Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.lang.forth > #28906 > unrolled thread

Announcing the first stable release of Mecrisp-Stellaris for ARM Cortex M

Back to article view | Back to comp.lang.forth

Contents

  Announcing the first stable release of Mecrisp-Stellaris for ARM Cortex M Matthias Koch <matthias.koch@hot.uni-hannover.de> - 2014-03-04 12:04 +0100
    Re: Announcing the first stable release of Mecrisp-Stellaris for ARM Cortex M Matthias Koch <matthias.koch@hot.uni-hannover.de> - 2014-03-26 15:57 +0100
      Re: Announcing the first stable release of Mecrisp-Stellaris for ARM Cortex M rickman <gnuarm@gmail.com> - 2014-04-05 00:55 -0400
        Re: Announcing the first stable release of Mecrisp-Stellaris for ARM Cortex M Paul Rubin <no.email@nospam.invalid> - 2014-04-04 23:04 -0700
          Re: Announcing the first stable release of Mecrisp-Stellaris for ARM Cortex M rickman <gnuarm@gmail.com> - 2014-04-05 04:16 -0400
            Re: Announcing the first stable release of Mecrisp-Stellaris for ARM Cortex M Paul Rubin <no.email@nospam.invalid> - 2014-04-05 12:52 -0700
              Re: Announcing the first stable release of Mecrisp-Stellaris for ARM Cortex M rickman <gnuarm@gmail.com> - 2014-04-05 19:39 -0400
                Re: Announcing the first stable release of Mecrisp-Stellaris for ARM Cortex M Paul Rubin <no.email@nospam.invalid> - 2014-04-05 18:27 -0700
                  Re: Announcing the first stable release of Mecrisp-Stellaris for ARM Cortex M rickman <gnuarm@gmail.com> - 2014-04-05 22:28 -0400
                    Re: Announcing the first stable release of Mecrisp-Stellaris for ARM Cortex M Paul Rubin <no.email@nospam.invalid> - 2014-04-05 20:25 -0700
                      Re: Announcing the first stable release of Mecrisp-Stellaris for ARM Cortex M rickman <gnuarm@gmail.com> - 2014-04-06 00:37 -0400
                        Re: Announcing the first stable release of Mecrisp-Stellaris for ARM Cortex M Paul Rubin <no.email@nospam.invalid> - 2014-04-05 22:59 -0700
                          Re: Announcing the first stable release of Mecrisp-Stellaris for ARM Cortex M rickman <gnuarm@gmail.com> - 2014-04-06 03:08 -0400
                            Re: Announcing the first stable release of Mecrisp-Stellaris for ARM Cortex M Paul Rubin <no.email@nospam.invalid> - 2014-04-07 19:26 -0700
                              Re: Announcing the first stable release of Mecrisp-Stellaris for ARM Cortex M rickman <gnuarm@gmail.com> - 2014-04-08 00:13 -0400
                                Re: Announcing the first stable release of Mecrisp-Stellaris for ARM Cortex M Paul Rubin <no.email@nospam.invalid> - 2014-04-07 21:41 -0700
                              Re: Announcing the first stable release of Mecrisp-Stellaris for ARM Cortex M Mark Wills <markwills1970@gmail.com> - 2014-04-08 02:20 -0700
                                Re: Announcing the first stable release of Mecrisp-Stellaris for ARM Cortex M Paul Rubin <no.email@nospam.invalid> - 2014-04-08 05:59 -0700
                                  Re: Announcing the first stable release of Mecrisp-Stellaris for ARM Cortex M rickman <gnuarm@gmail.com> - 2014-04-09 00:41 -0400
          Re: Announcing the first stable release of Mecrisp-Stellaris for ARM Cortex M Mark Wills <markwills1970@gmail.com> - 2014-04-08 02:24 -0700
            Re: Announcing the first stable release of Mecrisp-Stellaris for ARM Cortex M Paul Rubin <no.email@nospam.invalid> - 2014-04-08 06:15 -0700
              Re: Announcing the first stable release of Mecrisp-Stellaris for ARM Cortex M Mark Wills <markwills1970@gmail.com> - 2014-04-08 06:46 -0700
              Re: Announcing the first stable release of Mecrisp-Stellaris for ARM Cortex M Matthias Koch <matthias.koch@hot.uni-hannover.de> - 2014-04-08 17:00 +0200
                Re: Announcing the first stable release of Mecrisp-Stellaris for ARM Cortex M Paul Rubin <no.email@nospam.invalid> - 2014-04-08 08:55 -0700
                  Re: Announcing the first stable release of Mecrisp-Stellaris for ARM Cortex M Mark Wills <markwills1970@gmail.com> - 2014-04-09 00:50 -0700
        Re: Announcing the first stable release of Mecrisp-Stellaris for ARM Cortex M visualforth@rocketmail.com - 2014-04-05 10:37 -0700
        Re: Announcing the first stable release of Mecrisp-Stellaris for ARM Cortex M Matthias Koch <matthias.koch@hot.uni-hannover.de> - 2014-04-08 13:09 +0200
          Re: Announcing the first stable release of Mecrisp-Stellaris for ARM Cortex M rickman <gnuarm@gmail.com> - 2014-04-09 00:50 -0400

Page 2 of 2 — ← Prev page 1 [2]

#29441

Mark Wills <markwills1970@gmail.com> writes:
>> This would be an awesome Forth target because of the FRAM.
>
> Paul, do you have a link for this? I searched the TI site but no
> luck. TI's website is a bit of a nightmare to be honest.

http://www.ti.com/tool/MSP-BNDL-FR5969LCD

Typing "ti launchpad fram" into a search engine finds some more pages
too.

Matthias - sorry, I didn't mean to hijack the Mecrisp thread.  Maybe
this unrelated discussion should move to a new thread.  The FRAM device
really seems just about perfect for Forth though.

[toc] | [prev] | [next] | [standalone]

#29443

Thank you, Paul. And yes, apologies to Matthias for de-railing this thread.

[toc] | [prev] | [next] | [standalone]

#29445

> Matthias - sorry, I didn't mean to hijack the Mecrisp thread.  Maybe
> this unrelated discussion should move to a new thread.  The FRAM device
> really seems just about perfect for Forth though.

No matter, feel welcome :-) FRAM is a nice technology, but it appears to me as read cycles are limited because of a write-on-read operation like old days core memory. There is a cache for not wearing out FRAM cells in tight loops, but Forth jumps around a lot - maybe this causes cache misses continously and wearing of memory cells beyond of what is expected in datasheets.

One note: TIs FRAM devices contain a 16/20 bit mixed address length CPU core called "CPUX" which is currently not supported by Mecrisp. As the extended instruction set is downstream binary compatible to the classic MSP430 CPU, you may have luck, but you won't have a change to use the upper address range.

Matthias

[toc] | [prev] | [next] | [standalone]

#29447

Matthias Koch <matthias.koch@hot.uni-hannover.de> writes:
> FRAM is a nice technology, but it appears to me as read cycles are
> limited because of a write-on-read operation like old days core
> memory.

Oh that's interesting, I knew there was a small amount of write wear
(they mention 1e15 cycles) but I didn't realize every read also did a
write.  Still, at 10 mhz (1e7 hz), that's still several years of
continuously hitting a single cell.  By putting the innermost loops in
SRAM, I think things should be fine.

> One note: TIs FRAM devices contain a 16/20 bit mixed address length
> CPU core called "CPUX" which is currently not supported by Mecrisp. As
> the extended instruction set is downstream binary compatible to the
> classic MSP430 CPU, you may have luck, but you won't have a change to
> use the upper address range.

Aha, I knew there were some big 430's that have extended range.  The
early FRAM parts had 16k of FRAM and 1k or so of SRAM, so 16 bit
addressing would seem like enough.  The one on the Launchpad has 64k of
FRAM and 2k of SRAM.  If that means a tiny bit of the FRAM is
unreachable (or needs some CODE words to reach) then that's fine.  If it
means the SRAM is unreachable I guess that's more of a problem.

In this day and age anything needing more than 64k of address space
should probably use a 32 bit cpu.  I hope they make some ARM devices
with FRAM at some point.

[toc] | [prev] | [next] | [standalone]

#29484

On Tuesday, 8 April 2014 16:55:05 UTC+1, Paul Rubin  wrote:
> Oh that's interesting, I knew there was a small amount of write wear
> (they mention 1e15 cycles) but I didn't realize every read also did a
> write.  Still, at 10 mhz (1e7 hz), that's still several years of
> continuously hitting a single cell.  By putting the innermost loops in
> SRAM, I think things should be fine.

There's some info on that in the Launchpad Users Guide:

3.3.4.1 Understanding the Numbers Behind the FRAM Speed Application

Every 250 ms, 1.8MB of FRAM are programmed with a pattern. Hence the speed of FRAM writes is calculated as 7.564 MB/s. The FRAM is written to in blocks of 1K bytes; it is this 1KB block that is subject to the lifetime FRAM endurance specification. FRAM endurance of block = E = 10^15 write cycles. This is a minimum specification for FRAM endurance found in the device data sheet.

... Time taken to exceed FRAM endurance @1.8MB/250ms : >4000 years.

http://www.ti.com/lit/ug/slau535/slau535.pdf

This particular Launchpad board with this MCU looks yummy. So are we saying that Mecrisp runs on this chip? I'll go RTFM ;-)

[toc] | [prev] | [next] | [standalone]

#29301

On Saturday, April 5, 2014 12:55:55 AM UTC-4, rickman wrote:
> 
> Maybe I confuse easily, but I made the mistake of digging out my 
> Stellaris Launchpad and trying to find docs on it.  Seems like TI had 
> done a purge and there is nearly no trace of the Stellaris docs on their 
> web site.  The Stellaris Launchpad comes with no docs of any sort other 
> than the quick start flysheet.  I went online and found nothing at the 
> TI site.  

I am sure this is the NIH syndrom. 

Stellaris is/was a microprocessor brand of Luminary Micro, Inc:
http://www.crunchbase.com/company/luminary-micro

TI later took over in 2009:
TI acquires Luminary Micro
http://www.electronics-eetimes.com/en/ti-acquires-luminary-micro?cmp_id=7&news_id=217500258

ARM welcomes TI's acquisition of Luminary
http://www.eetimes.com/document.asp?doc_id=1253768

[toc] | [prev] | [next] | [standalone]

#29437

> Still I find the conversion a bit confusing.  Did they just rename the 
> Stellaris chips or are the Tiva C chip "equivalents" not exactly the 
> same devices?  It looks like the Launchpad is not exactly the same.

Pardon me for late reply, but you can basically substitute LM4F120 by TM4C123 datasheets. 
As far as I recall, the chips have upgraded USB capabilities, but are almost identical otherwise 
and binary compatible. If you like, I can send you the original datasheets for Stellaris Launchpad 
and LM4F120 chip per Email, I still have them. Contact me directly, my address is valid.

Yes, the Tiva board with TM4C1294 in the target list is the "Connected Launchpad" with Ethernet.

Matthias

[toc] | [prev] | [next] | [standalone]

#29480

On 4/8/2014 7:09 AM, Matthias Koch wrote:
>> Still I find the conversion a bit confusing.  Did they just rename the
>> Stellaris chips or are the Tiva C chip "equivalents" not exactly the
>> same devices?  It looks like the Launchpad is not exactly the same.
>
> Pardon me for late reply, but you can basically substitute LM4F120 by
> TM4C123 datasheets. As far as I recall, the chips have upgraded USB
> capabilities, but are almost identical otherwise and binary compatible.
> If you like, I can send you the original datasheets for Stellaris
> Launchpad and LM4F120 chip per Email, I still have them. Contact me
> directly, my address is valid.
>
> Yes, the Tiva board with TM4C1294 in the target list is the "Connected
> Launchpad" with Ethernet.

Thanks for the reply.  I found a manual for either the Stellaris chip or 
the Stellaris launchpad or both, I can't recall.  I'm sure the chip data 
sheet is around given all the data sheet web sites.  It can be hard to 
find the manufacturer's site for all the datasheet sites sometimes.

I think I understand the chips are not literally identical with new 
names, but have very minor differences which should only be improvements 
from what I have seen.  No loss of functionality.

-- 

Rick

[toc] | [prev] | [standalone]

Page 2 of 2 — ← Prev page 1 [2]

Back to top | Article view | comp.lang.forth

csiph-web