localroger

  1. Windmill: Resources

    In this post: How I am using memory overall, and how I'm actually creating code.

    I am deep into the coding of the Windmill cross-compiler, which I will use to compile the real self-sufficient Windmill compiler written in Windmill. As usual coding clarifies a lot, and I made some adjustments to what I thought was final language syntax to make the compiler a lot simpler.

    The astute reader might wonder why I am bothering to write a Forth-like language that, as it turns ...
    Categories
    Uncategorized
  2. Windmill: What Happened?

    If you were following my blog in December, you'll recall I was posting about a new native Propeller development system I was working on called Windmill. In this post, why I haven't posted about it for over three months...

    At work, Windmill is a long term project, and in late December a couple of very hot short term projects came up. One of those required a VGA driver with special capabilities, which got me deep into figuring out the mysteries of the video driver. And video driver ...
    Categories
    Uncategorized
  3. Windmill: Booting and Bootstrapping

    In this post: What the title says

    I spent last week manually testing the interpreter by hand assembling byte codes into DAT statements and launching the interpreter cog against them in Hub RAM. That of course gets old very fast, and I still have a couple of tokens and one major service cog behavior that aren't tested because it would be such a pain to set up. It's time to get the system started as a system, so it can be used to implement its own diagnostics.

    And ...
    Categories
    Uncategorized
  4. Windmill: The Service Cog

    In this post: How I decided to deliver EEPROM bytes and serial communications to the large code Windmill system -- with one cog.

    In my original vision for Windmill, the EEPROM drivers were going to be part of the interpreter, so Windmill would have the same cog overhead as Spin. However, as I started writing test code I realized a few things. First, if I wanted to support other memory systems such as FlexMem or one of the multiple-prop solutions, it would be a lot easier if the ...
    Categories
    Uncategorized
  5. Windmill: Byte Code Interpreter

    In this post: A very clever way if I do say so myself of cramming as much functionality as possible into a cog.

    Having decided to implement a Forth-like stack machine byte code system for running big programs out of EEPROM, I then had to figure out how to get a one-cog interpreter to do as much as possible. I was intent on getting the most common and loopiest functions that I'd be using a lot tucked into the interpreter and coded in PASM, to mitigate the relatively slow speed of ...
    Categories
    Uncategorized
Page 2 of 3 FirstFirst 123 LastLast