StrongARM SBC performance of Squeak?

Ned Konz ned at bike-nomad.com
Wed Jul 18 19:47:21 UTC 2001


On Wednesday 18 July 2001 11:48 am, you wrote:
> Ned Konz <ned at bike-nomad.com> is widely believed to have written:
> > We're looking at using Squeak on a small, low-power single-board
> > computer, perhaps a StrongARM machine.
> >
> > We will probably be running Linux, and using VNC for the display.
>
> This will be pretty similar to both my Acorn amchine and the NetWinder
> (RIP) except for the use of VNC. At 200MHz you can expect reasonable
> general performance (10-12million bc/sec, for what that's worth) but
> abysmal floating point performance due to lack of hardware. 

Hopefully, Squeak doesn't require too much float in general. But didn't 
someone say that it was doing float math in Morphic?

> There is
> also a problem right now with gcc being totally f*cked in the
> optimisation department, roughly halving the performance on most linux
> ARM machines.

Is this being fixed?

> With VNC for displaying, you will obviously be hugely affected by the
> details of the implemetation, but on my Acorn the Morphic performance is
> a bit slow to say the least. It's not at all bad on the NetWinder when
> you have one with the older gcc/kernel that doesn't have a terminal
> float/optimisation bug.

Is the older gcc and kernel available still? (though I may need USB, which 
will probably argue for a 2.4.x kernel).

> Given suitable encouragement (ie money) I hope to get to do an ARM
> jitter someday soon, which might make the general performance much
> better, maybe even boost the Morphic performance.

Great! I'll look forward to it!

-- 
Ned Konz
currently: Stanwood, WA
email:     ned at bike-nomad.com
homepage:  http://bike-nomad.com




More information about the Squeak-dev mailing list