forth, Cross compiler wanted

Cross compiler wanted

Author	Message
Brad Rodrigue #1 / 5	Cross compiler wanted Quote: > I'm trying to put together a Forth cross compiler for an 8-bit dual stack > machine > called the Neuron Chip. The hardware architecture is so Forth-like, that > it > makes sense to use in-line word expansion and direct threaded code, rather > than the traditional implementation. > It's also necessary to separate code, volatile data, and non-volatile data > into separate > segments to be located in ROM, RAM and EEPROM respectively. I've done a clumsy implementation of this in my Chromium cross-compiler. Unfortunately, this is still a work in progress, and currently runs under F83. (You can find it, and source for a 6809 Forth target, in <ftp://ftp.forth.org/pub/Forth/Camel/cam09-10.zip>.) Quote: > All of the PD implementations that I have seen run the compiler > on the target, generate indirect threaded code, and mix the dictionary > headers with the compiled code and the data declarations. Even the ANSI > standard itself seems to require the compiler itself to execute on the > target. This is, alas, true of the ANSI standard. The committee decided that NO cross-compiled Forth could meet the standard, due to the requirement for interpretive behaviors. But there are several cross-compilers which don't run on the target, generate direct-threaded or subroutine-threaded code, and omit the dictionary headers. One such is TCOM, included in F-PC. It'll probably be easier to modify TCOM than to modify my Chromium. You might also consider a commercial product, which would do all you ask and then some, with documentation and support to boot. Offhand, I don't know if any of the Forth vendors support the Neuron chip. -- Brad Rodriguez T-Recursive Technology : Computers on the Small Scale Contributing Editor, The Computer Journal... http://www.--.com/ ~tcj Secretary, Forth Interest Group.......... http://www.--.com/
Sun, 02 Apr 2000 03:00:00 GMT

Harold Rabbi #2 / 5	Cross compiler wanted I'm trying to put together a Forth cross compiler for an 8-bit dual stack machine called the Neuron Chip. The hardware architecture is so Forth-like, that it makes sense to use in-line word expansion and direct threaded code, rather than the traditional implementation. It's also necessary to separate code, volatile data, and non-volatile data into separate segments to be located in ROM, RAM and EEPROM respectively. All of the PD implementations that I have seen run the compiler on the target, generate indirect threaded code, and mix the dictionary headers with the compiled code and the data declarations. Even the ANSI standard itself seems to require the compiler itself to execute on the target. This target doesn't have standard character stream I/O or a mass storage device, so the associated words for compiling and file management would only be needed at compile time. Does anybody have experience with an ANSI or F83 cross compiler? It doesn't have to be written in Forth. TIA -- Harold Rabbie Saratoga, CA Remove .NOSPAM from my address before replying
Sun, 02 Apr 2000 03:00:00 GMT

b.. #3 / 5	Cross compiler wanted Quote: > I'm trying to put together a Forth cross compiler for an 8-bit dual stack > machine > called the Neuron Chip. The hardware architecture is so Forth-like, that > it > makes sense to use in-line word expansion and direct threaded code, rather > than the traditional implementation. Do you mean sub-routine threading using the CPUs native call/return instructions? Quote: > It's also necessary to separate code, volatile data, and non-volatile data > into separate > segments to be located in ROM, RAM and EEPROM respectively. > All of the PD implementations that I have seen run the compiler > on the target, generate indirect threaded code, and mix the dictionary > headers with the compiled code and the data declarations. Even the ANSI > standard itself seems to require the compiler itself to execute on the > target. > This target doesn't have standard character stream I/O or a mass storage > device, so the > associated words for compiling and file management would only be needed at > compile time. It sounds like what you could use is an umbilical system, where the compiler lives in a PC and emits code into an eprom emulator running on the target. Forth Inc and MPE build such systems for many CPUs. In exchange for money they can create one for the neuron chip. I only heard about this neuron chip yesterday and I am not aware of a PD Forth system for it. -- Brian Fox Web site: http://www.microtronix.ca/~bfox/ MICROTRONIX SYSTEMS LTD. Ph: (519) 659-9500, Ext 259 200 Aberdeen Dr. FAX (519) 659-8500 London Ont. Canada N5V 4N2
Sun, 02 Apr 2000 03:00:00 GMT

Harold Rabbi #4 / 5	Cross compiler wanted Quote: > > I'm trying to put together a Forth cross compiler for an 8-bit dual stack machine > > called the Neuron Chip. The hardware architecture is so Forth-like, that it > > makes sense to use in-line word expansion and direct threaded code, rather > > than the traditional implementation. > Do you mean sub-routine threading using the CPUs native call/return instructions? Either in-line copying for short code words, or else use the native call/return for longer words, or when space is at a premium. Quote: > It sounds like what you could use is an umbilical system, where the > compiler lives in a PC and emits code into an eprom emulator running on the target. The host-target communications are done already in silicon. One of the neat features of the Neuron Chip is that it includes a 7-layer network protocol stack and it can easily be interfaced to a variety of media such as twisted pair, radio and power line. The chip includes EEPROM, so you don't need a ROM emulator. Quote: > I only heard about this Neuron Chip yesterday and I am not aware of a PD Forth system for it. Exactly why I want to do it. The only available development system is C-based and costs at least $5K. -- Harold Rabbie Saratoga, CA Remove .NOSPAM from my address before replying
Sun, 02 Apr 2000 03:00:00 GMT

Page 1 of 1

[ 5 post ]