HOw long should it take the average person to learn ASM? We are talking just enough to add or modfy a table. I need to get an idea on how many hours a day to spend on it.....

 

It's really impossible to say. People's aptitudes are all over the board for stuff like that. I have years in trying to make sense of it, and still haven't much of a clue.

 

Makes me wonder how much time RBob has in.....

 

Just a week or two

RBob.

 

I think to get a basic understanding does not taketoo long.
The problems come from the abstract uses of the code for faster operation. Things that drive you nuts are the usage of the "status" bits to check if values are higher/lower/equal then indexed addressing makes you say "what was that number again"
Taking a line by line "execution" approach in one of the hacs helps to see what each of the opcodes does and gives you an indication of how long this may take.
Looking at how the code is executing and understanding what is going on is alot easier than reformatting to fit something else in.
That's where the mastery come in. (that will take at least the two weeks Rbob has put in )
My two weeks have taken about 6 months on and off and I have a grasp but by no means a mastery of it yet.
JP

 

I think the key to learning and understanding assembly code is to first know how a uProcessor works. Understanding how the registors are used, how memory is used, what addressing is, along with the accumulator goes a long way toward writing code.

Section 6 (Central Processor Unit) of the pink book describes a bunch of this. This would be the best place for a beginner to start out. Then read some simple GM code and try to understand it. There isn't a need to understand the code of a 3D lookup. It is enough to understand the basic operation, it isn't simple code.

Start where the SA lookups are first done, and move down through the code. For the most part the SA stuff is straight forward. May even want to start with a C3 mask, such as $61.

When it comes time to modifying code, try some easy stuff first. Once that is working move on to more items. Take a building approach to a modification. Make it simple at first, once that works, move the 'simple' up a notch and add to it.

RBob.

 

If you're interested in a good book to start, I highly recommend 'Bebop to the Boolean Boogie'.

Should be the second link:

http://www.amazon.com/exec/obidos/se...222422-8015159

Lame name I know, but this guy covers tons of material in an easy to read way.

 

Pink book.....sect 6....I'll read it. I think you are right, I need to step back and learn how everything works. I seen a book that decribed memory, how it's divided into sections and each section has an address range...or something like that(only read it for a second). Didn't buy the book because I was not sure the info could be directly applied to our ECM. Didn't want to gamble $56.

I'll buy Bebop to the Boolean Boogie within the next two weeks. Thanks!

-Tim

 

Another thing that might help (at least it does for me) is to start with something you want to "do" to the code so you can apply your "research". Your mention of adding a lookup table is an excellent place to start! As Rbob mentioned, you can start simple (i.e. copy an existing table) and make changes to see how things work from from there (i.e. change input variable, change table size, change breakpoints, etc.) If you start with the array/table routines, you should be exposed to a good portion of binary math and parameter scaling.

If you're interested in getting started "for free" here's what I'd do (if you haven't already):

1) Download the HC11 Pink book. You'll want this available for reference all the time!

2) Obtain a copy of some existing source code (i.e. $60) and a compiler. Without making any code changes, make sure you're proficient at getting the code to compile and produce a binary that exactly matches the "stock" code. Once you can do this, you should be in a good position to compile the "modified" code you'll be making later.

3) Obtain one of the free HC11 simulators. This will let you view/manipulate registers and memory as you step through portions of the code. Although you won't be able to run the entire binary from start to finish, you will be able to tell the simulator to start and step through the code at a particular area you're interested in manipulating (i.e. the start of one of the array/table lookup routines).

4) Identify a portion of the code that has an array lookup. I'd start with an array because it's a little simpler and is used as part of the larger "table" lookups. In order to help you understand the steps the algorithm performs, write down (or use your favorite programming language) a method of how you'd do an array lookup. This might help you get a feel for the big picture.
Simulate/step through the ECM logic with (3) above and your algorithm close at hand while trying various input values. (Here's where keeping the Pink Book around pays off big time!) If you're using $60, the PE AFR vs. RPM array would be a good candidate. Verify you get what is expected from the table for any RPM value you try. Look at how the code executes when you enter an input value that precisely falls on a breakpoint vs. when the value is between breakpoints. What happens when you enter 0? What happens when you enter 6375?

There are further variations on this theme, such as investigating array lookups that have different max/min input values, different breakpoints and different sizes. Ultimately, you can also tear into the "table" lookups, too.

I don't want to discourage you from picking up some books and doing further research. I just find I learn better if I can physically "see" results while doing this research! And, having a specific goal sometimes keeps me from jumping off on tangents. (Although, following those tangents at a later date is very often satisfying!)