Shop OBEX P1 Docs P2 Docs Learn Events
Propeller GCC Status Update - Page 2 — Parallax Forums

Propeller GCC Status Update

2»

Comments

  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 23,514
    edited 2011-09-26 13:52
    FWIW, an interesting take on software project management (from "someone" who ought to know):

    -Phil
  • jazzedjazzed Posts: 11,803
    edited 2011-09-26 21:24
    There has not been a significant breakthrough with Propeller GCC this week. Just a few notes ....

    Performance has improved (yet again ...) for heater's fft_bench in LMM and XMM modes.

    I was able to demonstrate 7 different ways to toggle LEDs in Rocklin with Propeller GCC using COGs programmed 5 different ways and 2 ways using LMM. There are other ways in progress .... :)
  • FredBlaisFredBlais Posts: 379
    edited 2011-09-27 06:49
    nice to hear some updates
  • jazzedjazzed Posts: 11,803
    edited 2011-09-27 08:26
    FredBlais wrote: »
    nice to hear some updates
    Just in case you missed them, most updates are posted in Jazzed Projects Blog.
  • Heater.Heater. Posts: 21,230
    edited 2011-09-27 11:37
    I'm beginning to wish I'd never put up that fft_bench. Looks like soon the GCC compiled C version is going to be running faster than my hand crafted PASM version. At which point I might have to give up and retire from programming all together.
  • jazzedjazzed Posts: 11,803
    edited 2011-09-27 16:35
    Heater. wrote: »
    I'm beginning to wish I'd never put up that fft_bench. Looks like soon the GCC compiled C version is going to be running faster than my hand crafted PASM version. At which point I might have to give up and retire from programming all together.
    If it makes you feel better, your PASM code will probably always be at least 2ms faster than LMM and 3ms faster than XMM.
  • Luis DigitalLuis Digital Posts: 371
    edited 2011-10-04 06:06
    Hello jazzed,

    What's new this week? Will we have a public release this month?

    Thank you.
  • jazzedjazzed Posts: 11,803
    edited 2011-10-04 07:53
    We do not have anything significant to report for this week.
    I hope we can preview at the end of this month. We'll see ....

    A preview release depends on these factors:
    1. A more complete set of GCC features - I.E. finish GDB.
    2. The GCC test suite should pass - this work is in progress.
    3. Continue testing with drivers, applications, and frameworks.
    4. Preview release package form needs to be decided.
  • Luis DigitalLuis Digital Posts: 371
    edited 2011-10-10 05:55
    Fingers crossed this week. lol
  • jazzedjazzed Posts: 11,803
    edited 2011-10-10 19:45
    Luis Digital wrote: »
    Fingers crossed this week. lol

    We still hope to preview at the end of this month ... or earlier.
    1. GDB and one other feature has been deferred until after preview.
    2. The GCC test suite passes more than 99% of tests - we are almost there.
    3. Continue testing with drivers, applications, and frameworks.
    4. Preview release packages for Linux and Windows have been tested. MAC OSX is next.
    Item 3 is the main issue right now. We want to deliver a quality preview.
  • __red____red__ Posts: 470
    edited 2011-10-13 06:42
    How do you see gdb working?
  • jazzedjazzed Posts: 11,803
    edited 2011-10-13 08:52
    __red__ wrote: »
    How do you see gdb working?
    Hi __red__ As I understand it:
    • Propeller will have a special debug interpreter in COG that communicates with the host computer.
    • Certain functions like fast-cache may be disabled in the Propeller debug interpreter.
    • The host will run a server that translates the GDB command set for the Propeller debug stub.
    • The GDB client will talk to Propeller via the host server.
Sign In or Register to comment.