Shop OBEX P1 Docs P2 Docs Learn Events
Interesting article on use of parallel processing — Parallax Forums

Interesting article on use of parallel processing

pgbpsupgbpsu Posts: 460
edited 2011-06-01 15:48 in Propeller 1
The following link is to a recent paper about how parallel/multi-core processing simplifies the unification of high-speed, deterministic, low-level tasks with time-critical but variable latency high-level tasks. A warning for those who have sensitive eyes, it uses and refers to that "other" chip. Nonetheless, I think it offers a valuable look at how they chose this platform and they've put into a scientific article what many on the forums already know from experience- working without interrupts can make things much easier.


http://dms.sagepub.com/content/early/2011/01/19/1548512910388197.full.pdf+html

One might ask why they didn't chose a propeller for this project. They clearly know about Parallax since they used a Ping sensor.

Furthermore, one might ask, given the author's explanation about what makes this particular platform such a good choice, how Parallax can address those issues with Prop II (or I) via Parallax Semiconductor. These authors didn't seem put off by the lack of hardware SPI, UART, etc. But they do highlight the fact that it's programmed in C.

If the moderators don't like this mention of other vendors please delete this thread. No harm intended. I just thought some folks here would find the concept interesting and apropos.

Peter

Comments

  • Heater.Heater. Posts: 21,230
    edited 2011-06-01 13:49
    From that link:
    You may access this article (from the computer you are currently using) for 1 day for US$32.00.

    Are they out of their freakin minds?
  • LeonLeon Posts: 7,620
    edited 2011-06-01 13:55
    They used XMOS devices, presumably because they offered: very high performance, low cost, lots of I/O, XC for programming parallel and multi-core applications, on-chip debugging, fast inter-thread and inter-core communication channels, and deterministic operation.
  • pgbpsupgbpsu Posts: 460
    edited 2011-06-01 14:20
    Heater- Sorry about the broken link. I'm at a University and our library must subscribe to that journal because I was able to get to it without shelling out $32. I can tell you it's not worth the money. Leon is mostly right although from my evaluation they weren't using it for its "very high performance" or "low cost". But the other issues were all mentioned.
  • Heater.Heater. Posts: 21,230
    edited 2011-06-01 14:29
    32 dollars to have someone explain to me that if I have multiple cores or hardware scheduled threads then I can use some for "high-speed, deterministic, low-level tasks" and others for "time-critical but variable latency high-level tasks" does seem rather over the top. I can explain it to anyone here in a single paragraph for free.
  • pgbpsupgbpsu Posts: 460
    edited 2011-06-01 14:45
    @Heater- I agree entirely. I thought it was interesting to see an academic article with industrial overtones lay out the reasons for using a parallel processor. More interesting yet is that they didn't chose a Prop. I considered it part of the ongoing discussion of how is the Prop perceived by people outside this forum, whether or not, by nature parallel processors are a niche market, etc. After considering it however, I think we've had plenty of those discussions here and I don't want to spawn an new prop vs "other vendor's parallel offering" thread. That will no doubt pop up again soon without my encouragement.

    p
  • LeonLeon Posts: 7,620
    edited 2011-06-01 14:49
    The Propeller is deficient in many of the areas I mentioned. That is why they didn't choose it. Also, XMOS gives free hardware worth $250 to any academic that requests it; a friend of mine who runs the embedded systems lab at ESIEE near Paris got several boards under that scheme which are being made available for student projects.

    Here is some more information on the project, including a YouTube clip:

    https://www.xcore.com/forum/viewtopic.php?f=27&t=983&p=6862&hilit=unmanned#p6862
  • Sal AmmoniacSal Ammoniac Posts: 213
    edited 2011-06-01 14:53
    pgbpsu wrote: »
    how is the Prop perceived by people outside this forum

    I was very surprised that Parallax Semiconductor did not exhibit at last month's Embedded Systems Conference in San Jose. This would have been the perfect opportunity to demonstrate the capabilities of the Propeller to thousands of engineers in their target market.
  • Heater.Heater. Posts: 21,230
    edited 2011-06-01 14:59
    Well, it's good to see academics looking at such things. Then the students will get exposure to those devices and then awareness of the benefits of multi-core and event driven processing will spread. And that will grow the market for Parallax as well. It's all good. The Prop II will find itself in a good position.
  • Kevin WoodKevin Wood Posts: 1,266
    edited 2011-06-01 15:48
    I don't think you should ask why they didn't choose the Propeller. I think you should ask why those that do choose the Propeller aren't writing similar articles.
Sign In or Register to comment.