IAR and memory banking for 128K flash and 8051

"Amir" skrev i meddelandet news: snipped-for-privacy@d4g2000prg.googlegroups.com...

You could also consider moving into the 21st century like the rest of us by selecting an MCU appropriate for the job.

--
Best Regards,
Ulf Samuelsson
This is intended to be my personal opinion which may,
or may not be shared by my employer Atmel Nordic AB

"I'm trying to extract information from a prisoner using the rack, but I'm having real trouble tying him onto it tightly enough. I checked the scrolls in the monastery, but didn't get any useful help. So I'm throwing it out to all my fellow executioners"

"You should move into the 21st century and waterboard him".

Translation: Yes, 8051 is old, but if it's what he's using...

True. He might even be using a AT89C51RE2 (new 128K Flash 89C51), from Umm/Err... who was that vendor again ?? :)

He may have a PLCC44 socket he needs to upgrade, with a new chip. Care to suggest alternatives for that ?

-jg

There are a lot of people who want to Bank with 8052's. I do not know why. I inherited on once, but it was with Keil. If you do not get an answer here, An I assume you Post at IAR and look for samples there, and in the sample folder that the compiler installed. try asking at

( they may also ask why?)

IAR has(had) a banked switched memory model in the 1980's 1990's vintage C compilers. I made a lot of money fitting large programs into small address spaces. The compiler option I am familiar with was a -m (or perhaps -m0).

The CPU's memory space needed one bank to alway be in place and not switchable and then another bank to be switchable.

The compiler/linker sorted out all the bank switching parameters. If a call was made from a routine in one bank to a routine in another bank, the new bank number and routine address was passed to a bank switching routine in the common bank. All returns from routines also were done through this bank switching routine.

The hardware needed a register to store the bank number and that register fed the memory decode logic in my designs.

Hope this helps George

Thank you Larwe...I assume u picked the easiest answer...lol. But just to let u know that I am using a SOC which is pretty recent (2006) and it has an 8051 embedded. Brief: If you dont know the answer, or cannot provide any help, u dont have to write useless comments!

Thank you for the useful information George. I appreciate it..I use a SOC (cc2430 ), and according to cc2430 datasheet, the banking has to be done in 4 banks as opposed to 2. I did setup the parameters as described by IAR compiler documents, but when I program the flash, the system doesnt seem to find the start up code, thus nothing happens.

Regards, Amir

hy.

FYI Neil..Many SOC vendors, use these microncontrollers as they are small, and well developped/fit for embedded designs using SOCs.

Regards, Amir

And what else has multiple suppliers?

--
 [mail]: Chuck F (cbfalconer at maineline dot net) 
 [page]: 
            Try the download section.

So, then, which other architecture offers 2nd source? FYI your own employer does not seem to believe that the 8051 is obsolete:

Quote: "Atmel offers a broad range of microcontrollers based on the 8051 architecture. ..."

For example, I have used an Atmel 89C51 in a design in the early 90's and that is still in production. No end in sight so far. Because it works.

--
Regards, Joerg

http://www.analogconsultants.com/

Maybe I should put the signature in BOLD letters...

--
Best Regards,
Ulf Samuelsson
This is intended to be my personal opinion which may,
or may not be shared by my employer Atmel Nordic AB

I know it's your personal opinion, just wanted to make a point about why the 8051 architecture is still widely used. And unless another

2nd-sourceable series appears it will remain in use. That would be my personal opinion, and that of a whole lot of others.

--
Regards, Joerg

http://www.analogconsultants.com/

Hasn't the ARM7 been call the "8051" of the new millennium.

Too bad the ARM7 is just the cpu and not the entire configuration i.e. peripherals.

donald

The prices of both cores are asymptotically approaching zero, but they really don't 100% occupy the same application space.

The 128K in this thread is at the top-end of the 80C51 range, (ASIX make some larger, with Ethernet) but there is a LOT of R&D effort going into 80C51 in the far-east now :

Companies like Winbond, Megawin, Myson, Micronas, Coreriver, STC, Syncmos, ABOV, (etc) are expanding their offerings.

Devices like the new 12Vin/12V I/O Micronas easyLIN will generate a lot of traction.

Many of these are one clock or turbo cores.

The highest ADC performance devices use 80C51 cores, and dominate metering.

The 80C51 will still be (easily) the worlds highest volume 8 bit uC core.

-jg

Yep, it sure will be. I even found one in our pellet stove.

--
Regards, Joerg

http://www.analogconsultants.com/

Only if 2nd sources are coming from companies like Winbond. Else, no. Just my humble thoughts.

--
Regards, Joerg

http://www.analogconsultants.com/

I like the CPU Core. Banking is a way to force it to do more than it was designed to do. Banking adds Issues and slows the CPU down. That said it does work. I would guess that SOC use it since it is popular, Has good tools and has an lot of space to add control register.

In the end it is the core you have. Go with it.

Here's a bit of unrelated free advice:

When you join a newsgroup, remember that you are joining an established community. A lot of the people here have been around for years - if you don't understand the banter between old hands like Ulf and Lewin, don't comment on it, as it just makes you look silly. And learn to spell "you" - you'll be amazed at the increased level of respect you get.

mvh.,

David