Beyond one's own nose (STM32 Primer)
deSilva
Posts: 2,967
This seems to be more than a toy: www.stm32circle.com/resources/stm32primer.php
The ARM CORTEX M3 is considered to be the new "32 bit reference"; does anyone know how it compares to the Propeller?
They say this Primer-Development System is $32 - all inclusive (128x128 screen, USB,...).
----
Edit: Looked up prices (w/o VAT):
- 25 British Pounds
- 50 US $
- 37 Euro
----
Edit2:
Some benchmarks here (C: ATmega/8MHz against CORTEX/72 MHz = 1 against 170) www.stm32circle.com/forum/viewtopic.php?id=122
They can toggle a pin in C @20 MHz (= 4 instructions?)
Post Edited (deSilva) : 3/25/2008 11:14:12 AM GMT
The ARM CORTEX M3 is considered to be the new "32 bit reference"; does anyone know how it compares to the Propeller?
They say this Primer-Development System is $32 - all inclusive (128x128 screen, USB,...).
----
Edit: Looked up prices (w/o VAT):
- 25 British Pounds
- 50 US $
- 37 Euro
----
Edit2:
Some benchmarks here (C: ATmega/8MHz against CORTEX/72 MHz = 1 against 170) www.stm32circle.com/forum/viewtopic.php?id=122
They can toggle a pin in C @20 MHz (= 4 instructions?)
Post Edited (deSilva) : 3/25/2008 11:14:12 AM GMT
Comments
Although I could only find it for $50.75 which still seems very reasonable compared to other dev kits.
As for comparison with the Prop, Baggers has done quite a bit of ARM development, I'm sure he will know how well it compares against the Prop.
Good to see you are still knocking around these forums anyway
Best regards,
Coley
Post Edited By Moderator (Chris Savage (Parallax)) : 4/1/2008 5:35:04 PM GMT
It is a nice looking dev system but what about general purpose IO? I'm also not sure if the processor has an external memory interface or just the standard uart, twi, spi etc.
And for a site with some info about the actual core www.realviewtools.com/products/CPUs/ARM_Cortex-M3_CM3Core.html
But it is interesting that these are on offer from RTOS in the UK: is there something about COGs that make the difference?
Ryton Mike
With the prop it is really easy to set-up serial ports and things compared to the AVR micros where you have to learn what a billion different registers do and which ones you have to change and in which order and what else it will affect and the list goes on. I would assume that this is also true to a large extent in the ARM chips.
It is interesting to note that the ARM core supports conditional execution on some instructions like the propeller. There are a couple of interesting examples of using the conditional instructions on wikipedia that show they can give a fair amount of increase in execution speed and smaller code size.
But I had also to learn that it will hardly enter the realm of so-called (!) professional microcontrolling, which is defined by:
- C programming
- fixation on recurrent cost, mostly disregarding needed development effort (*)
- need for comprehensive development environment, including version control, H/W supported debugging, and simulation, nearly regardless of cost.
(*) This might need an explanation: Project Managers just do not believe that one controller can be programmed "faster" than another. One reason is that 2/3 of development cost is considered to consist in top level design, documentation, and debugging - which is basically true... Re-usable components are considered a real asset to help here (which is also basically true), but this will increase the threshold for any change of H/W architecture considerably....
Jaja, but obsolescence helps a lot here. Most uC of the year 2000 have been replaced with newer versions or similar products. Just because the device is not cost-effective to manufacture or because the architecture reached a limit or because they wanted to sell more chips.
8086 -> 80186 -> 80286 -> 80386 -> 80486 -> Pentium -> Pentium II -> P3 -> P4 -> Centrino -> Duo -> Duo II -> ..
Or:
PIC12 -> PIC14 -> PIC16 -> PIC18 -> PIC24
The PIC32 is however something else
It looks pretty cool and includes an accelerometer, a pretty nice OLED display and two usb ports (one for programming, one for interfacing to a USB master). There are a few sample user built apps out there that look somewhat interesting. http://www.stm32circle.com/projects/project.php?id=31
I haven't really done much other than turn it on and try the fairly lame sample apps and check out the available IO pins. I am thinking of trying to get an Xbee module inside the case and use as a remote display/control device. The other nodes on the Zigbee network will probably be Propellers.
I don't think it will replace the prop for me, but it is an interesting self contained package.
--Nathan
$32 USD = £16 GBP ... Not bad for what one gets, cheapest way I can see of getting a full 128x128 colour display.
Anyone know where to buy these in the UK ? The online shop link from www.st.com doesn't seem to go anywhere.
uk.farnell.com/1502504/semiconductors/product.us0?sku=STMICROELECTRONICS-STM3210B-PRIMER
Wonder if it can be hacked to squezze a Prop inside
Post Edited By Moderator (Chris Savage (Parallax)) : 4/1/2008 5:34:48 PM GMT
http://www.rapidonline.com/Electronic-Components/Integrated-Circuits/ST-Microcontrollers/STM32-Primer/80672
Always had good service off them, I recomend,
Ryton Mike
gets added over dollar to sterling conversion rates ? Us folk in the UK sure do seem to get
Smile deals. It's the same on Propeller kit. Understandable for smaller companies with shipping
costs but those like ST must have an infrastructure to move their parts around the world. At
least Microchip usually sell at straight USD to GBP conversion prices.
Avnet (US) stock it at $33.98, Avnet (UK) don't appear to.
@tax, I thought all Europeans got taxed to death.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Paul Baker
Propeller Applications Engineer
Parallax, Inc.
Post Edited (Paul Baker (Parallax)) : 3/26/2008 2:27:14 AM GMT
The very last step in a product life cycle is called "retirement", and for some time known as being even more critical in our short-lived business...
Maybe better than a 10,000% national budget deficit.... But take care: one also can try both....
http://store.buglabs.net/
Gee, I've seen this happen so many times it made me sick ... literally ... chronic stress syndrome.
What I've also seen is that "programmers" care too much about getting it done they leave the "future enhancements" entirely for the next sucker rather than designing an extensible architecture. This of course in a later time·furthers the desire to just get it done because it's already such a mess. Some years later, you end up with something that is impossible for adding new features, constantly breaks with new commits, requires two days just to build, un-nerving times just to boot, three days for code reviews, a week of regression testing, and a month of "watchlist" followup. When someone tells you "yes, but the long build time gives me the opportunity to multitask ..." you know you're in trouble. The company often wonders why the best talent is impossible to recruit and keep [noparse]:)[/noparse]
A metamorphisis usually happens afterwards .... Leveraged buy-out because the company is worth more functionally split up than as one operation. Customers switch wholly to other company's product lines regardless of the institutional imperative (no one ever got fired buying from IBM). Steve Jobs comes back to give an extreme makeover after years of deterioration (best case).
Hewlett-Packard's signs on campus have (or had in 2002 at least) "invent" prominently displayed. This is a good mission statement with many layers of meaning.
One meaning is: "re-invent" ... if you don't re-invent to use emerging technologies or methods when the opportunity comes, you lose a competetive advantage that can be obtained with·the addition of fresh perspective from the bright young MSCS. Management not allowing for the "re-invent" is shorting it's own fate and hurting long term stock holder value. A machine in a manufacturing site needs good oil; so does the vehicle of technological advance: intellectual property.
Time to market is often key to secure the right to re-invent. So it's ok to let the product schedule drive when competition is overwhelming as it was for internet infrastructure in the 90's. The problem is that many people can not step back and define sustainable after being in a given situation so long ... stuck in a rut ... to understand the need to re-invent especially after it took so much hard work to achieve present results.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
jazzed·... about·living in·http://en.wikipedia.org/wiki/Silicon_Valley
Traffic is slow at times, but Parallax orders·always get here fast 8)
Future electronics has the primer for $35
http://www.componentsuperstore.com/Store/ProductDetails.aspx?ProductIdentifier=STM3210BPRIMERSTMICRO7117757
Thanks,
Doug
Some very good observations! Though a little bit off-topic
I might - also off-topic - add this anecdote from more than 10 years ago.... It's not "maintenance" but "change of paradigm". And it's Davd vs. Goliath.
Our competitor was a "global player". They could have summoned as much power as they wished for this development. They used - of course - "PC-technology" (in the way IBM started their PC: use parts from the winners). Plus tons of nearly working software. It was $1000+ per piece.
We used an at that time quite new concept (FPGA) for the most demanding functionality, put nearly everything in software, and used some very clever algorithms. It run on a tuned 8-bit processor. It was $100+ per piece, and worked. It was incredible, and it took a long time to explain to the customer that this WAS possible and that is was NOT a fake. He asked questions as: "But isn't it risky to use an FPGA? But is it professional to not use C++? But is it future-oriented to use an 8-bit processor? But is it safe to use an algorithm not described in the literature?"
We never had asked those questions... We wanted to get the thing done....
I learned there some of the many aspects of product management ....
It could become a success because we had been a small team , and my boss was the owner of that company, and he had a "vision", and he technically UNDERSTOOD what we were doing.
But I think there is more: When building things from today's S/W libraries, operating systems, OTS components, you are aggregating MUCH more than you need. My personal opinion is that you REALLY need only 5% of the code of a typical PC kiosk application. I understand well that we are no longer living in a time where you ask: "Why do you need that line of code?" (Though I had to, when doing code analysis for safety critical software
And the answer might well be this: "Well, it is called from that and that module, and it had been always there, and everybody needs it, and in fact I don't really know, and just look for yourself..."
1 It uses a fair amount of power so it has only run for an hour or so from when it says fully charged. I am guessing that the on screen battery indicator is not very accurate at all, and when I charge it overnight it will last quite a bit longer, but not as long as a typical PDA because the screen gets quite warm.
2 The batteries overheat if charged too long (over 10 hours)
3 The cover does not latch, and kind of just sits on there, This is remedied by a piece of tape, but not very elegant.
4 It's orange. I am in Knoxville, but I do not like orange.
5 The development platform only runs on windows.
Doug
There is a lot of step-up and step-down regulation
2,4 V -> 3,3 V -> 1,8 V
So I should expect 50% is lost, which means it needs 800 mA for operation... The clock is quite high, there are peripherial chips.
But no, this is not within the expected range, I should say...