Moving from 8051 to AVR

That's too bad. One always likes to think life could be better.

--
Grant Edwards                   grante             Yow!  .. I want to perform
                                  at               cranial activities with
                               visi.com            Tuesday Weld!!

Not true.

As usual, you forgot the caveat: "as long as you are a current customer, paid and up-to-date on the latest version, AND you have personal friend(s) at the compiler company".

_and_ they themselves still have it _and_ the company still exists to talk with you.

Jon

Yea, that was a funny one.

So, what you end up with is the product of an infinite series of probabilities less than 1.0. In the limit, the probability is 0. The only question is how fast it approaches 0.

--
Grant Edwards                   grante             Yow!  I will SHAVE and
                                  at               buy JELL-O and bring my
                               visi.com            MARRIAGE MANUAL!!

Impossible? What is the problem? You can just archive and copy the binaries. I do this all the time. E.g for me each version of the tools lives in its own directory.

/opt/arm-elf-3.3.2 /opt/arm-elf-4.0 /opt/avr-elf-3.3.2 /opt/c167 /opt/gnuarm-3.4.3

You just set the path appropriately (according to which version you want to run), in the Makefile. Since this is version controlled along with the rest of the source code, the correct compiler gets run automatically.

Actually I have been thinking of checking in the binaries of the tools, too, along with my own source code. (Has anyone done this with svn? Is it too slow?)

You know, I realise I have no idea how I would achieve this with a similar suite of comercial compilers, with their dongles, node locked licenses, key disks etc.

I have *got* the old versions. Why would I not? It is *running* them that is the problem!

--

John Devereux

That's easy: Because of customer demand.

They saw the Philips LPC doing very well, and many customers said "If you do the same, we will buy it" - not "Hey, yes we'd LOVE to jump to a core with short life cycles and a history of supply hiccups"

After all, what Atmel build is not decided by Ulf :)

There are a lot of interesting C51 things comming out of asia..

Of course - remember Ulf has a certain ( somewhat rose-tinted ) view of AVR vs C51, and it's ben a while since he was at the code-face...

-jg

WOW, from a lecturer, insane.

I consider NOT knowing analog electronics a HUGE weakness.It limits you from half the circuitry in the world ;). Unfortunately I have that weakness and am trying to rectify it.

Do you mean: it's almost impossible to get a tarball of a specific version of gcc? or: it's almost impossible to get the installation package of a specific version of gcc?

The tarballs seem to be widely available: a mirror of ftp.gnu.org has versions back to 1.42.

Because gcc is Free software, if you have an installation package, taking a copy and keeping it in your configuration management system is perfectly legal.

And then there are things like snapshot.debian.net, allowing you to acquire any version of an installable package from Debian.

So I'm not sure which bit of this is impossible.

Assuming the vendor is still in business, amongst other problems.

Anyone know how I can get, for example, Acorn C v3?

cheers, Rich.

--
rich walker         |  Shadow Robot Company | rw@shadow.org.uk
technical director     251 Liverpool Road   |
need a Hand?           London  N1 1LX       | +UK 20 7700 2487
www.shadow.org.uk/products/newhand.shtml

Computer *science*, note: these days, these are guys who purchase computers, not build them :->

It means you can't *debug* the stuff when it isn't working because of some weird fault, because you have no model of how the data is actually getting sent...

Reminds me of the chap we had here once from a UK University, with a degree in Mechanical Engineering, who hadn't been into an actual workshop during his degree...

cheers, Rich/

--
rich walker         |  Shadow Robot Company | rw@shadow.org.uk
technical director     251 Liverpool Road   |
need a Hand?           London  N1 1LX       | +UK 20 7700 2487
www.shadow.org.uk/products/newhand.shtml

Not necessarily. That strongly depends on who he's talking to. If the above read "the lecture told the auditorium of computer science

101", the statement could be judged to be on the boundary between eccentric and mischievously misinforming. If he said the same thing to an individual student whose area of interest/talent he knows to be firmly rooted on the computer science end of the school, it's perfectly correct. Having that student add more than the vital basics of analog electronics (Kids: you don't want to touch _this_ end of a voltage convertor, nor _this_ end of a soldering iron) to his course selection would only distract him from more important matters, yet still fail to make him a usable analog electronic engineer.

Not all students are the same, so they don't all require the same advice.

--
Hans-Bernhard Broeker (broeker@physik.rwth-aachen.de)
Even if all the snow were burnt, ashes would remain.

Don't think so. Intel sued Atmel for patent infringment, and got their face slammed. Atmel got the 8051 as a compensation, but quickly decided something better was needed - AVR...

Yes Atmel has employed some Jim Granville clones. But I do believe in the principle that managers should check with me before making important decisions...

Nope, still program regularily for customers. Both AVR and ARM for embedded and Linux.

8051? Nah, Life's to short... If The Guantanamo prisoners were forced to program 8051's they would talk, soon enough ;-)

--
Best Regards,
Ulf Samuelsson
ulf@a-t-m-e-l.com
This message is intended to be my own personal view and it
may or may not be shared by my employer Atmel Nordic AB

I think you have to remember Ulf works for Atmel; presumably they want to push their own technology (AVR) rather than one they second source. Could well be a bad call when the ARM takes over the lions share of the embedded market.

Ian

Atmel was the first company to promote ARM as a standard Microcontroller. Where's the problem ;-)

--
Best Regards,
Ulf Samuelsson
ulf@a-t-m-e-l.com
This message is intended to be my own personal view and it
may or may not be shared by my employer Atmel Nordic AB

Er, surely that's obvious - two of your three microcontroller product lines will be obsolete.

Ian

Since this is topical, note these new release from Infineon :

XC886/888 info is here ( adds to their existing smaller XC866 ) This is a divide by 2 core ( divide by 2 allows any clock duty cycle )

12 MIPS

as you would expect from Infineon, solid design into the Automotive/industrial sector.

Things not seen in most 8 bit uC ( including the AVRs )

** ECC FLASH [24/32K] ** 16/32 bit MDU ** Cordic ** MultiCAN with 2 nodes ** Mask models

and the more usual stuff too

** 2 UART, fast SPI ** 1.5uS ADC ** Advanced PWM unit all for indicated price from $2.68/10K

-jg

I also archive the complete toolchain source tree and the shellscript used to build it. That way even if you have to switch to a host that can't execute IA32 ELF binaries, you can still build the exact version toolchain you originally used way back when before IA32 went away as a host platform.

If you end up having to develop on a platform for which gcc won't compile, then you're screwed. That seems like a pretty remote possibility.

--
Grant Edwards                   grante             Yow!  I'm wearing PAMPERS!!
                                  at               
                               visi.com

That's ~15 and ~10 years ago, I was talking about the new Atmel single cycle AT89LP family. Nice devices - we have already released 2 designs in them.

Also just seen the release info on the Infineon XC888/886, (also very nice devices), ... and I have the freedom to choose ( same tools :)

-jg

Hey, what about the MARC4 ? [ and the AVR32 ? ].... :)

-jg

You may want to watch your back... Jim Granville could be behind you...

Don't think the AVR will be obsolete anytime soon.

Assuming the same technology, and not too high pincount, AVR chips should always be lower cost than ARM chips simply due to the fact that it has 8 bit datapath. Wide datapaths use quite a lot of silicon.

When you are looking at 100-200 pin chips this does not matter. When you are looking at 8-32 pin chips it does.

--
Best Regards,
Ulf Samuelsson
ulf@a-t-m-e-l.com
This message is intended to be my own personal view and it
may or may not be shared by my employer Atmel Nordic AB