Shop OBEX P1 Docs P2 Docs Learn Events
Question for Parallax staff from potential industrial user. — Parallax Forums

Question for Parallax staff from potential industrial user.

MicksterMickster Posts: 2,694
edited 2013-05-20 08:44 in Propeller 1
Questions that are not being answered that will affect future decisions:
  1. Are we going to see some support for PropBASIC or are we stuck with GCC for a high level language when PASM-type performance is required?
  2. What can we expect from the P2 in terms of quadrature counter capabilities?
I want the Prop to be in my future but am starting to look at alternative solutions.


Regards,

Mickster

Comments

  • cgraceycgracey Posts: 14,155
    edited 2013-05-16 19:39
    The P2 will have a version of Spin that will probably be 20x faster than P1's Spin. The P2 also has two quadrature encoders (via the CTR's) per cog, which sample at the main clock frequency (160MHz). If you need a chip right away, don't wait for P2, but be sure to check back later when it's available.
  • jmgjmg Posts: 15,173
    edited 2013-05-16 20:06
    Mickster wrote: »
    ]Are we going to see some support for PropBASIC or are we stuck with GCC for a high level language when PASM-type performance is required?

    Anything that runs on Prop 1, and is popular, will port across quite quickly to Prop 2.

    If you need QuadCtrs, and also need a solution before Prop 2 is in production, then you could use a CPLD alongside a Prop 1.
    The Ctr docs for Prop 2 are close, so when those are released, you could easily design such an external quad solution to port easily to the Prop 2.
    Something like the Lattice MachXO2-256, should manage a number of quad ctrs.
  • KC_RobKC_Rob Posts: 465
    edited 2013-05-17 09:35
    Mickster wrote: »
    Questions that are not being answered that will affect future decisions:
    1. Are we going to see some support for PropBASIC or are we stuck with GCC for a high level language when PASM-type performance is required?
    2. What can we expect from the P2 in terms of quadrature counter capabilities?
    I want the Prop to be in my future but am starting to look at alternative solutions.


    Regards,

    Mickster
    Are you using P1 for anything now? If so, what limitations/drawbacks have you experienced with it in practical use? I completely understand wanting to stick with something you're familiar with. I wonder if, from a business standpoint, it would make sense to do an interim design based on P1 (like that suggested by jmg) to get by until P2 arrives? Much as anything I'm just thinking out loud here ... I suspect more than a few of us could be faced with similar decisions.
  • max72max72 Posts: 1,155
    edited 2013-05-18 06:01
    This silence about Propbasic is worrying...
    While it is not considered a "professional" language it combines speed and user friendlyness.
    An official support in simpleide would push the development and offer the users a terrific tool.
    I would like to know if propbasic will never be supported for the same reasons.
    This silence from parallax is strange, considering the usual openness offered us. Moreover we could accept any answer, probably...

    Massimo
  • RsadeikaRsadeika Posts: 3,837
    edited 2013-05-18 06:30
    This silence about Propbasic is worrying...
    I think at one time bean did have some sort of rudimentary IDE for working with Propbasic, maybe he should get that up and available for everyone. I also have the feeling that the amount of Propbasic users is very small, so why complicate SimpleIDE with Propbasic.

    My personal feeling is that SimpleIDE should be kept simple, it already addresses Spin, which a very few people are using, within SimpleIDE. The problem I have is, it looks like their is going to be a lot of different IDE's - Eclipse, SimpleIDE, New Proptool, and whatever else comes along. So, maybe somebody, not me, should take up the challenge and choose, maybe Eclipse, and do the pluglins for Propbasic, Spin, C, and whatever else comes to mind, that will work with PI and PII.

    Ray
  • MicksterMickster Posts: 2,694
    edited 2013-05-18 08:54
    KC_Rob wrote: »
    Are you using P1 for anything now? If so, what limitations/drawbacks have you experienced with it in practical use? I completely understand wanting to stick with something you're familiar with. I wonder if, from a business standpoint, it would make sense to do an interim design based on P1 (like that suggested by jmg) to get by until P2 arrives? Much as anything I'm just thinking out loud here ... I suspect more than a few of us could be faced with similar decisions.

    I have been experimenting with closed-loop servo-motor control on the P1.

    However, I need 4 axes, each equipped with 2 quadrature encoders as I use feedback from both the motor shaft and the driven load (aka: dual loop) at frequencies exceeding 1.6M quadrature counts/sec.
    I have already been looking at dedicated quad-decode/counter chips that I could hook to the Prop via SPI. I also need lots of DIO which I could also expand in a similar way.

    PropBASIC is my favorite programming language for the Prop as it is sooo easy to use and results in code that I can still understand even after not seeing it for a couple of weeks. :smile:

    Chip stated that the P2 Spin language could be 20X faster than on the P1. But isn't PASM 100X faster than Spin right now (I could be wrong here)? PropBASIC generates PASM code.

    So, the easiest to learn, most readable language available for the P1 is already 5X faster than Spin will be on the P2?

    OTOH, I am sure that the P2 Spin will be more than adequate for my purposes and quad-counters in hardware will be wonderful.

    Regards,

    Mickster
  • MJBMJB Posts: 1,235
    edited 2013-05-18 14:36
    Mickster wrote: »
    I have already been looking at dedicated quad-decode/counter chips that I could hook to the Prop via SPI. Mickster
    If you could give the type codes of those chips, I would be interested as well. thanks MJB
  • KC_RobKC_Rob Posts: 465
    edited 2013-05-18 17:39
    Rsadeika wrote: »
    I also have the feeling that the amount of Propbasic users is very small, so why complicate SimpleIDE with Propbasic.
    I'm not sure it's that small; the topic comes up often enough around here. And there is no question that folks are using Spin and are likely to continue doing so (esp. if performace really is much much better on P2).

    Now back to something max72 said: I tend to think that if people are using PropBASIC for "professional" work it can fairly be called a "professional" tool; and it would appear that Mickster and others are doing (or are planning to do) simply that.
  • jmgjmg Posts: 15,173
    edited 2013-05-18 17:57
    MJB wrote: »
    If you could give the type codes of those chips, I would be interested as well. thanks MJB

    Likely it is this :
    http://www.usdigital.com/products/interfaces/ics/lfls7366r-s
    Nice package, and tolerable price.

    I rechecked my Quad in CPLD notes, and these are the fit reports for a QFN32 MachXO2
    This code is untested, but is used to check what fits in CPLDs..
    // Design is 4 Quad counters (CW:0) and a single Muxed (CW:0) SPI Capture/Shifter 
// can vary the width, to fit into a 32 pin XO2-256 
//
// parameter  CW = 23;	// 23..0 x 4 Quad Ctrs + 23..0 SPI Capture/Shift
// in XO2-256   SLICE     97/128   75% used    Max :  LCMXO2-256ZE : 77.730MHz 

// parameter  CW = 27;	 
// in XO2-256   SLICE    111/128   86% used   MAX : LCMXO2-256HC   141.044MHz 
 
// parameter  CW = 29;	 
// in XO2-256  SLICE   118/128     92% used   MAX : LCMXO2-256HC   133.815MHz 

 parameter  CW = 31;	 
// in XO2-256-tqfp100  SLICE  125/128      97% used   MAX : LCMXO2-256HC  138.947MHz
// in QFN32            SLICE  125/128      97% used   MAX : LCMXO2-256HC 133.316MHz

MachXO2 and SW seem able to fit very close to 100%.
Speed does not look like being an issue.

    So you can easily fit 4 channels of 24 bit Quad + SPI read into a 32QFN, and (just) fit 4 channels of 32 bit
    To manage 8 channels, in one package, you bump to a MachXO2-640, which comes in 100tqfp.
  • MicksterMickster Posts: 2,694
    edited 2013-05-19 15:48
    jmg wrote: »
    Likely it is this :
    http://www.usdigital.com/products/interfaces/ics/lfls7366r-s
    Nice package, and tolerable price.

    That's the one!

    Cheers!

    Mickster
  • pedwardpedward Posts: 1,642
    edited 2013-05-20 00:38
    For comparison, PASM is about 40x faster than SPIN on the Prop 1. Clock for clock, SPIN on the Prop 2 would be 8x faster than the Prop 1. The stated performance of SPIN is 500Kops, or about 40x slower than PASM.

    I'm fairly certain that you can track quad counters at ~2.5Mhz in PASM on the P1, you might be able to track 4 at that speed, definitely 1. This is a per-COG figure. A nice way to do this would be to do 2 encoders and servo control in 1 COG, then run 4 COGs to do what you need.

    In all fairness, PropBASIC is a 3rd party language that was developed and maintained by a non-Parallax entity. Parallax has only shed light on languages they directly support and maintain.
  • BatangBatang Posts: 234
    edited 2013-05-20 06:37
    New thread started.
  • LoopyBytelooseLoopyByteloose Posts: 12,537
    edited 2013-05-20 08:44
    For development purposes, I really like Forth on the Propeller. Two versions, PropForth and Tachyon Forth have said they would develop Propeller 2 versions.

    Please consider these a useful alternative to PropBasic. Why so? Well, they are interactive, not the write-compile-load cycle that can be more tedious in the developmental phase where you might be checking for hardware compatibility and wiring problems.
Sign In or Register to comment.