortex A8 (like 
>the old Beagle Board).  The Panda Board has a dual Coretex A9.  I've not
>benchmarked them, but I would expect the Panda to be a lot faster for 
>CPU-bound tasks.  The iMX53 has the advantage of SATA if you expect to be 
>I/O limited, but if I were building a "farm" I think I'd seriously consider 
>NFS and negligible local storage.

We were pondering the panda, yes. We've got a table of the possible
build machine hardware options at

  http://wiki.debian.org/ArmHardFloatBuildd

However, there is a fair amount of anecdotal evidence for problems on
the panda under heavy load at the moment. Jon Masters told me last
week that the Fedora folks are having major trouble with instability,
and others have similar tales. Also: panda availability is currently
approximately zero I'm told. The iMX53 is the next best option for us,
I feel. Especially with the discount vouchers we have, it's looking
very cheap.

>My choice of the iMX53 was based mainly on its video features (it has 
>analogue VGA out, which the Panda doesn't) and more than a bit of 
>prejudice because I seemed to be the interested only person at CELF last 
>year who didn't get a free Panda.

Aww... :-/

>Anyway, if you do go with the iMX53, do keep us informed.  I'm currently 
>working on mechanical/electrical things e.g. a box and power supply.  Note 
>that the board seems to have very little capacitance on its 5V power input, 
>so it can "brown out" if you connect something (via USB or otherwise) that 
>has non-trivial inrush current.  There is also the fact that it powers-on 
>"off" until you press a button, which I imagine is probably not the behaviour 
>that you'll want.

Yup, I'm aware of at least some of those. Thanks for the info, and
we'll keep folks informed as to how things are going.

-- 
Steve McIntyre                                steve.mcintyre@linaro.org
<http://www.linaro.org/> Linaro.org | Open source software for ARM SoCs


Reply to: