I hope it's a P1+ in same package. But, that's very unlikely... Maybe PropGCC release?
We've been told not to write another line of P2 code until we get a functional chip (or something like that). Of course you can still use the P1 PropellerGCC release.
I am guessing it will be a 64 I/O P1B, perhaps 200MHz/50MIPS per cog.
?? - I think Chip was meaning some software tools/flows work around P1, not new silicon variants.
IIRC there are some parts of the tool chain presently coded by Chip in x86 Assembler ?
This last week, I had to make a detour to accomplish some Prop1 work which will be announced next week. It's going to open doors to people innovating on the current design.
Not that you didn't already have our attention, but you certainly have it now!
refresh...refresh...refresh...
Definitely not expecting a new/variant P1.
Maybe its helping with the PropTool replacement code which Chip wrote in '86 assembler???
Or, its more specs on the counters a vga so we can make those sections do more amazing things
I understand that the entire Spin compiler and assembler in the Prop Tool is written by Chip in x86 assembler.
The PropTool replacement compiler is OpenSpin written by Roy Eltham in C/C++. Reverse engineered from Chip's x86 original Ibelieve.
It boggles my mind that Chip is continuing to work on the compiler for the PII in x86 assembler. I would have expected new developments to build on the new cross-platform code base.
But then it boggles my mind that Chip wrote the compiler in x86 in the first place. I suspect he can work on it faster in x86 than trying to get to grips with C++ and the new code base. Awesome!
There are no more alignment issues for words and longs. The entire memory is addressed using a 19-bit byte-level address. Words and longs can start anywhere.
Chip
Are these words and longs still stored in little endian format?
But that would mean we would have to wait a whole 'nother Christmas for the P2, up from the 2 we are stuck with now! A new P1 should come after the P2 is out of the way, not before!
A new P1 should come after the P2 is out of the way, not before!
Yup. One fish-hook lurking in the 'better P1' arena, is the process needs a core Vcc, which kills retrofit, and once that is gone, there is not much market-space left between a new-design pass, and a P2,
Any suggestion of 64 i/o reduces that market-space even further.
How about P1 as is just redone in Verilog? @Chip said something like 'P1 in Verilog would be surprisingly simple'.
This then done in a smaller process should result in a faster P1, right?
Same pins, same instructions, easy way to upgrade for existing customers. Most of existing tools, languages, OBEX, Documentation would work out of the box.
Maybe reducing rom size and make the rom loadable to have some more hub ram.
That could give some breathing room for existing applications and developers of them to wait for the P2.
A P1 with 128K of RAM would do for me thank you Chip. Everything else can be the same. Then at least I can run a complete stand alone Z80 emulator and CP/M system in a single device. Complete with 80*40 VT100 terminal display on VGA.
You always wanted that stand alone, self hosted development system for the Propeller didn't you Chip?
I'll look forward to placing my order on Wednesday!
A P1 with 128K of RAM would do for me thank you Chip. Everything else can be the same. Then at least I can run a complete stand alone Z80 emulator and CP/M system in a single device. Complete with 80*40 VT100 terminal display on VGA.
You always wanted that stand alone, self hosted development system for the Propeller didn't you Chip?
I'll look forward to placing my order on Wednesday!
Is there a C++ compiler for CP/M that can compile OpenSpin? :-)
I very much doubt it. Never heard of one. Did C++ even exist in CP/M days? Well, yes it did, just.
I don't know what's up with OpenSpin but it's huge and it uses a ton of memory. The stripped executable is 200K and it uses some megs at run time.
Given that C compilers work in the 64K of a CP/M system it must be possible to build a Spin compiler that fits in that space. Spin is a much smaller language. OpenSpin is not that compiler. There is a reason old hands think C++ is huge and bloated and produces huge and bloated code
But hey, this is a Z80 CP/M system, who needs Spin there?
I very much doubt it. Never heard of one. Did C++ even exist in CP/M days? Well, yes it did, just.
I don't know what's up with OpenSpin but it's huge and it uses a ton of memory. The stripped executable is 200K and it uses some megs at run time.
Given that C compilers work in the 64K of a CP/M system it must be possible to build a Spin compiler that fits in that space. Spin is a much smaller language. OpenSpin is not that compiler. There is a reason old hands think C++ is huge and bloated and produces huge and bloated code
But hey, this is a Z80 CP/M system, who needs Spin there?
Yeah, I was just kidding. C++ for CP/M is probably not practical. However, I seem to remember that someone else created an open source Spin compiler in C that was announced shortly after Roy announced his. Was I dreaming that? If not, maybe that could be used as the basis of a CP/M Spin compiler if we really want to go that way. I myself like CP/M back in its day but I don't have any desire to go back to it for anything resembling serious development work. It's fun as a retro hack though.
I don't recall anything about about another Spin compiler in C. May be. Anyone have any idea?
I don't seriously propose going back to CP/M but...
It amuses me that after three decades or more of SGML, XML, HTML, XHTML, back to HTML, and other over complex, tortuous and brain damaged mark up languages, a lot of the worlds blogs and other media is now using things like markdown. See for example github or indeed this forum.
Basically a return to the simplicity of the mark up used by the WordStar word processor of the CP/M days.
Turns out that nobody ever needed the complexity of MicroSoft Word in the first place!
....
I don't know what's up with OpenSpin but it's huge and it uses a ton of memory. The stripped executable is 200K and it uses some megs at run time.
Given that C compilers work in the 64K of a CP/M system it must be possible to build a Spin compiler that fits in that space. Spin is a much smaller language. OpenSpin is not that compiler. There is a reason old hands think C++ is huge and bloated and produces huge and bloated code
...
We have a Spin compiler that works right on the Propeller itself with its 32kB Memory. All you need is an SD card for the source files and the binary output. It is a self hosting system with editor and shell. It's called Sphinx and I don't understand why it gets constantly ignored....
The Sphinx compiler is now also included in Spinix - which is another self host
I have heard of Sphinx, I have not tried it out. It sounds like a totally amazing achievement.
I agree, I had my tongue firmly in cheek when I talking about CP/M on the Propeller.
Why does Sphinx get ignored? I don't know. I guess if you have Propeller and a PC and and the Prop Tool or SimpleIDE or whatever and a PropPlug and you want to get something done you don't need Sphinx.
Sphinx is great - it allowed me to do self-hosted development on PropCade.
Michael did a great service to the propeller community in writing it - and I always found it a pity that it was largely ignored.
I think one of the major reasons for a lack of uptake was the minimal sales of propeller boards that had enough resources to run it - specifically, boards lacking a uSD socket, or built-in video/kb interfaces. The highest selling propeller boards did not include everything needed, and the boards that had all the needed features tended to cost more than the most inexpensive boards from parallax... which is what people tended to buy the most.
Frankly, I was very surprised that Parallax did not push Sphinx, or adopt it.
I've seen very little evidence that people really want to do self-hosted Propeller development.
Maybe someone needs to run polls like:
Do you use a self-hosted development Propeller environment often?
Have you used a self-hosted development Propeller environment often?
Which self-hosted development Propeller environment do you use?
Doubt any of that would have much value though since full participation is rarely expected.
Comments
We've been told not to write another line of P2 code until we get a functional chip (or something like that). Of course you can still use the P1 PropellerGCC release.
?? - I think Chip was meaning some software tools/flows work around P1, not new silicon variants.
IIRC there are some parts of the tool chain presently coded by Chip in x86 Assembler ?
refresh...refresh...refresh...
Maybe its helping with the PropTool replacement code which Chip wrote in '86 assembler???
Or, its more specs on the counters a vga so we can make those sections do more amazing things
refresh...refresh...refresh...;)
The PropTool replacement compiler is OpenSpin written by Roy Eltham in C/C++. Reverse engineered from Chip's x86 original Ibelieve.
It boggles my mind that Chip is continuing to work on the compiler for the PII in x86 assembler. I would have expected new developments to build on the new cross-platform code base.
But then it boggles my mind that Chip wrote the compiler in x86 in the first place. I suspect he can work on it faster in x86 than trying to get to grips with C++ and the new code base. Awesome!
From what I understand, Chip has all the P1 design running in Verilog.
So, some musings of what could be done quickly for a new P1...
Where's the PROCEED TO CHECKOUT button?
With Chip's new 2 clock instruction cycle mechanism and simple P1 Cogs I would think 200+ MHz (100+ MIPS per cog) would be possible? Drool.....
Are these words and longs still stored in little endian format?
But that would mean we would have to wait a whole 'nother Christmas for the P2, up from the 2 we are stuck with now! A new P1 should come after the P2 is out of the way, not before!
Yup. One fish-hook lurking in the 'better P1' arena, is the process needs a core Vcc, which kills retrofit, and once that is gone, there is not much market-space left between a new-design pass, and a P2,
Any suggestion of 64 i/o reduces that market-space even further.
In this general space, one problem is the FPGAs with enough RAM, tend to be now in BGA only.
Yes.
How about P1 as is just redone in Verilog? @Chip said something like 'P1 in Verilog would be surprisingly simple'.
This then done in a smaller process should result in a faster P1, right?
Same pins, same instructions, easy way to upgrade for existing customers. Most of existing tools, languages, OBEX, Documentation would work out of the box.
Maybe reducing rom size and make the rom loadable to have some more hub ram.
That could give some breathing room for existing applications and developers of them to wait for the P2.
Just speculating here. Or dreaming?
Enjoy!
Mike
Ken tells me Wednesday is the day.
And that's Thursday here
So it's probably some additional software or specification info.
Would be nice to be a new P1 though.
You always wanted that stand alone, self hosted development system for the Propeller didn't you Chip?
I'll look forward to placing my order on Wednesday!
Nothing like twisting an arm or two haha
Only one question - above or below the spousal approval threshold?
I very much doubt it. Never heard of one. Did C++ even exist in CP/M days? Well, yes it did, just.
I don't know what's up with OpenSpin but it's huge and it uses a ton of memory. The stripped executable is 200K and it uses some megs at run time.
Given that C compilers work in the 64K of a CP/M system it must be possible to build a Spin compiler that fits in that space. Spin is a much smaller language. OpenSpin is not that compiler. There is a reason old hands think C++ is huge and bloated and produces huge and bloated code
But hey, this is a Z80 CP/M system, who needs Spin there?
I don't seriously propose going back to CP/M but...
It amuses me that after three decades or more of SGML, XML, HTML, XHTML, back to HTML, and other over complex, tortuous and brain damaged mark up languages, a lot of the worlds blogs and other media is now using things like markdown. See for example github or indeed this forum.
Basically a return to the simplicity of the mark up used by the WordStar word processor of the CP/M days.
Turns out that nobody ever needed the complexity of MicroSoft Word in the first place!
We have a Spin compiler that works right on the Propeller itself with its 32kB Memory. All you need is an SD card for the source files and the binary output. It is a self hosting system with editor and shell. It's called Sphinx and I don't understand why it gets constantly ignored....
The Sphinx compiler is now also included in Spinix - which is another self host
I have heard of Sphinx, I have not tried it out. It sounds like a totally amazing achievement.
I agree, I had my tongue firmly in cheek when I talking about CP/M on the Propeller.
Why does Sphinx get ignored? I don't know. I guess if you have Propeller and a PC and and the Prop Tool or SimpleIDE or whatever and a PropPlug and you want to get something done you don't need Sphinx.
Michael did a great service to the propeller community in writing it - and I always found it a pity that it was largely ignored.
I think one of the major reasons for a lack of uptake was the minimal sales of propeller boards that had enough resources to run it - specifically, boards lacking a uSD socket, or built-in video/kb interfaces. The highest selling propeller boards did not include everything needed, and the boards that had all the needed features tended to cost more than the most inexpensive boards from parallax... which is what people tended to buy the most.
Frankly, I was very surprised that Parallax did not push Sphinx, or adopt it.
I'd love to see a P2 version of Sphinx!!!
Besides OpenSpin, I know of only Sphinx as HomeSpun as Spin compilers, and HomeSpun is written in C#
If anyone has spare time... a nice plain ANSII C Spin compiler would be a great project
http://propeller.wikispaces.com/Homespun+Spin+Compiler
I've seen very little evidence that people really want to do self-hosted Propeller development.
Maybe someone needs to run polls like:
Do you use a self-hosted development Propeller environment often?
Have you used a self-hosted development Propeller environment often?
Which self-hosted development Propeller environment do you use?
Doubt any of that would have much value though since full participation is rarely expected.