I am no compiler writer and have no idea if that PASM plus hub stack model will ever prove to be useful.
But I imagine it's a good place to start for an anyone who has a experience retargetting GCC but has only just been introduced to the Prop. Isn't targetting raw machine code a natural instinct for a C compiler writer?
This is exactly the case. It is a stepping stone for porting the back end. An LMM VM will be integrated in the near future. Once we have a functional P1 product, we will revisit the first model and decide if COG "only-ness" is worth the extra effort.
If it was made available to developers, possibly under an NDA, lots of interesting stuff might be available by the time the P2 is launched, independent of this official project. It shouldn't impact progress on the P1 compiler.
If it was made available to developers, possibly under an NDA, lots of interesting stuff might be available by the time the P2 is launched, independent of this official project. It shouldn't impact progress on the P1 compiler.
Sure. You're welcome to try that approach. Good luck. The best way to such information would have been to accept invitations to participate.
...not unless you only ever expect to write programs less than 500 instructions long
I seem to remember that where I once worked there were guys building code with a C compiler for weeny little PICs with about that much code space. Crazy I know but it worked for some little tasks for them and they were happy. Those PICs were peripheral to ARMs where the heavy lifting occurred.
You should be able to compile the heater_fft to less than 500 instructions. Yeah, yeah, I know, is that useful?
Let's see if anyone else is interested. Perhaps groups of people could come up with projects they could work on together, as with this one, but on a more informal basis.
Ross, We were told not to release any P2 stuff yet. Sorry.
Yup ... and I suspect that's why Parallax is going to lose some of its long time supporters. This site may soon be full of nothing more than the odd fanboi .... and Leon, of course!
Hey Leon! ... tell us some more about the XMOS! I believe it supports GCC as well as your own proprietary version of C ...?
.. and I suspect that's why Parallax is going to lose some of its long time supporters.
It is said in the open source software world "release early and release often". And I also feel that any such P2 simulator or first stabs at GCC targets should be out in the open.
But here is the thing. This is not just software we are talking about its the instruction set of the P2 processor hardware. That instruction set may well not be fixed yet. So releasing a simulator for it can be seen as premature.
Now you could argue that the P2 instruction set should be debated in the open prior to it being nailed down. Well, a lot of that has gone on already. In fact with Parallax and the P2 it may be the first time in history that a processor producer has had such an open debate with its customers about the design of the next generation device.
This is so fascinating to watch that this long time supporter is not going anywhere. Apart from the odd XMOS detour of course:)
This is so fascinating to watch that this long time supporter is not going anywhere.
Yes, I agree. It's a bit like driving past a road accident - try as you might to avert your gaze, you just can't help slowing down and taking a peek ...
I have a few clarifications on the current GCC status. The first test program implemented a serial port and printed "Hello World". It also converted a binary number to ASCII, and printed that as well. The second test program was a minor modification of the first one, and it uncovered an issue with addressing stack variables. We haven't implemented the assembler or linker yet, so we are using a small program to convert the assembly output from GCC to a file that the Prop tool can compile.
The P2 simulator is only partially completed. We are waiting for more detail on the instruction set before work can continue on the simulator. For the most part, the simulator implements the instructions that have been openly published on this forum. There is no detail on how the new instructions affect the Z and C flags, so the current implementation either ignores the flags or is based on guesses on how they would work. Most of the I/O, video and counter instructions have not been implemented. We have run a slightly modified Spin interpreter on the P2 simulator. We have also run a test program that exercises the PTRA, PTRB, CLUT and index instructions.
From what I can tell, Parallax personnel have not spent much time on the GCC project. They provide guidance, but most of the work is being done by contractors and volunteers. Steve is doing a great job coordinating the efforts.
The current COG PASM model seems to be the right thing to do at this point. It is easier to generate native PASM code than to generate LMM PASM. The C language relies on the stack to save registers, return address, local variable, etc. Stack usage can be minimized by selecting a higher optimization level and avoiding deep levels of function calls.
As all others discussions was without any invitations
> NOW it is shame Parallax changed theirs MIND.
And as long IT needs invitations --- I'm not interested --- I can wait THAT can give me much more to Criticize.
As I don't like that some people will control Parallax what NEED be done. Even if them DON'T understand Propeller's working principles!
As all others discussions was without any invitations
> NOW it is shame Parallax changed theirs MIND.
And as long IT needs invitations --- I'm not interested --- I can wait THAT can give me much more to Criticize.
As I don't like that some people will control Parallax what NEED be done. Even if them DON'T understand Propeller's working principles!
Nothing has changed. How exactly are you qualified to port GCC to Propeller?
NOW as You ask --- You will know in time You publish and I start criticize.
That question on Hardware/Software on PropII I wanted with start of its design NOT in time it is not possible to change anything to work better!
So Build Yours Titanic -- If it will have small holes Maybe it sail ----> BUT if I find one big HOLE I will sink it!
You didn't answer the question.
I'm as sure now as I was then that you are not qualified to port GCC.
In more than one post in the past you complained bitterly about C programming.
Why would I want to invite anyone like this in to an effort involving C?
Maybe your attitude about that has changed? Show me some of your C code.
I can only say NOW --- I started my programing in 1976 - Write Monitor and Basic interpreter written in I8085 assembly language.
After that even ported Pascal to it.
Write some debugger's and other things
Have worked with some Industrial Control systems Build/Programing and after it service more that 10 Years.
BUT as I said before You publish Yours Compiler -- I will never say How much more I have worked on.
BUT reread last meaning from previous post - "BUT if I find one big HOLE I will sink it!" -- I have never said that if I have don't know what I talk on!
GCC is a very well defined tool-chain with specific requirements to meet.
We will meet the requirements for GCC with Propeller specific modes of operation.
Maybe you can tell us how GCC works? That way we'll all know you assessments are valid.
If you want to continue your violent ranting, you can talk about the holes in GCC.
I can only say NOW --- I started my programing in 1976 - Write Monitor and Basic interpreter written in I8085 assembly language.
After that even ported Pascal to it.
Write some debugger's and other things
Have worked with some Industrial Control systems Build/Programing and after it service more that 10 Years.
BUT as I said before You publish Yours Compiler -- I will never say How much more I have worked on.
BUT reread last meaning from previous post - "BUT if I find one big HOLE I will sink it!" -- I have never said that if I have don't know what I talk on!
BUT before You talk more --- Describe usable Programing MODEL on Propeller !!!.
Before that -- Not possible to talk more on this
Ps. Will never say You what experiences I have on Compilers/Interpreters. As it is not point in this discussion.
In time I need write my CW that is OK -- But as I'm to sick to seek any work that is not necessary for me.
GCC is a very well defined tool-chain with specific requirements to meet.
We will meet the requirements for GCC with Propeller specific modes of operation.
Maybe you can tell us how GCC works? That way we'll all know you assessments are valid.
If you want to continue your violent ranting, you can talk about the holes in GCC.
In fact with Parallax and the P2 it may be the first time in history that a processor producer has had such an open debate with its customers about the design of the next generation device.
Based on the way some of these discussions are going, it may very well be the last time, too. It seems that the open debate has caused some amount of emotional investment from the community, and that there may be some "negative waves" as a result.
So much negativity going on here. What is going on with the GCC project and P2 stuff is by Parallax's choice, they are doing what they feel is best for their business right now. I trust Ken, Chip, and everyone involved there enough, based on their track record, to produce something good. I assume most of you are hear because you like Parallax. So, why are you crapping on their lawn?
Not to read the mind of Parallax but I think their stance would/should be something along the lines of:
"The Official GCC porting team is a collection of invited and selected contractors/consultants and volunteers who access to pre-release, preliminary, volatile information due to their involvement in this effort and due to their participation to a special NDA. When the Propeller 2 specifications are to a point where Parallax feels comfortable with their stability, NDA's will be offered to other interested parties for their collaboration in the project and preparation for a successful launch of a new product."
Nowhere in the Parallax mind did I read anything about "the rest of us begging for scraps". They are managing a significant, expensive, (hopefully) well thought out product launch in support of a new division. It's a careful balance between enough ACCURATE information released and too mcuh information released outside the inner circle.
Is that Parallax's official stance, that the GCC-porting team gets special privileges and the rest of us can beg for scraps?
Seriously? Parallax is owned by the Graceys... not anybody else on these forums. Why is there such an entitlement mentality here, that somehow they "owe" it to the "forums" to share every bit of detail about their business. What's next? People demanding that they post their general ledeger?
Seriously people. If anybody (in general) feels they can do a better job than Parallax, then maybe the best course is to start your own company. That way, you can run it however you like. You can be as open or closed as you like, Smile on anybody's lawn that you like, bash Leon all you like... what.freaking.ever.you.like.
I'll say this... I wouldn't blame Parallax if they just said forget about it and called the whole show off. First people complained that there's no Parallax supported C tools. Now they're developing them, and people are complaining about that. Un.freaking.real.
Parallax put out a call for developers. Some people aren't qualified for the work... that's life. Other's didn't want to get involved... they made their choice. Get over it already.
Anyways, for those working on this project, thanks for the effort.
Parallax put out a call for developers. Some people aren't qualified for the work... that's life. Other's didn't want to get involved... they made their choice. Get over it already.
Anyways, for thsoe working on this project, thanks for the effort.
First people complained that there's no Parallax supported C tools. Now they're developing them, and people are complaining about that. Un.freaking.real.
Parallax put out a call for developers. Some people aren't qualified for the work... that's life. Other's didn't want to get involved... they made their choice. Get over it already.
Anyways, for those working on this project, thanks for the effort.
I am not complaining that they're developing tools. I just think there are people who might not have the skills or interest to be involved with GCC but could still make significant contributions, especially if given access to a P2 simulator or whatever. I don't really consider it a question of "entitlement" <eyeroll>.
mpark, don't take all of that as directed towards yourself. A lot of that is just my interpretation on what I've seen at times.
The best I can say is just go with the flow and get involved where you can, even if it's not the "best" opportunity. And if there's no opportunity that fits, make your mark down the road when the "chip currently codenamed the Propeller2" hits the streets.
My only point of contention, one which I relinquish, was the idea that the GCC team is getting special privileges. Well, they're being paid to do a job, one for which they have the skills. I would have liked to be involved in that capacity, but I don't have the experience, so I didn't offer. I offered to help where I do have the experience. Some people didn't offer. That's okay.
I'm one of those people who did not offer to help with the GCC thrust. I have no experience with C and have, thus far, not encountered a need for it. But the GCC project interests me intensely. Why? Because I need to see where Parallax's dev resources are receiving priority, and I want to make sure that Spin and PASM receive an equitable mindshare. So far, I've been given no cause for alarm.
My only critique might be for jazzed to set a more inclusive, rather than exclusive, tone. Frankly, some of his comments of late seem a bit imperious. (Sorry, man, but I had to say it.) As Parallax's front man for this effort, diplomacy has to trump everything else. (Yeah, I know: who am I to be preaching diplomacy, right? ) There's a lot riding on this, along with an abundance of forumistas with a huge investment in the outcome. So it's not surprising that some of us can be a little touchy about perceived slights.
Comments
Then why not publish it?
Ross.
Ross, We were told not to release any P2 stuff yet. Sorry.
I seem to remember that where I once worked there were guys building code with a C compiler for weeny little PICs with about that much code space. Crazy I know but it worked for some little tasks for them and they were happy. Those PICs were peripheral to ARMs where the heavy lifting occurred.
You should be able to compile the heater_fft to less than 500 instructions. Yeah, yeah, I know, is that useful?
Yup ... and I suspect that's why Parallax is going to lose some of its long time supporters. This site may soon be full of nothing more than the odd fanboi .... and Leon, of course!
Hey Leon! ... tell us some more about the XMOS! I believe it supports GCC as well as your own proprietary version of C ...?
Ross.
It is said in the open source software world "release early and release often". And I also feel that any such P2 simulator or first stabs at GCC targets should be out in the open.
But here is the thing. This is not just software we are talking about its the instruction set of the P2 processor hardware. That instruction set may well not be fixed yet. So releasing a simulator for it can be seen as premature.
Now you could argue that the P2 instruction set should be debated in the open prior to it being nailed down. Well, a lot of that has gone on already. In fact with Parallax and the P2 it may be the first time in history that a processor producer has had such an open debate with its customers about the design of the next generation device.
This is so fascinating to watch that this long time supporter is not going anywhere. Apart from the odd XMOS detour of course:)
Yes, I agree. It's a bit like driving past a road accident - try as you might to avert your gaze, you just can't help slowing down and taking a peek ...
Ross.
The P2 simulator is only partially completed. We are waiting for more detail on the instruction set before work can continue on the simulator. For the most part, the simulator implements the instructions that have been openly published on this forum. There is no detail on how the new instructions affect the Z and C flags, so the current implementation either ignores the flags or is based on guesses on how they would work. Most of the I/O, video and counter instructions have not been implemented. We have run a slightly modified Spin interpreter on the P2 simulator. We have also run a test program that exercises the PTRA, PTRB, CLUT and index instructions.
From what I can tell, Parallax personnel have not spent much time on the GCC project. They provide guidance, but most of the work is being done by contractors and volunteers. Steve is doing a great job coordinating the efforts.
The current COG PASM model seems to be the right thing to do at this point. It is easier to generate native PASM code than to generate LMM PASM. The C language relies on the stack to save registers, return address, local variable, etc. Stack usage can be minimized by selecting a higher optimization level and avoiding deep levels of function calls.
> NOW it is shame Parallax changed theirs MIND.
And as long IT needs invitations --- I'm not interested --- I can wait THAT can give me much more to Criticize.
As I don't like that some people will control Parallax what NEED be done. Even if them DON'T understand Propeller's working principles!
Nothing has changed. How exactly are you qualified to port GCC to Propeller?
This question are some month to late!
NOW as You ask --- You will know in time You publish and I start criticize.
That question on Hardware/Software on PropII I wanted with start of its design NOT in time it is not possible to change anything to work better!
So Build Yours Titanic -- If it will have small holes Maybe it sail ----> BUT if I find one big HOLE I will sink it!
I'm as sure now as I was then that you are not qualified to port GCC.
In more than one post in the past you complained bitterly about C programming.
Why would I want to invite anyone like this in to an effort involving C?
Maybe your attitude about that has changed? Show me some of your C code.
I can only say NOW --- I started my programing in 1976 - Write Monitor and Basic interpreter written in I8085 assembly language.
After that even ported Pascal to it.
Write some debugger's and other things
Have worked with some Industrial Control systems Build/Programing and after it service more that 10 Years.
BUT as I said before You publish Yours Compiler -- I will never say How much more I have worked on.
BUT reread last meaning from previous post - "BUT if I find one big HOLE I will sink it!" -- I have never said that if I have don't know what I talk on!
GCC is a very well defined tool-chain with specific requirements to meet.
We will meet the requirements for GCC with Propeller specific modes of operation.
Maybe you can tell us how GCC works? That way we'll all know you assessments are valid.
If you want to continue your violent ranting, you can talk about the holes in GCC.
Yes -- I'm are dumb --- Only You know all !!!
BUT before You talk more --- Describe usable Programing MODEL on Propeller !!!.
Before that -- Not possible to talk more on this
Ps. Will never say You what experiences I have on Compilers/Interpreters. As it is not point in this discussion.
In time I need write my CW that is OK -- But as I'm to sick to seek any work that is not necessary for me.
Based on the way some of these discussions are going, it may very well be the last time, too. It seems that the open debate has caused some amount of emotional investment from the community, and that there may be some "negative waves" as a result.
So much negativity going on here. What is going on with the GCC project and P2 stuff is by Parallax's choice, they are doing what they feel is best for their business right now. I trust Ken, Chip, and everyone involved there enough, based on their track record, to produce something good. I assume most of you are hear because you like Parallax. So, why are you crapping on their lawn?
Note to self... avoid UPEW...
I'm sorry that you are ill. I hope you can get well soon.
So, why are you crapping on their lawn?
Originally Posted by Kevin Wood
Maybe not AVOID, just watch carefully where you step!
http://www.youtube.com/watch?v=d-diB65scQU&ob=av3e
C.W.
Is that Parallax's official stance, that the GCC-porting team gets special privileges and the rest of us can beg for scraps?
"The Official GCC porting team is a collection of invited and selected contractors/consultants and volunteers who access to pre-release, preliminary, volatile information due to their involvement in this effort and due to their participation to a special NDA. When the Propeller 2 specifications are to a point where Parallax feels comfortable with their stability, NDA's will be offered to other interested parties for their collaboration in the project and preparation for a successful launch of a new product."
Nowhere in the Parallax mind did I read anything about "the rest of us begging for scraps". They are managing a significant, expensive, (hopefully) well thought out product launch in support of a new division. It's a careful balance between enough ACCURATE information released and too mcuh information released outside the inner circle.
Seriously? Parallax is owned by the Graceys... not anybody else on these forums. Why is there such an entitlement mentality here, that somehow they "owe" it to the "forums" to share every bit of detail about their business. What's next? People demanding that they post their general ledeger?
Seriously people. If anybody (in general) feels they can do a better job than Parallax, then maybe the best course is to start your own company. That way, you can run it however you like. You can be as open or closed as you like, Smile on anybody's lawn that you like, bash Leon all you like... what.freaking.ever.you.like.
I'll say this... I wouldn't blame Parallax if they just said forget about it and called the whole show off. First people complained that there's no Parallax supported C tools. Now they're developing them, and people are complaining about that. Un.freaking.real.
Parallax put out a call for developers. Some people aren't qualified for the work... that's life. Other's didn't want to get involved... they made their choice. Get over it already.
Anyways, for those working on this project, thanks for the effort.
I am not complaining that they're developing tools. I just think there are people who might not have the skills or interest to be involved with GCC but could still make significant contributions, especially if given access to a P2 simulator or whatever. I don't really consider it a question of "entitlement" <eyeroll>.
The best I can say is just go with the flow and get involved where you can, even if it's not the "best" opportunity. And if there's no opportunity that fits, make your mark down the road when the "chip currently codenamed the Propeller2" hits the streets.
My only point of contention, one which I relinquish, was the idea that the GCC team is getting special privileges. Well, they're being paid to do a job, one for which they have the skills. I would have liked to be involved in that capacity, but I don't have the experience, so I didn't offer. I offered to help where I do have the experience. Some people didn't offer. That's okay.
My only critique might be for jazzed to set a more inclusive, rather than exclusive, tone. Frankly, some of his comments of late seem a bit imperious. (Sorry, man, but I had to say it.) As Parallax's front man for this effort, diplomacy has to trump everything else. (Yeah, I know: who am I to be preaching diplomacy, right?
-Phil