Shop OBEX P1 Docs P2 Docs Learn Events
GCC, Prop II future of Propeller and SPIN — Parallax Forums

GCC, Prop II future of Propeller and SPIN

mumcs01mumcs01 Posts: 3
edited 2012-07-01 16:08 in Propeller 1
With the beta of Propeller GCC, and in the announcement of the beta of Propeller GCC the mention of 'all the development will be fed into Propeller II', I have some questions about the future that perhaps someone from Parallax can answer. I apologize in advance if this is answered somewhere else, but I couldn't find it.

- With the advent of GCC for the Propeller, will further development continue on SPIN for the propeller, and more importantly will Propeller 2 be a GCC / Assembler only chip. Is SPIN going to continue on to the next generation of chip? I know its been a somewhat controversial language depending on who you talked to. I actually like it.

- Will Propeller II EOL Propeller I? I've opted to start some new projects on Propeller instead of the the PIC, and it would be pretty frustrating to see all my work not be worth anything as the project has a 5 year life. (It happened to me on the SX, so I'm skidish).

- I'm not a C guy. I know I need to learn, but how is the whole object library going to work for Propeller with GCC? Can Spin objects work in the GCC environment, or will there need to be separate C objects that will only be for use with GCC, and SPIN objects that will only work in the SPIN environment.

Just curious. Thanks!
Mike.

Comments

  • Dave HeinDave Hein Posts: 6,347
    edited 2012-06-29 12:43
    Prop II will not contain the Spin interpreter in ROM. Spin programs will have to include the Spin interpreter as part of the binary file. Parallax has said that it will continue to support Spin on the Prop II. There has been some mention of enhancing Spin to take advantage of some of the features of Prop II.

    I think Prop I will be around for a long time. It will be lower in cost, consume less standby power and require only one power supply. There will still be a lot of applications where Prop I works just fine.

    There is a conversion program that converts Spin objects to C++. This can be used to link Spin code to C/C++. The drawback is that this requires more memory, but on the other hand it will run faster. Another approach is to run C code in one or more cogs while running Spin code in other cogs. They would communicate through mailboxes in hub RAM.
  • groggorygroggory Posts: 205
    edited 2012-06-29 12:44
    I'm not an expert, but here's what I've learned as I've been lurking in the prop2 status thread.

    1) SPIN is important and will continue to be fully supported.

    2) Prop1 and Prop2 target two totally different markets. Both will continue production into the foreseeable future.

    3) I'm not 100% on this one, but there are ways of getting spin into GCC, but it doesn't seem rock solid quite yet. I know they're working on it!
  • 4x5n4x5n Posts: 745
    edited 2012-06-29 13:01
    mumcs01 wrote: »
    With the beta of Propeller GCC, and in the announcement of the beta of Propeller GCC the mention of 'all the development will be fed into Propeller II', I have some questions about the future that perhaps someone from Parallax can answer. I apologize in advance if this is answered somewhere else, but I couldn't find it.

    - With the advent of GCC for the Propeller, will further development continue on SPIN for the propeller, and more importantly will Propeller 2 be a GCC / Assembler only chip. Is SPIN going to continue on to the next generation of chip? I know its been a somewhat controversial language depending on who you talked to. I actually like it.

    - Will Propeller II EOL Propeller I? I've opted to start some new projects on Propeller instead of the the PIC, and it would be pretty frustrating to see all my work not be worth anything as the project has a 5 year life. (It happened to me on the SX, so I'm skidish).

    - I'm not a C guy. I know I need to learn, but how is the whole object library going to work for Propeller with GCC? Can Spin objects work in the GCC environment, or will there need to be separate C objects that will only be for use with GCC, and SPIN objects that will only work in the SPIN environment.

    Just curious. Thanks!
    Mike.

    As others have already mentioned there currently aren't plans to drop spin or pasm from the prop2.

    As far as "having" to learn C don't sweat it. I've "had" to learn a lot of different programming languages in my life and have come to realize that in the end programming is programming and while some languages are suited to particular tasks the skills needed to be good at programming transfers quickly and easily between languages. If you can program in spin you'll do fine with C and C++!!
  • jmgjmg Posts: 15,183
    edited 2012-06-29 13:09
    mumcs01 wrote: »
    - With the advent of GCC for the Propeller, will further development continue on SPIN for the propeller, and more importantly will Propeller 2 be a GCC / Assembler only chip. Is SPIN going to continue on to the next generation of chip?

    It is only software, why would it go away ?
    There are many more reasons for it to continue.
    Prop 1 Spin is in Rom, so that makes it a locked standard, but prop 2 will allow Spin 2....
    - Will Propeller II EOL Propeller I?

    Is is the same package, price and power profile ? Then the answer is obviously no.
    Any future EOL on a Prop 1 will more likely be FAB driven - when no one can build it anymore...
  • Invent-O-DocInvent-O-Doc Posts: 768
    edited 2012-06-29 13:44
    Fair questions.....

    I don't see any reason for the propeller II to supplant the propeller 1 in the things it does well. Although the prop II has new capabilities, it also draws more power and is more complicated. The prop II is likely to create entirely new applications - some will be more computer-like due to graphics capability and more memory, but some will be more electrical engineering focused with unknown applications because there has never been a microcontroller with 90+ pins that has A/D, D/A, video scaler and multiple oscillators before - so who knows.

    Prop II will use an enhanced spin also because it was designed to take advantage of the architecture of an 8-cog parallel computer. I was really a proponent of having a C, and now that I see what the code looks like, I see advantages but somehow SPIN seems simpler....

    Finally, I think that the real gap in the parallax portfolio is really on the low end - something that the SX and STAMP fulfilled in the past and is being filled in by the arduino. They could probably use a lower pin count prop 1.5.
  • PublisonPublison Posts: 12,366
    edited 2012-06-29 15:27
    mumcs01 wrote: »
    With the beta of Propeller GCC, and in the announcement of the beta of Propeller GCC the mention of 'all the development will be fed into Propeller II', I have some questions about the future that perhaps someone from Parallax can answer. I apologize in advance if this is answered somewhere else, but I couldn't find it.

    Just curious. Thanks!
    Mike.

    Mike welcome to the forums.

    I'm really surprised that many people, like you, do not direct these questions to http://www.parallaxsemiconductor.com/

    Parallax does not always surf these forums, but a question to one of the Application Engineers at the above link, will get you an answer in a much more timely manner.

    That said, many of use can try a direct you to recent activity, but it tends to get buried in the forum threads just because of all the activity that goes on here. Hang in there. In the mean time, try both means of getting an answer.

    Jim
  • mumcs01mumcs01 Posts: 3
    edited 2012-06-29 19:38
    Thanks all! Somehow, if nothing more I feel reassured. :smile: I still can't decide if i Want to dive now into Gcc or later, but I'm certainly going to use the prop for this project. (again, I love spin)

    Oh, and invent-o-doc, I totally agree with you. I would love a scaled down prop for small embedded use. Thats why I loved the sx so much. I switched over to pic when sx discontinuation was announced, but I missed the parallax support and comprehensive learning and teaching tools.

    Anyway. Thanks again!
  • smbakersmbaker Posts: 164
    edited 2012-06-30 00:07
    The biggest problem that I see with C is that it's not as good of a learning language as spin (or python, which is quite spin-like). It's easier to hang yourself in some non-obvious way with C.

    As someone who has done more C programming in his life than any other language (followed closely by pascal and python), I really do find programming the prop in spin to be refreshing. I hope spin continues to be actively supported and extended for the Prop II.

    I also join in the calls for a prop 1,5, perhaps somewhere in the range of an 18 to 28 pin footprint. Maybe 4 or 6 cores. Lower power, smaller pin count, and easier to embed in smaller projects/products. It would complete the line.
  • JonnyMacJonnyMac Posts: 9,194
    edited 2012-06-30 08:41
    I also join in the calls for a prop 1,5, perhaps somewhere in the range of an 18 to 28 pin footprint. Maybe 4 or 6 cores. Lower power, smaller pin count, and easier to embed in smaller projects/products. It would complete the line.

    I would love that, too.
  • Mike GreenMike Green Posts: 23,101
    edited 2012-06-30 09:14
    "If wishes were ..."

    It's not like you could take the Prop I design and just cut it in half. Chip design doesn't work that way. There have been very very long threads posted that have hashed over this sort of thing and you're talking about mostly starting over from scratch, maybe not all the way since you can salvage most of the building blocks, but a huge amount of the work is in the interconnections and that would have to be redone. The Prop I is a large chip and, although cutting the number of cogs would make it smaller in terms of total area, the dimensions required might still cause problems in a smaller package (the chip might be long and thin). In any case, you're talking about a couple of years of work.

    Some interesting questions that have been discussed before include: 1) How many cogs would be 'enough' for a small Prop? Would that reduction be enough to reduce the chip area to where there's an advantage to making a smaller chip? 2) Would reducing the amount of SRAM help the chip area enough? 3) Would reducing any of the features of the Prop I cripple the chip enough that it couldn't compete at all with the PICs and AVRs and ATMegas of the world, particularly given that they will continue to evolve while a new Prop design is underway?
  • JonnyMacJonnyMac Posts: 9,194
    edited 2012-06-30 09:42
    you're talking about a couple of years of work.

    That's being very optimistic considering the small size of the design team. Still, I think it's appropriate to remind our friends at Parallax that a smaller Propeller has a market and should resources become available it would be worth pursuing.
  • potatoheadpotatohead Posts: 10,261
    edited 2012-06-30 10:12
    I agree, and perhaps P2 will generate revenue that could be used to explore some niche scenarios. Doing that would compete with building P3. When the time comes, I suspect that's going to be a very interesting discussion.

    Re: Future of SPIN. Let's hope SPIN gets a few extensions to plug some holes we know about now. I love SPIN. It might not be the most expressive language right now, but it sure is fun.
  • pik33pik33 Posts: 2,397
    edited 2012-06-30 11:47
    P1.5?

    It should be simply P2 chip in DIP40/44pin SMD
  • Mike GreenMike Green Posts: 23,101
    edited 2012-06-30 12:06
    @pik33, The P2 cannot be put into a DIP40 or 44pin SMD package. There's not enough room for the chip in these packages.
  • Dave HeinDave Hein Posts: 6,347
    edited 2012-06-30 13:11
    potatohead wrote: »
    Re: Future of SPIN. Let's hope SPIN gets a few extensions to plug some holes we know about now. I love SPIN. It might not be the most expressive language right now, but it sure is fun.
    I suspect there will be a number of SPIN variants for P2 since the interpreter will not be hardcoded. P2 will have the same cog RAM limitation as P1, so I don't think we'll be able to cram much more in it than P1 already has. It's too bad that P2 doesn't have more cog RAM. Even an extra 1KB per cog would have helped tremendously.
  • Heater.Heater. Posts: 21,230
    edited 2012-06-30 13:39
    smbaker,
    "spin is like python"
    OK not a direct quote but really, yes they both use white space for block indenting, after that there are very few similarities.
  • potatoheadpotatohead Posts: 10,261
    edited 2012-06-30 13:39
    On P2 LMM will perform much better, and I suspect that will do a lot to marginalize the COG size limit.

    I hope we get in line LMM PASM as one extension.
  • cavelambcavelamb Posts: 720
    edited 2012-06-30 19:59
    On the small end, I could easily see a 4 cylinder being viable.
    A smaller cheaper chip for cost sensitive projects.

    Dunno about a two-hole, though.
    That's kind of a curious creature, isn't it?

    Two cogs and a hub?

    A high-performance bottom-end jelly-bean controller?
  • Dave HeinDave Hein Posts: 6,347
    edited 2012-06-30 20:09
    smbaker wrote: »
    The biggest problem that I see with C is that it's not as good of a learning language as spin (or python, which is quite spin-like). It's easier to hang yourself in some non-obvious way with C.
    I have a bias toward C, but I really think it is a better learning language than Spin. The main reason is that the C compiler will warn the programmer about common programming errors. Programmers will be told when they are using an int as a pointer. This is something that the Spin compiler cannot do because there are no type declarations.

    A subset of C can be taught to beginners that includes most of the features of Spin. This would shield the beginner from having to worry about structs, function pointers and other features that may be confusing to a new programmer.
  • simonlsimonl Posts: 866
    edited 2012-07-01 15:18
    I keep trying to get going with C, but each time it just feels over complicated after using Spin.
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 23,514
    edited 2012-07-01 15:25
    Simon,

    Once you're comfortable with Spin and PASM, I see no technical advantages from moving to C. The main advantage it has is bringing current C users into the Propeller fold, not bringing current Propeller users into the C fold.

    -Phil
  • ratronicratronic Posts: 1,451
    edited 2012-07-01 15:55
    After reading this thread I still haven't figured out is Parallax going to provide a Spin interpreter for the Prop2 when it is released?
  • LeonLeon Posts: 7,620
    edited 2012-07-01 16:03
  • ratronicratronic Posts: 1,451
    edited 2012-07-01 16:08
    Whew! Thanks Leon not that I can't do C but that is a relief.
Sign In or Register to comment.