Shop OBEX P1 Docs P2 Docs Learn Events
Multiple propellers — Parallax Forums

Multiple propellers

pmartinpmartin Posts: 10
edited 2006-06-18 05:29 in Propeller 1
Anyone know if it would be possible to gang multiple propellers together, having them all communicate and work together at relatively high speed? Just curious if anyone who has more experience than myself has any ideas if that could work. I was thinking the I2C interface could be used; if not maybe the serial port, or something more exotic using the I/O pins.

Also, does anyone think that you could add an external scratch pad RAM to a propeller?

PM

Comments

  • PyrotomPyrotom Posts: 84
    edited 2006-06-15 15:01
    Take a look at Chip's "Propeller Loader" thread.....
  • Ryan ClarkeRyan Clarke Posts: 738
    edited 2006-06-15 15:59
    Pmartin,

    Are you heading in the direction of a Propeller cluster?

    Just curious.

    Ryan

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Ryan Clarke
    Parallax Tech Support

    RClarke@Parallax.com
  • Mike GreenMike Green Posts: 23,101
    edited 2006-06-15 16:14
    Adding scratchpad RAM to a propellor should be easy. Ramtron makes a line of EEPROMs in both I2C and SPI forms that are fast, have unlimited write capability and the write times are the same as read times. They also make their ferroelectric memories in standard parallel form (32K x 8), but that would take a lot of I/O pins unless you multiplex the address and data lines.
  • cgraceycgracey Posts: 14,133
    edited 2006-06-15 17:06
    You need to read the old PROPELLER SUPERCOMPUTING thread:

    http://forums.parallax.com/showthread.php?p=582511

    Yes, all this is possible. I think that within the next two months, we'll have an experimental 8x8 array (10,240 MIPS in 512 COGs·and 2MB of global RAM). I've been wanting to do this for a while, and after working on a bunch of video drivers, I think it would be fun to be able to dedicate a whole COG to just one scan line per frame.

    The PropellerLoader will get augmented a bit so that it could load the same program into a whole array of chips, instead of just one. Each Propeller would receive his own ID into BYTE[noparse][[/noparse]5] which is normally reserved for checksum during normal downloading. We'll hybridize the download process a bit by first loading an intercessor program which will then suck up and launch the final code with the ID insertion. We can bus them together however we want, so that some lines are common throughout, and others connect rows and columns.

    I really look forward to writing one short program and then seeing it run up to 512 COGs at once, keying off its ID to differentiate its data awareness or function.
    pmartin said...
    Anyone know if it would be possible to gang multiple propellers together, having them all communicate and work together at relatively high speed? Just curious if anyone who has more experience than myself has any ideas if that could work. I was thinking the I2C interface could be used; if not maybe the serial port, or something more exotic using the I/O pins.

    Also, does anyone think that you could add an external scratch pad RAM to a propeller?

    PM
    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔


    Chip Gracey
    Parallax, Inc.
  • Kaos KiddKaos Kidd Posts: 614
    edited 2006-06-15 18:13
    Chip...
    Oooooo... that just shot shivers up and down my spine...
    Its one HECK of a video card / driver...
    I'm very interested in how you would build the final outputs that would be feed the VGA/SVGA monitor.
    I would imagin the software would simply "wait its turn" to output it's line of data to the video device, or a dedicated "manager" to toggle the various inputs (video outputs from the others in the array) to the video device.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Just tossing my two bits worth into the bit bucket


    KK
    ·
  • bambinobambino Posts: 789
    edited 2006-06-16 18:42
    Witnessing 512 COGs would be interesting to see on a video screen and even more exciting to build, but would the video be possible with the ram Mike mentioned?
    I'll be lucky to talk my other half into one Propeller, not to mention sixty three more!
  • Kaos KiddKaos Kidd Posts: 614
    edited 2006-06-16 19:03
    LOL WITH bambino... I know that feeling... but still.... I think it would be awesom...
    Well, I do hope chip keeps us "little budgeted guys" in mind and makes the project scallable to say a 2X2 array... just to see how it works...

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Just tossing my two bits worth into the bit bucket


    KK
    ·
  • cgraceycgracey Posts: 14,133
    edited 2006-06-16 19:05
    bambino said...
    Witnessing 512 COGs would be interesting to see on a video screen and even more exciting to build, but would the video be possible with the ram Mike mentioned?

    The main memory would come from the global RAM in each Propeller chip. There'd be plenty. Well, 2MB in 64 chips. The COGs' RAM would total to another 1MB.
    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔


    Chip Gracey
    Parallax, Inc.
  • MuncherMuncher Posts: 38
    edited 2006-06-17 02:39
    And I have to settle for just one propeller. Why don't you try making a ray-(tracing or casting, I forget which one) program? You could recreate old Atari games.
  • Kaos KiddKaos Kidd Posts: 614
    edited 2006-06-18 03:56
    Funny you should metion ray-tracing... I'v been toying with raytracing and 3d modeling for quite some time as was trying to work out the math for the propeller... twood be very interesting...
    Welp, I just went off topic, but I'll add this and the drop the intrusion... The idea is real time 3d ray-tracing of the sourrounds, I'll expand on this once I get the sram, and some other thinggies working correctly... But the multi cog idea will make it work...

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Just tossing my two bits worth into the bit bucket


    KK
    ·
  • MuncherMuncher Posts: 38
    edited 2006-06-18 04:29
    i can't wait to see whatever you come up with.
  • RinksCustomsRinksCustoms Posts: 531
    edited 2006-06-18 05:29
    Grant it i'm no hacker or bill gates, but might it be easier to make an AGP adapter, to something like I have (nVida6800GT w/256MB of ram)? although the idea's presented herin do offer a possible alternate route for something powerful and small...



    I SAY MORE POWER TO YA'S!!! way outa my league...

    ~post edit~

    I retract this previous statement after a bit of enlightenment on the prop and an actual brainwave!
    An 8x8 array would afford the ability of a screen comprised of RGB leds's/current drivers (RGB's mounted on edge of PCB, current drivers mounted perpendicular, all SMT of course) for each color (this can be done very cheaply per pixel as you know) and have theorectically muliple colums @ once and all rows @ once. What would be the refresh rate of this super huge, super fast display?
    assuming 6 pins for programming/eeprom/inter-prop communication leaves 1,664 pins free for the display and up to another 256 pins if you omit/disconnect the two for inter-prop comm and two more for serin/serout and utilize the I2C buss for powerup loading/inter-prop comm for basic sync'ing off a central Xin clock durring runtime, having each prop assigned to generate a part of the screen and having a timed power supply to the 64 EEproms that power down after loading each prop with it's own "group" code so that data integrity is not comprimised leaving the i2c buss for inter-prop comm. I knw, i know, stupid polok! Chips are for nerds! lol just a thought anyway

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Definetly a E3 (Electronics Engineer Extrodinare!)
    "I laugh in the face of imposible,... not because i know it all, ... but because I don't know well enough!"

    Post Edited (RinksCustoms) : 8/14/2006 2:18:05 AM GMT
Sign In or Register to comment.