> AVR is a nice architecture, and I used it for years, but the problem
> is that you have to deal with Atmel which doesn't have a quality
> control program like others do. Also, Atmel almost broke my last
> company when they denied that they had a problem with their Mega32
parts.

You ain't kidding. I had to kill a two month project (1/6 of my
yearly income) a few years ago because their 'mature' jtag programmer
on their 'mature' mega128 was zorking some internal debug registers
and making the parts dead after 2-5 programming cycles. After sending
parts to Norway to have them disected, someone finally admitted there
was a 'problem' on thier end. Project was long dead by then. Course
I still keep using them since I know the hardware and have a bunch of
drivers...fast becoming an ARM lover, though. Went through the hell
of .ld and startup.S mismatch issues for awhile, now full speed ahead.

One thought on the pic's: I've worked at alot of 5-man companies, all
owned by one guy who's been doing it since the 60's or 70's - they
pretty much all use pics (or 68 hc08/11/12) for the same reason I keep
using avr's - who wants to learn something new when something old is
resonably priced and works fine (of course, the pic32 bears no
resemblance to a pic11 in either architecture or toolset, so I'm not
sure about the validity of the argument). I'd never touch them
without a completely gpl gcc package myself, though. I actually did
dig though avr-libc once to fix a bug. That wasn't pleasant, but I
was running again in days instead of weeks or months. Brought up my
first *nix system in the process - way easier than installing cygwin,
imho.

Steve

An Engineer's Guide to the LPC2100 Series

> The "PH" book describes the MIPS I architecture. Not quite the same as that
> employed in PIC32, but the same heritage. It may be that there are
> thousands of graduate developers in the world who have the skills needed to
> understand the way that the PIC32 operates ...?
>
> I suspect that this is not the case with ARM (I may be wrong)

A pity that book does not cover the ARM architecture as well!

We use ARM chips (or rather, boards) in our courses, because it is one
of the cleanest 'modern' architectures to learn (comparable in spirit to
golden oldies as 68k and PDP-11) and the chips are good enough and
relatively easily available. And the chips are also used with C and C++.

(We also use PIC (14-bit core) assembler to give the students a taste of
how bad an often-used architecture can be. Note that I don't say PICs
are a bad choice, just a bad CPU architecture.)

--

Wouter van Ooijen

-- -------
Van Ooijen Technische Informatica: www.voti.nl
consultancy, development, PICmicro products
docent Hogeschool van Utrecht: www.voti.nl/hvu

...
> not a fair comparison, IMHO, a better one would be a generic CM-3 vs.
> M4K.
> I would suggest reading this WP: http://www.bdti.com/articles/m4k.pdf

I agree. But, note that the article has a number of mistakes regarding
Cortex-M3. That is, things like not realizing it has CLZ (count
leading zeroes), not knowing it has saturation support (SSAT and USAT
instructions), not realizing it has optimized branches, ignoring the
value of optimized back-to-back loads/stores, and not realizing it has
early out HW divide (divide ends often in 3 or 4 cycles, with a worse
case of 13). Further, forgetting that quoted speed of the core is
meaningless as the only measure has to be speed when connected to a
real system (ie. eFlash and real SRAM and peripherals). More
importantly, area and power are far more affected by eFlash and SRAM
(and peripherals) than the core. A core running 100MHz needs memory
running at that speed, and that memory will be larger and burn a lot
more power. If you want embedded Flash (which you do in an MCU), the
situation gets far worse. So, any comparison has to also consider what
speed is actually system sustainable (as Microchip has learned) and at
what cost.
That said, the bigger problem is that the article was centered on
someone building their own MIPS 4K core and not PIC32. Microchip,
through PIC32, picked many of the weaker options from the perspective
of the article (which is focused on signal processing).
But, I agree that one should be comparing PIC32 against Cortex-M3,
since that is what Microchip is trying to defend against.

--- In l..., "roboticsbcn" wrote:
>
> Hello,
>
> Someone can make a comparison between the architectures of the ARM7 and
> PIC32?
>
> Thanks!
>

Hi,
not a fair comparison, IMHO, a better one would be a generic CM-3 vs.
M4K.
I would suggest reading this WP: http://www.bdti.com/articles/m4k.pdf

Roger

At 11:39 PM 5/10/2008 +0800, Xiaofan Chen wrote:
>Many of the firmware developer or hardware developer do not come
>from the computer engineering or computer science department.
>So many of them have not attended computer architecture course.

I would venture the largest group are probably from EE with large numbers
from the sciences as well.

Robert

" 'Freedom' has no meaning of itself. There are always restrictions,
be they legal, genetic, or physical. If you don't believe me, try to chew
a radio signal. " -- Kelvin Throop, III
http://www.aeolusdevelopment.com/

At 07:22 AM 5/10/2008 -0400, J.C. Wren wrote:
>I've worked at one company where the purchasing department made the
>decision...

I don't suppose they then took responsibility for any problems that caused?

Robert

Another sign of the end of civilization, our technical magazines are
getting chatty
From an EETimes product descriptions 2006/08/09
".... systems that can sample gobs of inputs simultaneously"
Now just what is the technical definition for gobs again?
http://www.aeolusdevelopment.com/

Richard,
I do not know a whole lot about the history of Microchip
marketing - once I heard Microchip described as a company whose success
resulted from significant marketing rather than superior engineering.
I'm not really sure what that means, in concrete terms.

I would love to hear your perspective and history on this - if
only so as to understand how Microchip's direction fits into the larger
context that they are now playing it (i.e. ARMs, etc.)

*Brian

-----Original Message-----
From: l... [mailto:l...] On Behalf
Of FreeRTOS.org Info
Sent: Saturday, May 10, 2008 6:27 AM
To: l...
Subject: RE: [lpc2000] ARM7 vs PIC32

> When I see statements like

You have to look at the history of the Microchip marketing to know why
that
was said. It has nothing to do with confidence.

> that I wonder how many customers
> really know so little about their
> requirements,

LOL - you should read some of the emails I get.

Regards,
Richard.

--- In l..., "Paul Curtis" wrote:
>
> Bob,
>
> > > Someone can make a comparison between the architectures of the ARM7
> > and
> > > PIC32?
> > >
> > > Thanks!
> > >
> > Hi,
> >
> > lets not forget one critical value when comparing two CPUs, how many
> > clocks per instruction are needed in the average of a large program to
> > complete an instruction.
> >
> > afaik, this value is 1.9 for an ARM7 but I could not find the number
> > for MIPS, it is either called something else or not easy to find.
> >
> > ARM core has the reputation to be particular low in power consumption,
> > a reputation that the MIPS core does not share. However, I have seen
> > the AVR32 running a benchmark versus an ARM7 and the AVR easily beat
> > the ARM7, once again because it needed a lot less cycles to do the
> > same job.
> >
> > I do like the ARM devices, not primarily for the great core but for
> > the availability of so many derivatives from all the vendors at very
> > good prices.
>
> None of that is a comparison of architecture. That's comparing
> implementation. As for AVR32 vs ARM7, um, like looking at an ARM in the
> mirror. It's hardly an earth-shattering architecture, unlike the ARM
> architecture when it came out. Again, comparing some AVR32 to some
ARM7 is
> comparing implementation, not architecture. Sure, the architecture
might
> help things along a bit, but implementation is also key.
>
> --
> Paul Curtis, Rowley Associates Ltd http://www.rowley.co.uk
> CrossWorks for ARM, MSP430, AVR, MAXQ, and now Cortex-M3 processors
>
AVR is a nice architecture, and I used it for years, but the problem
is that you have to deal with Atmel which doesn't have a quality
control program like others do. Also, Atmel almost broke my last
company when they denied that they had a problem with their Mega32 parts.

Support is an issue for me. NXP and Philips have been good to me so far.

boB

On Sat, May 10, 2008 at 11:36 PM, Pont, Michael J.
wrote:
> One observation.
>
> Many students around the world learn about computer architecture via the
> book:
>
> Patterson, D.A. and Hennessy, J.L. (2004) "Computer Organization and Design:
> The Hardware / Software Interface", (3rd Edition). Elsevier /
> Morgan-Kaufmann. ISBN: 1-55860-604-1.
>
> The "PH" book describes the MIPS I architecture. Not quite the same as that
> employed in PIC32, but the same heritage. It may be that there are
> thousands of graduate developers in the world who have the skills needed to
> understand the way that the PIC32 operates ...?
>
> I suspect that this is not the case with ARM (I may be wrong)
>

Many of the firmware developer or hardware developer do not come
from the computer engineering or computer science department.
So many of them have not attended computer architecture course.

So ARM or MIPS are in general equal here.

Xiaofan

> Someone can make a comparison
> between the architectures of the ARM7 and
> PIC32?

One observation.

Many students around the world learn about computer architecture via the
book:

Patterson, D.A. and Hennessy, J.L. (2004) "Computer Organization and Design:
The Hardware / Software Interface", (3rd Edition). Elsevier /
Morgan-Kaufmann. ISBN: 1-55860-604-1.

The "PH" book describes the MIPS I architecture. Not quite the same as that
employed in PIC32, but the same heritage. It may be that there are
thousands of graduate developers in the world who have the skills needed to
understand the way that the PIC32 operates ...?

I suspect that this is not the case with ARM (I may be wrong)

Michael.