Assembly Help/Challenge <Basic to ASM Conversion>

[Trogdorgar]

Hi there guys, I have been using basic for a long while now, and want to start getting into assembly.
I have read some pretty good tutorials, especially the "Learn Assembly in 28 Days" Tutorial by Sean McLaughlin
It was an excellent read, and a good start for anyone wanting to learn the basic of assembly and beyond :biggrin:

Want I am wanting to do to get a grand understanding of Assembly is to have one of my programs, which is my Sudoku Solver, converted from TI-BASIC into assembly,
so that i can understand the basics of assembly more fluently.

I will attach the program to this post and hope that someone can convert it for me,
because there is no chance that I can without screwing it up. <insert completely stupid joke here> haha....

So if theres anyone out there looking for a good challenge for assembly or would like to help me out, that would be great, the program itself is pretty much as completely optimized
as i can get it.
The basic overview of the solver is in the mapping file.
The solver was made for use with the TI-84+
The solver programs are all in group names SKDUSLVR

and before I forget, the "probability of choice" grid is written in basically a reverse binary format
a.k.a 0.111111111 = 0.(Prob of 1)(P of 2)(P of 3) etc...
the conversion method I used in the programming to get the probability integer of a number from the format is this code
int(fpart(I_10^(N-1))10)

As I complete more methods of solving for the solver, I'll try to append them to the post.


Updates:

[Galandros]

bah, I need to install GFM or use the TI-Connect in my laptop. Today I don't have time for this. Exam tomorrow...

But BASIC and ASM are in so different levels that the ported code can be unexpected to you.
And in ASM, because of the speed, you can have the chance to use simply use a pure brute force algorithm. But I will try to make an equivalent code.


ASM in 28 days is a great tutorial. But you can also try later Cobb's tutorial or some TI documentation about b_calls and handling TI-OS variables and such. And practising is really needed!


Another programming enthusiast, just like me! Welcome to UTI forums.

[Trogdorgar]

Yeah right now I use both TI-Connect and TI-Graph Link, TIGL to create and edit the BASIC programs and TI-Connect to send them, since GL doesn't support the usb port sending for the 84+

Yeah and the reason I would like the program to be in logical format, instead of brute force is because of a few things.
1. I know the exact coding of a logical solver, since they are based on equalities instead of guess-and-check like in brute forcers.
2. The logical solver is easier to code and maintain, unlike the brute forcer, which if something goes wrong, you have to go through hours of searching the code to figure out what screwed up.
3. If I just have the program I have converted as it is, I won't have to read thousands of tutorials trying to figure out what each little command does, and I can use code relation to get a fair basis of what a patch of code does.
4. I would come to believe that a logical solver would be faster, due to the fact that brute forcing has to go through every single number, and if there is an error, has to backtrack until the error is fixed, where logical solvers don't.
5. I don't exactly get how a brute forcer works, so I would have no idea what the code is doing until I become fluent in assembly, where if i found a brute forcer in BASIC, I would basically know what is happening without even trying, BASIC is like a second language to me :P

In relation to the tutorials, I will check out Cobb's tutorials to see what they are all about, since right now I am basically on a "language learning" binge, and trying to learn multiple languages at once lol, right now im learning C / C++, ASM, HTML, PHP, and trying a little bit of Java.

Yay for enthusiasts lol, been doing BASIC for the past 2 years now, I can't get enough of it, started doing it on my old CASIO FX-9750G, but then took it apart and the keypad stopped working so I got a TI-84+ and kept doing BASIC ever since, and recently saw the stuff people were making in Assembly on ticalc, and wanted to get in on the action, so far my fav asm game is either ZTetris or Degeneration.

[magicdanw]

Please don't learn assembly by trying to compare it to basic. It won't work. Assembly is a completely different beast, and must be approached as such. Read 28 days, and all will be explained.

[Trogdorgar]

[Galandros]

Multiple languages at once is definitely bad idea. I just know TI-BASIC and z80 ASM. Great distance between both but once I master (somewhat) ASM I want try another one :biggrin:

And comparing both is not good idea... I though you wanted the code as curiosity or see some code in action.
Anyway, GFM is giving me too much trouble. At night I will use my other computer...



just a little code to see the point. (And action equivalent to BASIC: [A](R,C) )
;Do you remember using data tutorials in ASMin28days?
retrieveabyte:
;input: b=row c=column
;output: a in the position wanted in the matrix
ld hl,Matrixdata
dec b ;just to give the BASIC feel of [A](1,1
dec c ;very dumb ASM because you just assume 0 as the first byte
ld a,b
add a,a
add a,a
add a,a
add a,b ;a=b*9
add a,c
ld b,0
ld c,a
add hl,bc
ld a,(hl)
ret

Matrixdata: ;9x9 matrix
.db 0,0,0,0,0,0,0,0,0
.db 0,0,0,0,0,0,0,0,0
.db 0,0,0,0,0,0,0,0,0
.db 0,0,0,0,0,0,0,0,0
.db 0,0,0,0,0,0,0,0,0
.db 0,0,0,0,0,0,0,0,0
.db 0,0,0,0,0,0,0,0,0
.db 0,0,0,0,0,0,0,0,0
.db 0,0,0,0,0,0,0,0,0

The following example can explain why doing kind of BASIC in ASM is bad. ASM is not optimized if you use call's to the retrieveabyte routine every time you need a number in a certain position the matrix. You can just have hl always pointing to the matrix and retrieve every time you want. No need to do what the second code will do.
For example search for a value of 0 in the matrix and solve it with a subroutine, just:

ld b,255 ;loop 255 times is sufficient to pass the matrix many times...
ld hl,Matrixdata
loop:
ld a,(hl)
or a
call z,solve ;this calls a routine to find out which value can be given and we assume it preserves hl
inc hl
;I am lazy enough to not write a code that will restore hl to the start of matrix
;just assume it is done here:
<code that load hl with the start of matrix when hl has passed through it>
djnz loop

;ASM guys don't look if you don't need.
;I don't know how I was capable of this...
No need to:
ld e,255
ld b,1
ld c,1
loop:
call retrieveabyte
or a
call z,solve
inc c
ld a,c
cp 9
jr nz,loop
ld c,1
inc b
ld e,a
or a
jp nz,loop

This code lacks a check for if the sudoku is already solved after some passes... I hope there is no error.
So using the BASIC method in ASM is wasting ASM speed, space and your time. (note: I wasted time )


You may want to know the solve routine itself. Later I will see your basic code and see what I will do...

[Trogdorgar]

[magicdanw]

[simplethinker]

When you're just loading a constant in, it's always best to just do "ld A, 9" (or whatever). One of the key differences between thinking in Asm and BASIC is that with BASIC you want to do as much as possible with one command and stuff like multiplication and boolean expressions are more efficient. With Asm, you can only do one thing at a time and trying to combine many things and do them once is often more cumbersome and complicated than doing them one at a time. Multiplication is considerably slower than "and" in Asm, so those "optimizations" in BASIC considerably slow Asm down.

You really have to change your mindset when switching (doing that is what finally made Asm click in my head), so instead of converting your BASIC program into Asm, start from scratch and try and formulate how your algorithm could be implemented in Asm (remember, at the highest level you're still only dealing with data 8 bits at a time ). I think Iambian put it perfectly in his tutorials

[FloppusMaximus]

The things that are easy to do in assembly are rather different from the things that are easy to do in BASIC. Start out with something simple and highly interactive, like a game. Nothing that requires complicated algorithms.

I must disagree that Z80 assembly is the "easiest", but I agree with the spirit of that statement

[Trogdorgar]

Alright, thanks, I'll check out more tutorials and try out Cobb's tutorial, since apparently this has a good documentation about b_calls, since i have already red "Learn assembly in 28 days" and didn't learn much about b_calls from it.

You guys just continue to work on the Assembly sudoku solver, since i know that will be a good challenge for you, and would be a cool UTI project, like a assembly version of the scanraid.com sudoku solver, for ppl that dont have laptops :lol:

I will make sure to add all the names of the people who helped me put this together into the Credits whenever the program is finished, or as it is being worked on.

[cjgone]

Heh. Assembly is definitely very different from bASIC. You probably want to get the best feel of asm through smaller programs and then progress to this sudoku solver. ;o

[Trogdorgar]

Yeah, thats what i'm planning on, just gonna start slow, so i don't annihalate my ti-84 lol.

Sorry for the delay in reply, having computer troubles, a virus took out my main computer, and now somethin went wrong with our backup PC and now it wont even start, so i'm replying from the laptop... kinda sucks lol.

[Galandros]

I haven't patient lately to put your program on my calc and see how it works deeply. (and also I have school work...)

I could make an solver like the attached. I did it and I know very well how it works. Interested?