Even if you can’t understand everything that Shannon wrote about, I recommend taking a look. He throws in all sorts of observations. I believe at one point when discussing entropy he made a comment about the written English language and its impact on crossword puzzles.
I believe he was really into robotic maze solving mice too.
I don't understand what you mean by "compile feature", but other than that, it already has everything you listed.
It has debugging hardware, and people already have written debuggers, tested on the FPGA version, that use the debugging hardware to debug other cogs. It has 512 KB HUBRAM, which could be increased to 1MB if there were room on the die. Each of the 8 cogs (there could be 16 cogs if there were room on the die) has 2 KB COGRAM, which can't be increased due to the architecture (we've had this discussion many times) and 2KB LUTRAM (which was settled on long ago). Chip Gracey worked with the authors of the PropGCC compiler to make sure compilers would be able to generate efficient code, and yet it still has one of the two simplest assembly languages I've ever seen (the other being P1 assembly - I haven't yet had a chance to properly compare them and see which is actually easier because I don't have an FPGA).
Chip Gracey doesn't need any more stress. He already survived one heart attack. We don't want any more features - we just want the P2 to be done and released already so we can get our hands on it.
Of course there is a "compile" feature in the P1 and will be in the P2. In Spin or C and no doubt other languages.
Please do specify what you mean by "hardware debugging"? I have never found the traditional "run, stop, single step, break point" debuggers very useful. Often incapable of finding the bugs I have or even getting in the way of finding them. Even more so with multi-processor systems.
The P2 design has lots of program and data memory.
Of course optimized C will be supported. And assembler.
If there is anything else you need some help with you let me know!
That depends on what you mean by a simulator ... There has already been a hardware simulator for many months ... runs on any of several FPGA systems (depending on tradeoffs between capability (number of cogs and amount of memory) and cost and has been used for debugging hardware and software. Gets updated each time the logic (in high level language) gets updated. Parallax neither needs to write a simulator nor has the resources available to do so. There are many other pieces that are needed more. If you're volunteering, by all means do the work. You'll get to use it first ... and design and implement it the way you like.
There are real tradeoffs between die size and costs. A given package will only hold a die up to a specific size. Beyond that and the packaging costs go way up. Stacked dies can be used, but the cost again goes way up. We're not talking here about processors selling for hundreds of dollars or more where these large stacked dies and correspondingly huge packages can be used. These subjects have been discussed before in these Propeller 2 threads along with simulators, debuggers, power requirements and distribution across the chip, heat load, etc.
Before you make too many other suggestions, you might do a little more background checking into what has been suggested or discussed before. How big is Parallax really and what kinds of things have they done before and how? How are they spending their time and energy and who else is helping? How do they compare to some of the other names you've mentioned (Intel, AMD, and ???).
FYI There are a set of P2 specs, info about the instruction set, a set of different FGPA boards that emulate the current P2 (with options that fit those FPGA boards). The discussions for P2 features have been going on for 12? Years.
There was a list of 5 requirements that existing P1 customers wanted. Speed, Hub RAM, I/O Pins, Security were four of those requirements. IIRC ADC was the fifth. Security has just been scuttled.
You mention "parallel processing". The Propeller 1 is a parallel processor. It has 8 independent cores / cogs. These are normally used to implement I/O devices like UARTs, video displays, servo motor controllers, etc., but can be used easily to implement "user processes" as well. The Propeller 2 currently has 8 cogs although the design originally was for 16 ... scales by powers of 2 ... 16 just wouldn't fit on the die with all the other stuff (mostly smart I/O pin controllers). All the programming languages (Spin, C, Forth) support the multiple cogs, but you have to do it explicitly.
"simulator in the editor" ... why do you need it? If it's to shorten the modify / try it again cycle, Parallax has always emphasized the equivalent cycle using real hardware, often because you want to see how real hardware behaves since most simulators stop at the edge of the chip ... you don't have attached LEDs or switches or other devices / sensors / motors. There's even a Forth interpreter that runs on the Propeller 1 ... directly supports the multiple cogs ... extremely high performance ... used in commercial products.
"Working out sections of code" ... Remember that the Propeller is not a Raspberry Pi or a PC. A Propeller is usually hooked up to something in the way of hardware that it has to interact with in real time for your "sections of code" to work. It's much more helpful to set aside an I/O pin for a video (NTSC) output and assign two cogs for a memory / status real-time display that's started as the first thing in your program initialization and will continue running even if your program crashes. You could use ViewPort, but often that's too generalized for realistic debugging. I've used the video output scheme for debugging assembly as well ... a memory card I/O driver ... not really doable with a simulator.
I don't know about "free" as in you don't have to pay. As far as I know Viewport is a Windows only program and not open source. As such it it off my radar.
ViewPort is not free. I believe there's a limited time demo mode though.
"Already made my 'clean sheet' comments about Propeller 2"
What 'clean sheet' comments? Have you looked at the instruction set? What about the functionality of the smart I/O pins? What kinds of specialized I/O or execution are supported? These all speak to what markets the Propeller 2 would appeal to. Generalized suggestions without context are not very helpful. A simulator for just the processor of an embedded system type device without timing-accurate simulation of a wide range of external (and internal) peripherals is not useful in practice and, if done properly, is a huge undertaking.
The Propeller 1 is in fact used in some commercial products and is not just an educational tool.
I do very much hope the P2 is actually decided. We have seen too many changes of direction and delays already.
I was wondering about that 'derivative' thinking thing. Well, that is what humans do. We are social animals and we learn from each other. If you were brought up in isolation you would not even have a language, you would probably not think anything much. Everything we think, say and do is derived from those that went before us.
Comments
I'm so inadequate with only my DEO Nano
It's a bit much to say Shannon destroyed thermodynamics.
I believe he was really into robotic maze solving mice too.
Maybe the Propellor 2 team is motivated by stress.
Here are some stress risers for you.
Could you please include:
Compile feature
Hardware debugging
Lots of program and data memory.
For the hub and the cores.
Optimized for C programming and retain Assembler programming.
That's about it for now!
If there is anything else you need some help with you let me know!
It has debugging hardware, and people already have written debuggers, tested on the FPGA version, that use the debugging hardware to debug other cogs. It has 512 KB HUBRAM, which could be increased to 1MB if there were room on the die. Each of the 8 cogs (there could be 16 cogs if there were room on the die) has 2 KB COGRAM, which can't be increased due to the architecture (we've had this discussion many times) and 2KB LUTRAM (which was settled on long ago). Chip Gracey worked with the authors of the PropGCC compiler to make sure compilers would be able to generate efficient code, and yet it still has one of the two simplest assembly languages I've ever seen (the other being P1 assembly - I haven't yet had a chance to properly compare them and see which is actually easier because I don't have an FPGA).
Chip Gracey doesn't need any more stress. He already survived one heart attack. We don't want any more features - we just want the P2 to be done and released already so we can get our hands on it.
The Propeller 2 team is exactly one person. See here: https://www.parallax.com/company/about-us and be prepared to be amazed. A page that does not even mention the Propeller chip. But see here http://www.parallax.com/propeller/qna/Content/HomeTopics/WhyWorks.htm to be even more amazed.
Of course there is a "compile" feature in the P1 and will be in the P2. In Spin or C and no doubt other languages.
Please do specify what you mean by "hardware debugging"? I have never found the traditional "run, stop, single step, break point" debuggers very useful. Often incapable of finding the bugs I have or even getting in the way of finding them. Even more so with multi-processor systems.
The P2 design has lots of program and data memory.
Of course optimized C will be supported. And assembler. Yeah, right.
If it is made from scratch why is there a limit?
Second solution:
Use two of these 'dies'
Or three.Ask Intel or AMD how it's done. Beats me if Arm does it.
They could make different versions.
The 'as is' current thinking version with these limits just to get it done and a Propeller 3 or 2 Professional version.
Parallax can use these ideas for free.
Next one is going to cost them some SX-Blitzes and SX-Key's.
BTW SX-Key is a 'hardware debugger'.Class is over for today.
Now your getting in the swing of things!
Now a message for our young readers.
Technology sector is 'Innovation without limits'.
If you can imagine it you can make it.
That's enough of that!
Note on multi hardware cores.
It is a teaching platform.
It does not have to fit in anything like something with an industrial use or
compete in a market segment (automotive for example).
So it can be any size.
It's sort of pure research with the teaching angle thrown in.
A simulator written by Parallax.
In the real world they could just 'ship it' the way they have it figured out and
it would probably work out okay.
EDIT A hardware emulator for the Propeller 2. Just to let you know we think big down here in Southern California.
There are real tradeoffs between die size and costs. A given package will only hold a die up to a specific size. Beyond that and the packaging costs go way up. Stacked dies can be used, but the cost again goes way up. We're not talking here about processors selling for hundreds of dollars or more where these large stacked dies and correspondingly huge packages can be used. These subjects have been discussed before in these Propeller 2 threads along with simulators, debuggers, power requirements and distribution across the chip, heat load, etc.
Before you make too many other suggestions, you might do a little more background checking into what has been suggested or discussed before. How big is Parallax really and what kinds of things have they done before and how? How are they spending their time and energy and who else is helping? How do they compare to some of the other names you've mentioned (Intel, AMD, and ???).
Seems like you think Chip has the P2 all wrong?
Maybe he should start again ???
Cluso
Let me tell you what I know.
They are coming out with a Propellor 2. Period.
Any more info about it is a mystery to me.
Just making conversation. They can build whatever they want.
I know nothing about how it is projected to be built.
Maybe these are ideas for a Super Propeller.
Anybody connected some Propeller 1's together?
Gotten some parallel processing going?
Okay. It's just I heard die size.
If it is 'set in stone' why even bring it up?
I mean simulator in the editor. Okay. If they don't want to do it for whatever reason then 'That's that!'.
I am still kidding about adding extra features at what seems like this late date.
Sounds like the Propeller 2 in near ready.
If anybody wants to learn about Emulation that AVR ICE basic PCB version for $50 is a good deal.
That Cypress has some reasonable ones too. Don't know if they are FPGA's or what.
Anything else is probably pretty brutal on the budget.
Remember it's all the same thing.
There was a list of 5 requirements that existing P1 customers wanted. Speed, Hub RAM, I/O Pins, Security were four of those requirements. IIRC ADC was the fifth. Security has just been scuttled.
"simulator in the editor" ... why do you need it? If it's to shorten the modify / try it again cycle, Parallax has always emphasized the equivalent cycle using real hardware, often because you want to see how real hardware behaves since most simulators stop at the edge of the chip ... you don't have attached LEDs or switches or other devices / sensors / motors. There's even a Forth interpreter that runs on the Propeller 1 ... directly supports the multiple cogs ... extremely high performance ... used in commercial products.
IF I were to look at what the proposed Propeller 2 is tentatively intended to be.
That would be 'derivative' thinking.
Take a look in a parking lot of cars at a shopping center.
See how they all look the same.
That's 'derivative' thinking.
I prefer 'clean sheet' thinking.
Yes. You might do a search on "Parallel Processing" here in the forum.
Good. 'Put the power to the pavement!'.
'"simulator in the editor" ... why do you need it?'
There are situations where it is useful.Working out sections of code.
'Yes. You might do a search on "Parallel Processing" here in the forum.'
Makes sense that somebody else has thought of it.
In terms of your car analogy the P2, or even the P1, does not look like all the other cars in the lot.
Or did you mean that if you looked at the P2 design then your own thinking would be derived from it? No doubt it would have an influence.
We look forward to seeing your own 'clean sheet' micro-controller design.
'You could use ViewPort'
Is there a free version of Viewport on the Parallax website?
Like Stamp plot which we have.
Okay. Get your drift about just use the hardware.
Back to the task at hand.
Yeah team! Get that Propeller 2 built!
Back to Simple Learn tutorials for me.
'We look forward to seeing your own 'clean sheet' micro-controller design.'
Already made my 'clean sheet' comments about Propeller 2.
However they build it is fine with me.
Okay. How about a dip version or at least solder the thing to a header board for us.
A Propeller 2 Flip breadboard friendly would do. $19.95 to forum members.
With all these Propeller boards I am not to keen about buying anymore.
"Already made my 'clean sheet' comments about Propeller 2"
What 'clean sheet' comments? Have you looked at the instruction set? What about the functionality of the smart I/O pins? What kinds of specialized I/O or execution are supported? These all speak to what markets the Propeller 2 would appeal to. Generalized suggestions without context are not very helpful. A simulator for just the processor of an embedded system type device without timing-accurate simulation of a wide range of external (and internal) peripherals is not useful in practice and, if done properly, is a huge undertaking.
The Propeller 1 is in fact used in some commercial products and is not just an educational tool.
They are all here in the post.
Sounds like the thing is decided so that's it for me.
Will test Viewport a little with 30 day version.
They do have different priced versions.
See if it is worth getting $29.95 one.
I do very much hope the P2 is actually decided. We have seen too many changes of direction and delays already.
I was wondering about that 'derivative' thinking thing. Well, that is what humans do. We are social animals and we learn from each other. If you were brought up in isolation you would not even have a language, you would probably not think anything much. Everything we think, say and do is derived from those that went before us.
I'm sure there will be breakout boards available to make it friendly for us.
Heater
'We have seen too many changes of direction and delays already.'
Been watching the Halloween Frankenstein movie.
Has the mob with torches.
Should we do that to move things along?
They are used to the protests at Berkeley.
Maybe we would be communicating like they do things up there.
Nah! Just do it the Southern California way.
Just keep busy doing something else why we wait for the big moment!
We could dress up like those IFTA protestors or whatever they are.
They have black outfits and sticks.
Instead we could have soldering irons and pocket protectors.
Have an old fashioned get together in front of Parallax.
Chant 'Propeller 2, Propeller 2'.