DPRG
DPRG List  



[DPRG] C Compiler

Subject: [DPRG] C Compiler
From: Rick J. Bickle rbickle at swbell.net
Date: Thu Aug 28 16:14:01 CDT 2003

Kip,

I didn't realize that you could purchase just the chips.
I'll check it out.

Thanks,
Rick J. Bickle
Integrated Control Systems
Voice:  (972) 522-1593
Fax:    (972) 660-7950
Mail:   PO BOX 541058
        Grand Prairie, TX. 75054-1058
Email:  rbickle at intconsys.com 
Web:    www.intconsys.com


-----Original Message-----
>From: dprglist-admin at dprg.org [mailto:dprglist-admin at dprg.org] On Behalf
Of Kipton Moravec
Sent: Thursday, August 28, 2003 3:31 PM
To: Rick J. Bickle
Cc: DPRG List
Subject: RE: [DPRG] C Compiler


I fully understand your problem.   I hate banked memory as a general 
rule.  But hear me out.

The Rabbit 2000 has a 1 MB linear space for program and data memory with

glueless interfaces for flash and SRAM.

If you use 128K for Flash you can use the rest of the 1 MB memory for
data 
(unbanked).

If you need more memory, then the R2000 has an I/O space. Separate from
the 
program and data space.  What I was proposing was to use the program and

data space for your programming and variable location. But put an 
additional 1MB in the I/O space for only the table.  You only access it 
when you need a value from the table. All other variables are in the 
regular (unbanked) data memory.  What it means is that you have accessor

functions for the table data but the rest of the system works like
normal. 
(Think disk access, only faster)  No banking for 97% of the code.

You can use the rabbit chips and not use the modules.  I built
prototypes 
with modules, and am going to switch to the processors (no modules) when
we 
go into production.  It allowed me to get the prototypes and software
ready 
faster.  I was able to test most code with just the development system,
and 
I was able to make a cheap prototype board for the custom HW to test
that 
out.  When I was sure everything would work, then I made the production 
board.  The other nice thing about the modules is that they give you the

full schematic, so you can move the module circuit onto your production
board.

Again my opinion.

Kip



At 12:20 PM 8/28/03, you wrote:
>Kip,
>
>Those rabbit modules look really cool for fast development - short run 
>applications. Unfortunately, the product I am designing is projected to

>sell several thousand, which might make the rabbit stuff cost 
>prohibitive. Once the development tools are paid for, the XA-G49 chips 
>are about $8.00 each. The whole board should cost under $100.00.
>
>The reason I'm trying to stay away from banked memory, is that I've 
>done it before. I took a DS80C320 processor and accessed 1MB with it in

>32Kx8 banks. The actual hardware for this was simple, but the software 
>got very complex. The banked memory was used for a database 
>application, and records were stored, sometimes overlapping memory 
>banks. Furthermore, performing the bank calculation on every byte 
>stored slowed the entire process to a crawl. I eventually had to write 
>a special algorithm which monitored the distance from the beginning or 
>end of a bank, used direct access with bank offsets to read/write table

>data if the record did not overlap. If a record did overlap banks, it 
>had to write the number of bytes up to the end of one bank, switch 
>banks, and write the remaining data into the next bank. All of this for

>variable length records with variable length key fields.
>
>I got tired of beating my head on the computer. :)
>
>Rick J. Bickle
>Integrated Control Systems
>Voice:  (972) 522-1593
>Fax:    (972) 660-7950
>Mail:   PO BOX 541058
>         Grand Prairie, TX. 75054-1058
>Email:  rbickle at intconsys.com
>Web:    www.intconsys.com
>
>
>-----Original Message-----
>From: dprglist-admin at dprg.org [mailto:dprglist-admin at dprg.org] On 
>Behalf Of Kipton Moravec
>Sent: Thursday, August 28, 2003 11:35 AM
>To: Rick J. Bickle
>Cc: DPRG List
>Subject: RE: [DPRG] C Compiler
>
>
>I was just curious.
>
>The Rabbit can address 1M total (without bank switching)  The compiler 
>is included with the development kit. $279.  It is very fast to develop
>code
>with. I was surprised.
>
>The R2000 has 4 serial ports and the R3000 has 6.  I would probably 
>recommend the RCM2100 and use the IO space for a bank selected 1M 
>memory
>
>for just the table.  Use the regular RAM for the program data.  You can

>battery backup the SRAM (and external SRAM) with the circuitry built 
>in.  It would depend on how many you are building.  If it is less than 
>100 I would probably go with the rabbit module.  You will save a lot of
>time.
>(which translates into money).
>
>The data in the table will not be accessed as often, and can go in a 
>banked memory.  Just have accessor functions that read and write the 
>banked memory.  Then only the table will be banked and the rest of the 
>memory is
>normal.  (You can also do this with a 64K processor.  Use 32K as
regular
>
>RAM and bank the high 32K segment.)
>
>My opinion.
>
>Kip
>
>At 09:32 AM 8/28/03, you wrote:
> >Kip,
> >
> >I have selected the XA for several reasons. The application that I'm 
> >building requires large amounts of SRAM (1Mx8) to store table 
> >information. The XA can access this much memory directly, without 
> >resorting to bank switching circuits. Also, the XA-G49 has internal 
> >64Kx8 of flash, which can be programmed via the main UART. This will 
> >be
>
> >a really nice feature for my customers to use for upgrading the 
> >firmware if necessary. My project also requires 2 UARTS, which are 
> >built into the XA.
> >
> >If you know of another device with the same or similar capabilities, 
> >please let me know.
> >
> >I just got off the phone this morning with a guy selling the 
> >raisonance
>
> >compiler. This one works with both 8051 and XA for about $1300. This 
> >hitech compiler is about $800, and it is XA only.
> >
> >Rick J. Bickle
> >Integrated Control Systems
> >Voice:  (972) 522-1593
> >Fax:    (972) 660-7950
> >Mail:   PO BOX 541058
> >         Grand Prairie, TX. 75054-1058
> >Email:  rbickle at intconsys.com
> >Web:    www.intconsys.com
> >
> >
> >-----Original Message-----
> >From: Kipton Moravec [mailto:kip at kdream.com]
> >Sent: Wednesday, August 27, 2003 6:07 PM
> >To: Rick J. Bickle
> >Cc: DPRG List
> >Subject: Re: [DPRG] C Compiler
> >
> >
> >I have not used that compiler.  I have used the Keil and it is pretty

> >expensive.
> >
> >I have looked at the XA family and can not justify it.  I was 
> >wondering
>
> >why you are choosing it.  What about it do you like, that would 
> >justify spending that kind of money on a compiler?  How much is the 
> >Hightech compiler now?
> >
> >I have been using the 8051 variants from Philips for a long time, 
> >never
>
> >could go to the XA.
> >
> >Kip
> >
> >
> >At 05:10 PM 8/27/03, you wrote:
> >
> > >Hey Everyone,
> > >
> > >I was wondering if anyone had experience with the Hightech C 
> > >compiler. I was looking at purchasing it for use with the Philips 
> > >XA family. It's about half the cost of other compilers I have seen.
> > >
> > >Thanks,
> > >Rick J. Bickle
> > >Integrated Control Systems
> > >Voice:  (972) 522-1593
> > >Fax:    (972) 660-7950
> > >Mail:   PO BOX 541058
> > >         Grand Prairie, TX. 75054-1058
> > >Email:  rbickle at intconsys.com
> > >Web:    <file://www.intconsys.com>www.intconsys.com
> >
> >
> >_______________________________________________
> >DPRGlist mailing list
> >DPRGlist at dprg.org http://nimon.ncc.com/mailman/listinfo/dprglist
>
>
>_______________________________________________
>DPRGlist mailing list
>DPRGlist at dprg.org http://nimon.ncc.com/mailman/listinfo/dprglist
>
>_______________________________________________
>DPRGlist mailing list
>DPRGlist at dprg.org http://nimon.ncc.com/mailman/listinfo/dprglist


_______________________________________________
DPRGlist mailing list
DPRGlist at dprg.org
http://nimon.ncc.com/mailman/listinfo/dprglist


More information about the DPRG mailing list