return to ChibiAkumas.com Choose a Color Scheme: Dark
Print Mode

Learn Multi platform Z80 Assembly Programming... With Vampires!

Table of Contents
Introduction
The Z80
Numbers in assembly

Beginners Series - lets learn the basic Z80 commands by example!
    Lesson 1 - Getting started with winape
Lesson 2 - Memory copy, Symbol definitions, Loops and Conditional Jumps!
Lesson 3 - 'Case Statement' , 8 bit basic Maths, Writing to Ram and Reading from basic
Lesson 4 - Stack, Strings,Compiler Directives, Indirect registers, CPC Call
Lesson 5 - Bit level operations, Self modifying code
Lesson 6 -  Lookup table, Screen Co-ordinates, Vector Tables, Basic Parameters Byref
Lesson 7 - DI EI, RST x, Custom Interrupts, IM1/IM2, HALT, OTI / OTIR, HALT
Lesson 8 - Unwrapped Loops, Stack Misuse for speed & rarer Z80 commands

Advanced Series
- Lets learn some more useful Z80 examples that may help you in your programming
    Lesson A1 - Binary Coded Decimal
Lesson A2 - Interrupt Mode 2

Multiplatform Series - Using code provided that will talk to the hardware, lets write programs that works instantly on multiple systems!
    Lesson M1 - Z80 Monitor/Debugger
    Lesson M2 - String Reading and Memory Dumping!
Lesson M3 - String Matching for command reading
Lesson M4 - A Basic Text Adventure
Lesson M5 - Arkosplayer for Music and SFX!
Lesson M6 - Advanced Interrupt handler template
Lesson M7 - Multiplaform Font and Bitmap Conversion
Lesson M8 - Keyboard processing and redefinable game control input

Platform Specific Series - Lets learn how the hardware of the systems work, so we can get it to do what we want... Covers Amsrad CPC,MSX,ZX Spectrum, TI-83,Enterprise 128/64 and Sam Coupe!
    Lesson P1 - Basic Firmware Text functions
Lesson P2 - More Text Functions, Improvements... and the Sam Coupe!
Lesson P3 - Bitmap graphics on the Amstrad CPC and Enterprise 128
Lesson P4 - Bitmap graphics on the ZX Spectrum and Sam Coupe
Lesson P5 - Bitmap graphics on the TI-83 and MSX
Lesson P6 - Keyreading on the Amstrad CPC, ZX Spectrum and Sam Coupe
Lesson P7 - Keyreading on the MSX, Enterprise and TI-83
Lesson P8 - Tilemap graphics on the Sega Master System & Game Gear
Lesson P9 - Tilemap graphics on the Gameboy and Gameboy Color
Lesson P10 - Tilemap graphics on the MSX1
Lesson P11 - Tilemap graphics on the MSX2
Lesson P12 - Joypad reading on Master System,GameGear, Gameboy and Gameboy Color
Lesson P13 - Palette definitions on the Amstrad CPC and CPC+
Lesson P14 - Palette definitions on the Enterprise and Sam Coupe
Lesson P15 - Palette definitions on the MSX2 and V9990
Lesson P16 - Palette definitions on the Sega Master System and Game Gear
Lesson P17 - Palette definitions on the Gameboy and Gameboy Color
Lesson P18 - Making Sound with the AY on the Amstrad CPC, MSX,ZX Spectrum.... and NeoGeo + Atari ST!!
Lesson P19 - Sound on the Elan Enterprise
Lesson P20 - Sound on the Sam Coupe
Lesson P21 - Sound on the Gameboy and GBC
Lesson P22 - Sound with the SN76489 on the Master System, GameGear, Megadrive (Genesis) and BBC Micro!
Lesson P23 - Sound with the 'Beeper' on the ZX Spectrum and Apple II
Lesson P24 - Bankswitching and hardware detection on the Amstrad CPC
Lesson P25 - Bankswitching and hardware detection on the MSX
Lesson P26 - Bankswitching and hardware detection on the ZX Spectrum

ChibiAkumas Series - Lets look at the chibiakuams sourcecode, see how it works, and how to change it!
    Lesson Aku1 - Screen Co-ordinates and Text Drawing
Lesson Aku2 - Movements
Lesson Aku3 - Sprite Basics
Lesson Aku4 - The Star Array!
Lesson Aku5 - The Object Array!

Appendix
Details on the Amstrad CPC
Details on the Elan Enterprise
Details on the Gameboy and Gameboy Color
Details on the MSX and MSX2
Details on the Sam Coupe
Details on the Sega Master System and the Game Gear
Details on the ZX Spectrum
    Useful info
VASM - Z80, 6502 and 68000 Multiplatform Assembler
Links & Resources for download

Other series - No need to just limit yourself to the Z80... still want more, check out these series!
    Learn Multi platform 6502 Assembly Programming... For Monsters!
Learn Multi platform 68000 Assembly Programming... By Magic!


Platforms covered in these tutorials
Amstrad CPC
Elan Enterprise
Gameboy and Gameboy Color
Master System & GameGear
MSX & MSX2
Sam Coupe
TI-83
ZX Spectrum

Introduction
Welcome to my Assembly programming tutorials, These will be split into parts, the first will teach you the bare basics of assembly language, then we'll jump into some simple programs, once you've learned the basics we're going to jump straight into real game development!

Using my 'Chibi Akumas' (Little Monsters) Multi platform shooter game as a foundation... I'll teach you how to  tweak, enhance, or completely rewrite my game into something of your own, with as much fun and as little theory as possible! As my game engine is open source, once you're feeling ready to go it alone, you can take as much or as little from my game to make your own as you want!

At first we'll begin learning on the Amstrad CPC, as soon as we've covered the basics and moved to the full 'game engine' you'll be able to develop for either the CPC, the MSX or the ZX spectrum as you prefer!

If you want to learn Z80 get the Cheatsheet! it has all the Z80 commands, and useful info on the CPC, Spectrum and MSX!
It will give you a quick reference when you're stuck or confused - and it's what the author used to develop ChibiAkumas!
Print it in color at high resolution on 2 sides of A4 for maximum performance!
The next few chapters are quite technical and confusing, but if you want, just skip them for now, and jump straight into the coding!
This tutorial is designed so you can do that!
You'll need to know the technical stuff explained below one day, but you can come back to it later when you feel you want to!

Feel the power... of the Z80!
The Z80 is a 4mhz 8 bit processor from the 1980's... now by modern standards, it's slow and ridiculously out of date, so why would you want to learn to develop for it?
Well, you can learn a lot about modern computer concepts from the simple Z80, and 4mhz is 4 million commands a second! Which is a heck of a lot when you know how to use them!  
These old 8 bits give a simple system with a lot of potential for the creative person... and while one person could never create a game up to the standards of the latest 'AAA' titles... you could very easily create a game that's as good or better than the best games of the 80's!


Whether you're a fan of the CPC, MSX or Spectrum, these 8 bits have all the power and potential for you to show what you can really do - and you'll learn things doing assembly that you would miss out on with years of C++ or Java development!

If you want to make a game with the latest graphics, of course go download Unity... but if you really want to be in control, and to understand everything that's happening in your code, Assembly gives you the power! No operating system, no drivers, you can take control of everything and make anything you want with it!

Assembly development can be confusing at first... but it has very few commands to learn, Everyone has to start simply, so try not to compare what you're doing to others... just look at what you're achieving, and knowing however 'simple' what you're doing is... it's something you made yourself!

What is the Z80 and what are 8 'bits'
The Z80 is an 8-Bit processor with a 16 bit Data bus!
What's 8 bit... well, one 'Bit' can be 1 or 0
four bits make a Nibble (0-15)
two nibbles (8 bits) make a byte (0-255)
two bytes (16 bits) make a word (0-65535)

And what is 65535? well that's 64 kilobytes ... in computers Kilo is 1024, because four bytes is 1024 bytes
64 kilobytes is the amount of memory a basic 8-bit system can access

Z80 is 8 bit so it's best at numbers less than 256... it can do numbers up to 65535 too more slowly... and really big numbers will be much harder to do! - we can design our game round small numbers so these limits aren't a problem.

You probably think 64 kilobytes doesn't sound much when a small game now takes 8 gigabytes, but that's 'cos modern games are sloppy, inefficient,  fat and lazy - like the basement dwelling losers who wrote them!!!
Z80 code is small, fast, and super efficient - with ASM you can do things in 1k that will amaze you!

Numbers in Assembly can be represented in different ways.
A 'Nibble' (half a byte) can be represented as Binary (0000-1111) , Decimal (0-15) or  Hexadecimal (0-F)... unfortunately, you'll need to learn all three for programming!

Also a letter can be a number... Capital 'A'  is stored in the computer as number 65!

Think of Hexadecimal as being the number system invented by someone wit h 15 fingers, ABCDEF are just numbers above 9!
Decimal is just the same, it only has 1 and 0.

In this guide, Binary will shown with a % symbol... eg %11001100 ... hexadecimal will be shown with & eg.. &FF.

Assemblers will use a symbol to denote a hexadecimal number, some use $FF or #FF or even 0x, but this guide uses & - as this is how hexadecimal is represented in CPC basic
All the code in this tutorial is designed for compiling with WinApe's assembler - if you're using something else you may need to change a few things!
But remember, whatever compiler you use, while the text based source code may need to be slightly different, the compiled "BYTES' will be the same!
Decimal 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 ... 255
Binary 0000 0001 0010 0011 0100 0101 0110 0111 1000 1001 1010 1011 1100 1101 1110 1111   11111111
Hexadecimal 0 1 2 3 4 5 6 7 8 9 A B C D E F   FF

Another way to think of binary is think what each digit is 'Worth' ... each digit in a number has it's own value... lets take a look at %11001100 in detail and add up it's total

Bit position 7 6 5 4 3 2 1 0
Digit Value (D) 128 64 32 16 8 4 2 1
Our number (N) 1 1 0 0 1 1 0 0
D x N 128 64 0 0 8 4 0 0
128+64+8+4= 204            So %11001100 = 204 !

If a binary number is small, it may be shown as %11 ... this is the same as %00000011
Also notice in the chart above, each bit has a number, the bit on the far right is no 0, and the far left is 7... don't worry about it now, but you will need it one day!

If you ever get confused, look at Windows Calculator, Switch to 'Programmer Mode' and  it has binary and Hexadecimal view, so you can change numbers from one form to another!
If you're an Excel fan, Look up the functions DEC2BIN and DEC2HEX... Excel has all the commands to you need to convert one thing to the other!

But wait! I said a Byte could go from 0-255 before, well what happens if you add 1 to 255? Well it overflows, and goes back to 0!...  The same happens if we add 2 to 254... if we add 2 to 255, we will end up with 1
this is actually usefull, as if we want to subtract a number, we can use this to work out what number to add to get the effect we want

Negative number -1 -2 -3 -5 -10 -20 -50 -254 -255
Equivalent Byte value 255 254 253 251 246 236 206 2 1
Equivalent Hex Byte Value FF FE FD FB F6 EC CE 2 1

All these number types can be confusing, but don't worry! Your Assembler will do the work for you!
You can type %11111111 ,  &FF , 255  or  -1  ... but the assembler knows these are all the same thing! Type whatever you prefer in your ode and the assembler will work out what that means and put the right data in the compiled code!

If you need more help, check out 'Convert.bas' on the tools.dsk in the sources file... this is a Hex/Bin calculator that's designed to help you see how binary and hexidecimal compare to decimal - it should help you understand how Hex & Bin work!

Working with numbers in the Z80
We know what numbers we can have on the Z80, but how do we work with them?
When we need to save our results, we will store them to that 64k of memory mentioned earlier, but the memory is some distance away from the processor... to be fast we want to use the memory built in to the processor.
This z80 memory is called the 'Registers'... there's not much of it, but its really fast - so try to use them to do as much as you can! Each Register can only store one byte (0-255)... but some registers can be paired ... so HL together are 16 bits, and can strore 0-65535!

C
Each of the registers has a 'purpose' it is intended for... Of course you can use any register for anything you want! but they all have 'strengths' because many commands will only work with certain ones... and some commands may be slower or need more code if you use the wrong one!

A This is used during all our main calculatons - you'll use it all the time when adding up (Accumulation)!
HL This often stores memory locations, as there are a lot of special commands that use it to quickly read or write whatever memory locations.. it's also good at 16 bit maths, so if you want to add two numbers above 255, you'll probably need it!
BC These are often used as a bytecount or loop counter... sometimes you'll use B and C together, for a count of up to 65535, or just B on its own for up to 255...
DE Destination - if you're reading from one place and writing to another, you'll probably use HL as the source, and DE as the destination
IX Sometimes we want to get to memory by specifying a relative position - Indirect Registers allow us to do this ... for example if we have sprites, and each 4 bytes for X,Y,Width,Height.. just point IX to the start of the data for the sprite we want - and read the rest out as IX+1 , IX+2 etc...
Don't worry about this - we'll explain it later... IX is actually a pair of two registers  called IXH and IXL - we can use them alone for whatever we want - but they are slower!
IY IY works the same as IX
   
PC This is the place in memory that the Z80 is running - we dont' change this directly
SP This is the stack pointer - it's points to a big temporary store that we'll use for 'backing up' values we need to remember for a short while
R This is the Refresh register - the system uses it to know when to refresh the memory... don't change it ! you could mess something up - but it can be used for getting simple 'random' numbers!
I This is the Interrupt point... on the CPC and MSX it's pretty useless so you can use it as a 'temporary' store... but on the spectrum it's really important  and you'll have to leave it alone!

The Gameboy is not a Z80, but it's similar in some ways... however, the Gameboy 'GBZ80' CPU has no IX or IY, and no shadow registers! 
We'll be covering the gameboy's weirdnesses more later!
The shadow registers are a 'copy' of the main ones... on the CPC the system firmware uses these during 'Interrupts' so you can't have them!
But if we stop the interrupts then we can use them for even more power!
Interrupts do things like read the keyboard and update the system - but we canreplace the firmware and  take over these jobs ourselves and take total control of the Z80 so we have all the power!



Lesson 1 - Getting started with Winape!
For our Z80 development we'll be using the WINAPE Amstrad CPC Emulator - it's free and does everything you need for development in one place, so I highly recommend you use it.
Winape works great in Windows XP - so you can run it in a virtual machine if you are a Linux or Mac fan, and I believe it works with WINE too.

There's a video of this lesson,  just click the icon to the right to watch it ->

You may want to develop for the ZX spectrum or MSX, but you should start with the CPC and WINAPE!
Winape has a built in Assembler and Debugger that are second to none!  In the early days you will make a lot of mistakes, and having the compiler , emulator and debugger in one place will save you a lot of time and problems! You can even develop MSX and SPECTRUM code with it - the MSX and Spectrum versions of ChibiAkumas are compiled in winape too!.... If you dont't use widows, Winape works OK with WINE!


To get started, Download Winape from this Link (I'm using V2.0b2)  and extract it to a folder,
Start Winape by clicking on the icon
You should see the emulator window open and show the amstrad CPC blue screen... the CPC starts straight away into Basic... and Amstrad's basic is great for testing ASM code!

Don't worry if you've never used a CPC before, this tutorial assumes you know nothing about basic or the CPC!

Now, Let's get straight into it and code something!

Click on the Assembler menu and select "Show Assembler" - or press F3 on your keyboard!
Select the File Menu, then New to create an empty ASM program
Type in all the lines shown to the right - Don't worry about what they do, We'll look at that in a bit!

Also, don't worry about the case as winape isn't case sensitive.... ORG &8000 is the same as org &8000
Select Assemble from the Assemble menu to compile the program!
You will see the Assembler output.  This shows the actual bytes of code that what you typed ends up as!

Check there are Zero Errors ... if there are then you've made a mistake typing!

Click on OK!

Now this is the magic of Winape - your code is now immediately in the emulators memory - so lets go run it!
Now go back to the blue screen... and type in Call &8000 and hit your enter key

Basic should return the message Ready

if something else happened, check your code matches the screenshot above!

What did it do? well we'll look at that in more detail next!
Congratulations! You just wrote and ran your first assembly program - and you now officially Kick Ass!
Now we'll look in detail at what the program did, and how to use Winapes debugger to look at it in detail!

Now you've had a a quick go at assembly, lets look in detail at what that program does! We'll look at each line of the program, and explain what it means.
 
org &8000 ORG sets the ORIGIN of the code in memory - the Z80 has a 16 bit address bus, so memory goes from &0000-&FFFF - but we can't just use anywhere, as some bits are used by other things - for example &C000-&FFFF on the CPC is used by the screen.
&8000 is a good 'safe area' on the Amstrad for your code to start
ld a,4 load A (the accumulator) with the number 4 - The Accumulator A is the main register used for calculations - registers are used for very short term memory storage
inc a INCrease A by 1... A will now contain 5
inc a INCrease A by 1... hopefully you can guess it now contains 6
ld b,10 Set B to 10 ... B is another register... it can't do as much as A, but it can do a lot... notice the destination is on the left, and the source (the number) is on the right... this is always the case with Z80 assembly
add b this adds B to A ... although A isn't mentioned in the command, since the Accumulator A is used for almost all maths any time a destination isn't mentioned, it will be A
this command could be written as ADD A,B... winape will compile this, but the "A," is superfluous, so you're better off learning not to need it
LD (&7000),A When Brackets () are used in assembly they define a MEMORY location... &7000 is the a number (&7000) is the content of that memory point!
So this command puts A into memory location &7000!
RET RET returns back to whatever called the program... in this case basic!
 
Ok, we've read over the code, lets use the debugger to see the Z80 run the code!

If you still have your assembler open, just click on its window
If you closed it, just click  Show Assembler from the Assembler menu
Click on the Grey Area next to the command "LD A,4"

A red blob will appear! This means before the Z80 runs the command "LD A,4" it will stop, and the Debugger will appear!
Select Assemble from the Assemble menu to recompile the program!
Type Call &8000 again into basic, and hit Enter
If you've done everything right, the debugger will immediately pop up! Lets take a look at what it offers!

At the top of the screen you can see the Compiled code... the line the Z80 is running is highlighted.

On the Right you can see all the Registers - you should recognize the names! remember our code uses A and B... A is the first half of AF and B is the first half of BC

In the middle of the screen you can see the Memory

At the bottom of the screen you can see the Rom and other debugger options... don't worry about them now!

In the bottom right is the Stack... we'll cover that very soon!
The Debugger has tons of options we're not going to use right now, but feel free to try them later! Try right clicking on things for more options!

You can even change the registers and memory by typing new values in! and remember - The worst you can do is crash your emulator, so there's no real harm you can do - so don't be afraid to mess!

If you crash your emulator, the memory will be erased, so you'll have to re-assemble your code from the Assembler before Call &8000 will work again!
Now Lets look at what your code does!

Press the F7 key on your keyboard (or the single step button on the main window)

Keep an eye on the A and B registers in the debugger!

You should see A change to 4... press F7 a few more times and watch what happens
keep pressing F7 again and again until you get to the RET command!

You can press it even more if you want, but remember... the debugger isn't just debugging your code, but the whole CPC... so you'll end up debugging the whole of basic! 

When your code has finished running we have one thing left to do....
Remember the program wrote to &7000 ?

Scroll through the memory browser to find &7000 - and you should see our final value for A  has been saved there!
You can open the debugger by selecting Pause from the Debug menu at any time,  or by pressing F7 - and just close it when you've seen enough and want your emulator to run normally!

Think back on what you've just done, You've written a program, compiled it, run it, and watched everything the program did at the CPU level in more detail than you've probably ever seen any computer work before! Not bad for a days work! - and this is just lesson 1!

Adding a few numbers may not seem much, but it's just the start, and you'll soon find you can make the computer do amazing thing!
Also, don't forget, once you understand the basics, you can use pieces of other people's code to do the work you don't want to! These tutorials will show you how to build on the open source code of the Chibi Akumas game to allow you to make big progress super-fast!

I highly recommend you type the programs in yourself, but you can download the source code with comments Here (Contains source for all lessons)



Lesson 2 - Memory copy, Symbol definitions, Loops and Conditional Jumps!
Now you've got Winape up and running, and had a go at programming, we can get on with learning some more commands!
We're still going to do simple things, but let's use the CPC's screen this time, so you can see the results of your code!

The CPC screen memory is at &C000 and takes up &4000 bytes!

Open up the assembler
     (Remember: Press F3 or select Show Assembler from the Assembler menu)
Create a new document
      (File... New)
Type in exactly what you see to the right!... there is a typo in this code  - so we're going to have to debug it

Once you've typed it in assemble it! 
     (Ctrl-F9 or Assemble from the Assemble menu)
Oh no! there's an error!... what a surprise!

Click OK to close the assembling window, and lets sort that error out!
The error that occurred will be show at the bottom of the screen... double click on it, and the cursor will jump to the error!

Whoops, we've put BD ... theres no such pair! correct the error to BC

Now reassemble the program, and you should get no errors!
Type Call &8000 and hit Enter

The screen will clear, and the top line or two will have some weird junk on it!


You probably think this is weird, but when we look at the code, This is exactly what we'd expect to happen... so lets break it down line by line!

Our code is at &8000 - that's where we call to run it org &8000
Set HL to &0000 - in this case HL is a source memory address ld hl,&0000
set DE to &C000 - in this case DE is a Destination, and &C000 is the start of the CPC screen ld de,&C000
set BC to &4000 - in this case BC is a byte count - and &4000 is 16k, the size of the CPC screen! ld bc,&4000
LDIR means " Load, Increment Repeat"... in effect it copies BC bytes from memory pos HL to DE
ldir
return to basic ret

So we've copied 16k from &0000 to the screen - and what is at &0000 - well some system junk, and any basic progs!
don't believe me?... type:
    10 print "moo"
    Call &8000
and you'll see some more junk appeared! that new junk is your program as it appears in memory!
Of course if you set DE to &0000 and HL to &C000 - you'd copy the screen to your system area - and your machine will hang!
it will also hang if you set BC to &5000... why? well when the destination gets to &FFFF it rolls over to &0000 - again overwriting your system area - but give it a go if you want, there's no harm crashing an emulated machine!

By default the CPC screen starts with &C000 at the 'top' but if you make basic scroll up and down, the 'Top' will move somewhere else, if you press the down cursor until the screen scrolls - then call &8000 again you'll see the junk appear somewhere else
You can always reset the cpc by typing "Mode 1" in basic (reset to screen mode 1)

If your crazy enough to try this on a different system, you'd need to make a few changes, The ZX Spectrum screen starts at &4000 not &C000, and the size is &1800
On the MSX the screen isn't in memory so you can't see the effect on screen - and you'll crash the system if you write to &F000 or higher, but you could set the size to &1000, and see the change in memory between &C000-&D000 with a debugger if you want!

Ok, you've had a quick look at LDIR,
Now lets try another more useful example,

Type in the example code to the right, and assemble it as before
Assuming you got no errors, lets run the program.

The ORG was different this time, so type Call &8100

The screen will clear!

What's interesting is that this simple ASM code is faster at clearing the screen than the firmware's own CLS command!... and there's a lot of tricks we can do to speed it up even more... but we'll leave them for another day!

What did that code do? well lets break it down!

This line defines a 'symbol' ... we're telling the assembler 'ScreenSize' EQUals &4000
A symbol is a constant in your code - basically every time the assembler sees 'ScreenSize' it will put in &4000
there's a couple of reasons to do this, firstly it's makes your code easier to read, and secondly, if you need to change it to &2000 - then you can just change this definition in one place, rather than all the places you've used it!
ScreenSize equ &4000
set the origin to &8100 (I'm using a different one, as all these examples are in the same 'lesson2.asm' in the download file) org &8100
Set the source HL to the start of the screen memory... yes, you read that right, the SOURCE ld hl,&C000
set the destination DE to one byte after the source... note that the assembler calculates what &C000+1 actually equals - not the Z80 ld de,&C000+1
set the Byte count to our defined symbol -1 (this will compile to ld bc,&3FFF) ld bc,ScreenSize-1
set the first byte of the source to 0 ld (hl),0
run our LDIR copy... now here's the trick...
at first LDIR copies the byte from HL (&C000 we just set to 0) to DE (&C001)
next  LDIR copies the byte from HL (now &C001 which IT just set to 0) to DE (now &C002)
after that  LDIR copies the byte from HL (now &C002 which IT just set to 0) to DE (now &C003)... and so on!
We tricked LDIR into copying it's own data, and acting as a quick(ish) FILL command!
ldir
return to basic ret

Mathematics in ASM code in Winape code is pretty dumb... it can't do brackets like 2*(5+1)... and it doesn't do multiplication first.... Usually you'd expect 5+1*2 to be equal 7... but in winape this equals 12!
Why? well winape does each command from left to right, so 5+1=6... then 6*2=12...it takes some getting used to, but it does work fine!

Conditions and Loops!

We have one last version of this program to try!

Type in the code to the right, and compile it... you should get no errors.

There's some new commands in here, but take a look at them in a minute!
Type Call &8200 and hit enter
The screen will do a strange 'fading clear'... Weird huh?

Ok, it's not very useful, but it teaches us a lot of good stuff!

Lets take a look at the code, and see why it does what it does!

Define our constant symbol "ScreenSize" as &4000 ScreenSize equ &4000
Start our program at &8200 org &8200
Set A to %00001111 - this sets all 4 pixels of a screen byte to cyan on the cpc screen ld a,%00001111
This is the definition of a label...
it's like a the constant "screensize" in that the assembler converts future mentions of 'FillAgain' to a number when it compiles... however unlike 'screensize' that was defined with 'equ'...  a label points to a memory location in the compiled code... if you take a look at the assembled code you'll see it is in the position &8202... that's because 'ld a,%00001111' takes 2 bytes
FillAgain:
set HL to the start of the screen ld hl,&C000
set DE to one byte later (the same as last time) ld de,&C000+1
set BC to ScreenSize -1 ... you should recognize all this from last time! ld bc,ScreenSize-1
load the first byte with A - remember it will set  all the pixels Cyan ld (hl),a
Just like last time LDIR will fill the whole screen ldir
Decrease A by 1 dec a
Compare A with 255... when A=0 and we do DEC A... A will become 255... cp 255
if A does not match 255 , jump back to label 'FillAgain'
The next command the Z80 will run will be 'ld hl,&C000'
jp nz, FillAgain
Now you know how to do a loop! in fact JP Z and JP NZ can be used like LOOPs, IF statements and even CASE statements!
Actually - you don't have a lot of choice as assembly has so few commands - but remember... all other programming languages compile down to Assembly - so anything Basic or C can do is possible in ASM - and ASM will always be fastest if you do it right!


So we've used a Label, and a jump (JP) to create a loop!
there are three kinds of Jump command that you should know!

Command Meaning Example
JP ## This is like a GOTO command in basic,  it will just jump to the label or memory address you specify every time. 
this command takes 3 bytes
JP &4000
JP c,## This is like an 'IF X THEN GOTO' command in basic. if condition is met the jump will occur, otherwise it will continue
otherwise program execution will continue... there is no such thing as an ELSE, but you can always immediately do another jump!
this command takes 3 bytes
JP Z,&4000
JR # do a Jump Relative to the current location...
This is a bit tricky, but all you need to know is it takes 1 byte less than JP, but can only jump nearby - so it saves memory, but you can't always use it!
JR label
JR c,# Same as above, a 2 byte relative jump - saves one byte over JP c,## - but can't always be used if you need to jump far away JR Z,label
djnz # This is a special  'quick small loop' command... it automatically Decreases B and Jumps if b is NotZero
DJNZ only takes up 2 bytes, so it's small if you only need a basic loop.
a jr jump can only jump to a label that is nearby - so if you get an error use this alternative which can jump anywhere
     dec b
     jp nz,label
DJNZ label
JP (HL) Jump to memory location in HL, this is quite advanced, so don't worry about it now... but you can load HL with a label, then use this to perform a jump if you need to JP (HL)


So those are the Jumps we have available, but some of them need a condition too!...
we have 4 main conditions we need to so lets take a look at an example - you don't need to type it in

load the accumulator with 4       LD a,4       
Compare A to 10 CP 10

There are 4 basic conditions we can use in this situation - it's annoying, but what the condition  officially means, and what it does in this case are different

Flag example Official meaning *** Basic program equivalent What it means when using CP
Carry JP C,label Carry  is used with bitshifts and addition - if a byte goes over 255, it will go back to zero, but Carry will be True A<CP if A < CP Value then C is true and JP C,label  will make the jump to label
NoCarry JP NC,label NC is true when there is no carry A>=CP if A > CP Value then NC is true and JP NC,label  will make the jump to label
Zero JP Z,label Z is true when the last mathematical operation resulted in zero A=CP if A=CP then Z is true and JP Z,label  will make the jump to label
NonZero JP NZ,label NZ is true when the last mathematical operation did not result in zero A<>CP or A!=CP if A<>CP then NZ is true and JP NZ,label  will make the jump to label
*** Note: while CP always works, not all mathematical commands affect all flags - see the cheatsheet for full details! - eg "inc hl" does not set the zero flag!

NC NC Z and NZ don't make a lot of sense for < > = and <> ... and they're a real pain to remember!
Z and NZ are = and <> ... so try to think of them as 'is the difference Zero - or Non Zero?'
C and NC are < and >= ... so think of them as Chibi (smaller) and NonChibi (bigger or equal)
They're on the Cheatsheet, but if you can come up with a way of remembering which is which - all the better!

Well done! You finished Lesson 2! But there's no reason to finish with these programs if you don't want to!
Try changing some of the numbers!
What happens if you change 'ld a,%00001111' to 'ld a,%11110000' or 'ld a,%11111111',
or change 'ScreenSize equ &4000' to 'ScreenSize equ &2000'...
Have some fun! and try lots of things!...
There's a video available describing more about how CPC & Spectrum screen memory work if you want to learn more now!



Lesson 3 - 'Case Statement' , 8 bit basic Maths, Writing to Ram and Reading from basic
We used a Jump to effect a loop last time, but sometimes we'll need to jump to different places depending on a value.

Also, lets take a look at how to do some everyday maths in 8 bit... and finally, we'll use a simple basic program to act as a 'frontend' to our assembly

Lets get straight into coding!

We're going to write a little Assembly calculator

Create a new assembly document, and type in the program to the right.

Compile it - you should get no errors!

You'll notice that the program takes it's input and output from 4 memory locations:

&9000 Command Num
&9001 Var 1
&9002 Var 2
&9003 Result

We're going to access these from a basic program - which we'll write now!
Type the program in to the right in basic

if you're not familiar with Basic, Don't worry about what the commands do, we'll look at them in a moment

ADD and SUB in assembly can add or subtract up to 255 from a register, but if you only need to add or subtract one, use INC or DEC, they increase or decrease a register by 1...
INC and DEC commands take only 1 byte, so they're faster than ADD and SUB... and they work on 16 bit registers like HL
ADD and SUB only work on the 8 Bit Accumulator, but we'll learn how to do 16 bit equivalents later!

CPC basic is really easy, just start typing the lines in after the emulator starts.
If you make a mistake it's easy to fix- for example to edit line 30, just type "Edit 30" and hit enter!
CPC basic also has a strange 'copy' command, which allows you to 'copy' text already onscreen... just hold down shift and use cursor keys to create a 'shadow cursor' and use Alt to copy the letters under the shadow cursor... on the CPC the 'Alt' key was marked 'Copy'

Type RUN to start the program and hit enter.

Enter the values for the Variables as 20 and 5, hit enter after each value

When asked if you want to Add or Subtract, Enter 1 for subtract

The result will be show onscreen!
Feel free to try other values in the program, but it only uses 8 bit registers, so it can only do up to 255, and can't do negative values right!
Right now it'll only do Addition and Subtraction because there aren't any built in Multiply or Divide commands on the Z80 - we're going to work around that next!

Lets take a look at the ASM code!

program starts at &8000 org &8000
Read memory position &9000 into A ld a,(&9000)
read two bytes into 16 bit register BC... on the Z80, 16 bit registers are stored in memory in reverse
So C is loaded from &9001 and B is loaded from &9002
ld bc,(&9001)
We want to do a 'Case statement' where we do different commands depending on A, but no such command exists!
No problem! we just do is lots of CP x commands and JR Z,label commands
Remember, CP # always compares with A... and JR Z will do a jump if A matches with the compared value #
cp 0
Command 0 is add, so if this is what the user selected jump to the MathAdd label jr z,MathAdd
Compare A to 1 cp 1
Command1 is subtract, so if this is what the user selected jump to the MathSub label jr z,MathSub
if we got here, then A was something weird, so set A to 0 ld a,0
This is our SaveResult label, if a command was run then it will finish here, if the user put a strange number in, execution will also end up here SaveResult:
Load the result (in A) to memory point &9003 ld (&9003),a
return to basic ret
   
The start of out subtract routine MathSub:
we need our result in A, so load from C (Val1) into it ld a,c
Subtract B (Val2) from A sub b
we've finished, so Jump to our saveresult label jr SaveResult
   
The start of our addition routine MathAdd:
we need our result in A, so load from C (Val1) into it ld a,c
Add B (Val2) to A add b
we've finished, so Jump to our saveresult label jr SaveResult
Sometimes in ASM there's a smaller, faster command that  has the exact same result as another!
For example, you can use "OR A" instead of "CP 0"... and "XOR A" instead of "LD A,0"
The result is identical, but you'll save some speed and memory!... it'll just look a bit odd in your code!
It's something you'll want to learn to use.. so why not give it a try now!

In case you're not familiar with CPC basic, lets take a look at the basic code!

Clear the screen CLS
INPUT asks the user a question, and stores the users response in a variable
So this will show "Val1?" onscreen, and store input from user into a
     Note: a/b/c in basic is nothing to do with A/B/C in ASM!
INPUT "Val1";a
Show "Val2?" onscreen, and store input from user into b INPUT "Val2";b
ask the user what command to run INPUT "0=Add,1=Sub";c
POKE writes a byte into memory, this writes our command number into memory point &9000... this is how we'll get our values from basic to ASM POKE (&9000),c
Store Value from A into &9001 POKE (&9001),a
Store Value from B into &9002 POKE (&9002),b
Call our ASM program CALL &8000
PEEK reads a byte from memory, allowing us to get the result our ASM program produced.
PRINT just shows it onscreen
PRINT PEEK(&9003)

Using basic to 'launch' and test your ASM code is a great way to develop quickly and ease testing
Using PEEK and POKE to get data to and from your program is a good solution, but you can also pass variables using the CALL command, but it's a little tricky, so we'll cover it later!

Multiply and Divide

We want to add Multiply and Divide commands, but unfotunately the Z80 does not have these commands! but we can simulate them by repeatedly using the ADD or SUBtract commands!

Add These Lines to the bottom of your code below MathAdd
Add These Lines to 'case' condition block

use EDIT 30 to edit the line showing the options, make it the same as This
Run the program!
You'll now be able to do multiplication, but only if the result is less than 255! You'll also see that negative numbers don't work!

If you try a number that ends up too high, or below zero, you'll get a strange number, that's because the numbers 'roll around' from zero back to 255

Later we'll upgrade the program to use 16 bit numbers, so we can go from -32768 to 32767!

Take a look at the Hexadecimal tutorial at the start of this document if you want to know more about negative numbers now!

You can get the sourcecode for this lesson (and all the others) in the sources.7z file... the basic code can be found on the included disk image!

Repeately adding or subtracting a number to 'fake' Multiplication or Division is silly and slow, but if you only need to Double or Halve a number you can use bit shifts... we'll cover them soon!
You want to try avoid needing Multiplication and division if you can in your code, so design your game not to need anything except halving and doubling... of course you can do x4 or x8 by doubling twice or three times!




Lesson 4 - Stack, Strings,Compiler Directives, Indirect registers, CPC Call
That was a nice little program, but we still have some basics to cover!

You'll have noticed there's not many Registers, and you'll often wish you had more. So what can you do if you have a value you need later, but you need to do something else first?

We need some temporary storage, and we have something called the stack for that!

Lets suppose we have a value in A ... we need the value again later, but we need all our registers now too... what can we do? Well that's what the stack is for, it's a temporary store!
The stack is like a letter tray - we can put an extra sheet on the top of - and take it off later, but we can't get one from the middle, so we always get the Last one we put in - this is known as LIFO - Last in First Out
We use PUSH and POP to push a new item onto the stack, and pop it off later.

Lets look at two examples, and see why we want the stack!

Suppose we have a Call 'PrintChar' which will print the character in A ... and another call 'DoStuff' which will change all the registers - how can we keep A the same before and after this 'DoStuff' command?  Well, we could save A to some temp memory - or let the stack take the slack!

don't type this in, it won't actually work!

Without the stack
ld a,'1'
LD (Temp),a
call  PrintChar
call DoStuff
ld a,(Temp)
call  PrintChar
Temp: db 0
With the stack
ld a,'1'
push af
   call PrintChar
   call DoStuff
pop af
call Printchar

Both these do the same thing, but  commands like LD (temp),a takes 3 bytes, and PUSH AF takes only 1.. and it's faster! also you no longer need that Temp: db 0 ... so that's another byte saved!

the stack always works in 16 bit - so even if you only want to save B - you'll have to PUSH BC - but don't worry , it's so fast you won't mind!
note:  the F in AF is the 'Flags' (the Z NZ C NC in comparisons) - they're saved with A when you do a push.

You can push lots of different things onto the stack, but remember they will come out in the same order... you can even do  PUSH BC then POP DE to copy BC into DE
The order is important! if you're unclear on how the stack works, you can use the Debugger in winape to step through your program, and see what the stack does as your program works!
seeing things happen step by step in the Z80 is a great way to see how things are happening
The Stack

Lets do an example of the stack!

Type in the program to the right!
Type Call &8000 and hit enter.

The screen will 'Pause' so press A

You should see A|x|A onscreen

Run the program again, and press a different letter!

What was the point of that? well lets take a look at what the program does!  I've colored the PUSH, POP commands so you can see what POP gets the matching value that was PUSHED

This is a definition pointing at a command in the CPC firmware - it will print A to the screen as a character PrintChar equ &BB5A
This is a definition pointing at a command in the CPC firmware - it will wait until a key is pressed and save it in A WaitChar  equ &BB06
Start of the program at &8000 org &8000
Get a character from the user -this is why the screen paused - lets assume the user pressed A call WaitChar
Print the character to the screen call PrintChar
Push the character onto the stack - in this example 'A' push af
Load a bar symbol into A     ld a,'|'
Push the bar onto the stack for later     push af
Print the bar to the screen         call PrintChar
Load an 'X' into A         ld a,'x'
Print the X to the screen         call PrintChar
pop an item off the stack... we will get the Bar we just pushed     pop af
Print the bar onto the screen     call PrintChar
pop an item off the stack, we get the character the user entered - in this example  'A' pop af
print the character ('A') to the screen call PrintChar
Return to basic ret

So we can push items onto the stack, and get them back later so long as we need them in the same order!  But the stack doesn't just operate for us.. the Z80 uses it too
When we do a 'CALL label' command, the current running address is pushed onto the stack - effectively the Z80 does 'Push PC    JP Label'
When we do a 'Ret' command... the Z80 effectively does 'pop PC'
Pc is the program counter - the current address the z80 is running... now this Push PC and Pop PC command don't really exist, but that's what the Z80 does - and you need to know this so you know why this program won't work:
Command What happens
Ld a,'Z' Accumulator set to character 'Z'
push af AF pushed onto the stack
Call ShowIT the address of the next command (ret) is pushed onto the stack
Ret end of the program
ShowIT:
pop AF we wanted to get  'Z' back - but we actually got the address of the command after Callit
call PrintChar We wanted to print 'Z" but we actually printed half the address!
ret we wanted to return from the call - but we stole that off the stack - return jumps to the AF value we pushed - and the computer will crash!

If you don't understand the stack yet, try making some test programs, or editing the one you just typed in!
The stack's used so much you'll see plenty of examples - and you'll soon get used to it!

The stack is the fastest way to read and write memory on the Z80 - if you're clever you can use it in a 'tricky way' to quickly do things like fill the screen.

We'll learn how to do it later - it's an advanced trick, but if you want to make the Z80 as fast as possible - that's how to do it!

Compiler Directives

We'll move on from the stack - Lets have a look at another little program! - we're going to use the previous PrintChar command to make a string printing routine.

Type in the program to the right.

Assemble it - you should get no errors

We'll explain what each line does after we run it!
Use Call &8100 to run the program
It will print a little two-line message to the screen!
Chage the first line - put a Semicolon ; at the start of it - this marks it as a 'comment' - which means it does nothing in the code

Assemble it - you should get no errors
Run the program again - The message has changed!

Writing somthing as simple as a print string routine may seem a pain - and you're probably wondering why the firware can't do it for you,
But writing your own routines is the best idea - firstly - you'll know exactly what they do, and you can change them later to add special functions - and more importantly - you can port them to other systems and they will work the same!
The less you use the system firware the better! your Z80 code will work the same on a CPC / Spectrum or MSX - firmware calls will not!

That program probably seems rather long, but there's lots of good stuff in there! Lets take a look at it line by line!

Define a symbol called 'ThinkPositive' and set it to 1 ThinkPositive equ 1
Define PrintChar, and point it to the memory address in the Amstrad Firmware that Prints ascii character A to screen PrintChar equ &BB5A
Start of our program - the two lines above are instructions to the assembler, they do no compile to anything the Z80 sees org &8100
load the address of the Introduction string into HL ld hl,Introduction
call our PrintString function call PrintString
Call our NewLine function call NewLine
Load the address of the Message into HL ld hl,Message
call our PrintString function call PrintString
Return to basic ret
Start of out Printstring function PrintString:
Load a Byte (character) into A  from the address HL we were given ld a,(hl)
Was the byte 255? cp 255
If it was (the difference is Zero) then we're reached the end of the string - so return to the calling program ret z
increase the HL address counter inc hl
call our PrintChar routine to show the letter in A onscreen call PrintChar
Repeat the procedure. jr PrintString
A label defining the address of our introduction message Introduction:
a string of letters, ending with 255 - the assembler will convert these to their equivalent bytes according to their Ascii code. db 'Thought of the day...',255
remember that 'ThinkPositive' symbol? well we're telling the assembler that we only want to do the following if we've defined it! ifdef ThinkPositive
A label 'Message' and a message that will compile when the IF statement above is true. Message:    db 'Z80 is Awesome!',255
If 'ThinkPositive' is not defined (for example - when we put a semicolon in front of it) else
A label 'Message' and a message that will compile when the IF statement above is false - Note: Normally you can't have two labels with the same name, BUT because the IF statement means only one will compile there ARE NOT TWO in the final program Message:    db '6510 sucks!',255
End of the Assembler directive endif
NewLine:
A newline command ld a,13
Load Character 13 into A (Carriage return) call PrintChar
Print it ld a,10
Load Character 10 into A (New Line)  call PrintChar
Print it ret

Feel free to try other values in the program, It's important to note that the IFDEF is actually changing the Compiled code -the 'message' that is not shown DOES NOT EXIST in the compiled data!
This allows you to compile multiple versions of your program - Chibi Akumas uses this to compile different builds of the game for CPC, ZX spectrum and MSX - and for different languages - all with one code base!


The Indirect registers IX and IY

Suppose we have some bytes of data we want to read from a 'bank' of data - but we want to read those bytes by specifying an offset relative to the start address - we can use the Indirect register IX or IY to do this - lets look at an example

Type in the program to the right, and compile it.

It's pretty long, but The bottom part is identical to your previously entered one, so just copy and paste it, from your last example - or just add the new code to the bottom of your old one.

If it's too long, you can always download the sources file and just run it from there.
Run the program by typing Call &8200
It prints a message inside two kinds of brackets

Don't underestimate calls from basic! A great way to make your first game is to write the logic and input routines in basic, and call out to assembly for things like drawing sprites and music!
The important thing is the result you achieve, not the method - so why not make things easy for yourself and do some of the work in basic - you can always convert it to ASM later once you've worked out exactly what you need to do!

Lets look at the program and see how it works!

Start of the program org &8200
Load the Indirect register IX with the address of the square brackets ld ix,SquareBrackets
Load the address of 'Message' into HL ld hl,Message
load DE with the address of the Printstring function ld de,PrintString
Call the DoBrackets function - we'll take a look at it in a second call DoBrackets
Call the Newline command call NewLine
Load the Indirect register IX with the address of the curly brackets ld ix,CurlyBrackets
Load the address of 'Message' into HL ld hl,Message
load DE with the address of the Printstring function ld de,PrintString
Call the DoBrackets function - we'll take a look at it in a second call DoBrackets
return to basic ret
Start of the Dobrackets function DoBrackets:
Load A from the address in IX (plus zero - so just IX)  ld a,(ix+0)
Print character A call PrintChar
Run the function DoCallDE - we'll look at it in a moment. Call DoCallDE
Load A from the address IX plus 1 - note this does not change the value in IX ld a,(ix+1)
Print character A call PrintChar
ret
The function DoCallDE DoCallDE:
Push DE onto the stack push de
Return will take two bytes off the stack, and continue execution from that point - effectively we have done the command Call (DE) - the Z80 has no such command, but we have simulated it here ret

The Chibi Akumas game uses IX to point to the Player settings - the Player routine gets the Life - position - health etc of the player via IX+... references
This allows the same routine to handle both players just by changing the IX reference when the function is called - the first time it points to Player1's data - the second time Player2's


Call with parameters

That example of IX was rather useless, but the IX register has a far more useful function on the CPC - we can use it to get data from the Call statement! Lets take a look at an example!
Type in the example to the right - it's rather long - if it's too much trouble, remember all the examples are in the downloadable sources file.

The Printstring function is the same as before - so you can just copy and paste it.
Try typing Call &8300

Because you didn't give a parameter You will see a 'usage message'

Type Call &8300,12345

You will see '12345' converted to 16bit hexadecimal!

Feel free to try it with some other numbers!

We've created a Decimal to Hexadecimal converter! and it gets its data straight from basic!

Lets take a look at the code and see how it works
org &8300
When the program starts CPC basic will store the number of parameters passed in A cp 1
If the user did not pass 1 parameter, show how to use the program jp nz,ShowUsage
Print an & symbol onscreen ld a,'&'
call PrintChar
Numbers are passed as 16 bit integers... The first parameter location is passed by basic in IX - because the data is passed in little Endian it's backwards, so we load the larger part from IX+1 into A ld a,(ix+1)
if the high byte is not zero, call our ShowHex function or a
call nz,ShowHex
Load the smaller part from IX+0 (IX) int A ld a,(ix+0)
Call ShowHex call ShowHex
return to basic ret
Show the usage message to the user - as the last command is a JP there is no need for a return command. ShowUsage:
    ld hl,ShowUsageMessage
jp PrintString
Message string we show the user if they used the program wrong ShowUsageMessage:
    defb "Usage: Call &8300,[16 bit number]",255
ShowHex function - this shows an 8 bit byte in Hex ShowHex:
we want to Divide the number in A by 16 ld b,16
Call our Divide function call MathDiv
The remainder is returned in A - we need it later - so we Push it now push af
Load the result of the divide - this is how many 16's there were in the byte - so this is the first symbol in the hex string ld a,c
Print the hex string call PrintHexChar
get back A - this is now the second digit with a pop pop af
jump to the PrintHexChar - because we don't use CALL there is no need for a return jp PrintHexChar
This function prints a single hex digit 0-F PrintHexChar:
Compare A to 10 cp 10
if A is less than 10 we need to print a digit. jr c,PrintHexCharNotAtoF
Add 7 to A - this is the Ascii difference between 9 and A add 7
PrintHexCharNotAtoF:
Add 48 (0) to the digit - this converts A to an Ascii character add 48
Print it jp PrintChar
This mathDiv function is different from lastweeks, it stores the result in C and the remainder in B MathDiv:
Reset C - it will store the result ld c,0
ifA is zero then return cp 0
ret z
MathDivAgain:
Subtract B from A sub b
increase C (the result counter) inc c
Repeat if we've not gone below zero jp nc,MathDivAgain
we've gone over zero, so add b again, so A contains the correct remainder add b
we've gone over zero, so decrease C to get the correct result dec c
ret

Using IX is great - but it only works on the CPC - the MSX can pass one variable (see the basic documentation) but other systems cannot really do this - just use the POKE function in the previous example instead!
Other than CALL commands IX functions are great for settings data - you can use them for passing references to sets of data that you want to access and alter 'randomly'


Lesson 5 - Bit level operations, Self modifying code

Because memory and commands are limited, you'll quite often want to do things at the bit level, You can use bits to alter numbers, create patterns, and with conditions control actions from the different bits in a single byte

Lets use the CPC screen to see what bit commands do!
AND, OR and XOR

Type in the program to the right, and compile it
Type in Call &8000, and see what happens
The screen colors will have gone weird!

It's not clear what happened in 4 color Mode 1

Type in Mode 2... and Call &8000 again!
Try changing the XOR to AND or OR

Change %11111111 to other values like %11110000

Lets take a look at what that does to Mode 2 - where each bit is a pixel!

Sample XOR %11110000 
Invert Bits that are 1
AND %11110000 
Keep Bits that are 1
OR %11110000
Set Bits that are 1

Lets see how that works at the bit level!

Command LD A,%10101010
XOR %11110000
LD A,%10101010
AND %11110000
LD A,%10101010
OR %11110000
Result %01011010 %10100000 %11111010
Meaning Invert the bits where the
mask bits are 1
return 1 where both bits are1 Return 1 when either bit is 1

Each Bit is a pixel in Mode 2 - but in Mode 1 it takes 2 bits
the right half of the byte (%----XXXX) is color 1, the left half (%XXXX----) is color 2 - if both are set the result is color 3, eg (%00010001) will set the right hand pixel to color 3

It sounds weird, but just give it a try and see the results - and you'll soon understand it!

Single Bit Operations

You can use NOT AND and OR to do operations on all the bits, but you need to use A - BIT SET AND RES can check, set and reset bits but can be used on other registers without affecting A

Lets give it a go!
Type the program in to the right and run it.

If you remember from before, you'll know IX is used to get parameters.
Switch to Mode 2

Try Call &8050,&40FF

Also give Call &8050,&80F0 a go!

The first part of the parameter must be &80 or &40 - but try other values for the second part!

Lets take a look at the part of the program with new commands! we're going to skip over commands you should already know!

You should understand this now! org &8050
Return if 1 parameter was not passed   cp 1
  ret nz
Load the 'Operation' made up of the first two bytes of the passed parameter (eg XX in &XX--) - remember because of Little Endian these appear at IX+1   ld b,(ix+1)
load the bitmask from the second two bytes of the passed parameter (eg XX in &--XX)   ld c,(ix+0)
You should understand this now!   ld hl,&C000
Again:
  ld a,(hl)
Check bit 7 of B (bit 7 is the far left - so this is like CP &80 or CP %10000000 - but CP would check A, and this works with B)   bit 7,b
If it was zero, jump to our label   jr z,NoAnd
use the AND command with parameter C   AND c
NoAnd:
Check bit 7 of B (bit 7 is the far left - so this is like CP &40 or CP %01000000 - but CP would check A, and this works with B)   bit 6,b
If it was zero, jump to our label   jr z,NoOR
use the OR command with parameter C   OR c
NoOR:
  ld (hl),a
...

Each bit number is a position from Right to Left

Bit Number 7 6 5 4 3 2 1 0
EG: % 1 0 1 0 1 0 1 0

There are 3 types of single bit commands, they can work on almost any register, where as AND %00000001 or OR %00000001 and CP %00000001 only work on A - and AND or OR will change A - these will not

Command SET b,r RES b,r BIT b,r
Meaning Set Bit B in Register R Reset bit B in register R Check if bit B in register R is set
EG SET 7,A RES 6,A BIT 5,A
Equivalent to OR %10000000 AND %10111111 CP %00100000

These commands are great for using 'settings' variables where each bit has a different meaning - In Chibi Akumas the pressed joystick buttons are stored in a single byte and "BIT b,r" is used to test each button.

Things like object movements use different bits in a byte to allow a single byte to define all the possible move directions and types the game needs!

Bit Shifting and Rotating

Sometimes we want to move the bits around in a byte, this can be be to double or halve a value, or to take a couple of bits 'out' of a byte via the carry.

Lets try out the bit shifting commands!

Type in and compile these two examples! there are two versions because one is for shifting Right (&8100) and Left (&8200)
Type Call &8100 and see what happens!

You'll want to see it run continuously, so type in a little basic program to repeat the process

Try the Call &8200 version - and modify the program to Call &8200!
Try replacing "SRL A" with "RR A" or "RRC A" or "SRA A" - see what each does!

Try replacing "SLL A" with "RLA" or "RLC A" or "SLA A" - see what each does!

We'll take a look at them in detail in a second!

Carry Flag

So what do all those commands do? well first we need to understand the Carry Flag!
The Carry Flag is a single bit that stores the 'overflow' from 8 bit maths.

Question: What's 192 Plus 128?
Well in 8 bit maths - it's 64 - with a carry of 1... why? because, 8 bits can only count up to 255, and if there was a 9th bit it would be 1 - and the normal 8 bits would be 64 - confused?  well lets take a look at it in binary

Command Carry 7 6 5 4 3 2 1 0
Ld a,192 0 1 1 0 0 0 0 0 0
add 128 0 1 0 0 0 0 0 0 0
(result) 1 0 1 0 0 0 0 0 0

So the Carry allows us to store the 'overflow' from maths - and allows us to use 8 bit registers for 16 bits, or 16 bit register pairs for 32 bits... but they also allow us to do clever things with bits! Some commands use the carry to shift bits in and out of the register
Now lets take a look at what all those commands do!

Result Carry
Start Value (Keep an eye on the colors to see how the bits move) 10011001 0
RR A Rotates r right with carry - Carry is put at the left, right most bit is put in carry 01001100 1
RRC A Rotates r right with wrap (Carry unused) 11001100 1
SRA A Shifts r right, top bit is the same as previous top bit 11001100 1
SRL A Shifts r right, top bit is set to 0 01001100 1
RL A Rotates r left with carry - Carry is put to the right, left most bit is put in carry 00110010 1
RLC A Rotates r left with wrap  (RLCA is actually faster) 00110011 1
SLA A Shifts r left, bottom bit 0 00110010 1
SLL A Shifts r left, bottom bit 1 00110011 1

Shifting bits around can be used to change values in all kinds of ways -for example shifting left doubles a value, Shifting right halves it.

You can do repeated bit shifts to take a byte apart, and load it into other registers, or reverse it.


Self Modifying code

Lets change the subject a bit!
Remember the 2nd program? - it used a parameter passed by IX to decide what command to do?
The screen fill loop runs over 16,000 times, and using CP and Jump commands in that loop slows things down a lot!

What we really need is a way of comparing the parameter, and changing the action that takes no extra processing power... sounds impossible? well it's not... we make the program change its own code!

Because the program is in memory - and we can change memory, we can swap values or commands in our program whenever we want!

Type in the program to the right and compile it,

If it's too much trouble, remember, you can always just get it from the sources file!
Run the program with Call &8300,&xxyy

Where xx is a command number from 01-03 , and  yy is a bit mask from 00-FF

You'll be able to see the result best in mode 2
Try putting a breakpoint in before the loop (at ld hl,&C000 , for example)

Notice the NOP commands have disappeared  and been replaced with a different command!

(NOP means NO OPERATION - it  does nothing and is just a placeholder)

Lets take a look at the new commands in the program, and see how it works!

Check the user gave us 1 parameter cp 1
ret nz
Load the first part of the parameter (&XX--) into A ld a,(ix+1)
Load the address of our template AND command into HL ld hl,SMAND
if A =1 then jump to the start cp 1
jr z,Start
Load the address of our template OR command into HL ld hl,SMOR
if A =2 then jump to the start cp 2
jr z,Start
Load the address of our template XOR command into HL ld hl,SMXOR
if A =3 then jump to the start cp 3
jr z,Start
Start of the main program Start:
Load the byte of the template command into A ld a,(hl)
Write the byte to the label Selfmodify's address ld (SelfModify),a
Load A from the second part of the parameter (&--XX) ld a,(ix+0)
Write the byte to the one byte after label Selfmodify's address ld (SelfModify+1),a
ld hl,&C000
AgainD:
ld a,(hl)
NOP is a one byte command that does nothing - these two bytes will be replaced by the self modifying code and the command and mask the user's parameter chose will be put here SelfModify:
  nop
  nop
ld (hl),a
inc l
jp nz,AgainD
inc h   
jp nz,AgainD
ret
These three labels have a 'template command' which we pull the correct byte from to modify the code at 'SelfModify' SMAND: AND 1    ;1
SMOR:    OR  1    ;2
SMXOR: XOR 1    ;3

Different commands have different bytes, so when using self modifying code you need to know what bytes the commands compile to!
In this example we've read from a 'template' command at a label - but it would be faster and save memory to just replace ld hl,SMAND with something like ld h,&E6 (&E6 is the AND command in this case)
You'll never remember all the commands, but they're all on the cheat sheet, so just look them up when you need to!

Self modifying code doesn't just allow you to make slow conditions fast, you can save memory with temporary variables.

For example, both these do the same thing:
Normal: Self Modifying:
ld a,(temp)
inc a
ld (temp),a
ret
temp: db 0
ld a,0 :SelfModVar_Plus1
inc a
ld (SelfModVar_Plus1-1),a
ret
Total: 9 Bytes Total: 7 Bytes
See what we did? rather than storing the variable in a separate temp location, we modified the LD command - so now the variable is in the code, saving memory - and this is faster too!


Self modifying code is tricky, so don't worry about using it for now!
One day you may want to make your program as fast and efficient as possible, and Self Modifying code will be waiting to do that for you! Just get used to the normal stuff and remember this as something you need to know about, even if you don't use it!

Lesson 6 -  Lookup table, Screen Co-ordinates, Vector Tables, Basic Parameters Byref
We've covered most of the essential commands now, so we can do something pretty impressive this time!
We're going to create a program that grabs sprites, and prints them on screen - that you can use from your own basic programs!

Enough talk though... Lets make a start!

This lesson's code is quite big, so you may just want to download it from the Sources file. We'll enter it in sections, and look at what each section does, then run it!

Type in the code to the right

You can't compile it yet, there's a lot more work to do!

Copy-Paste the table below into your code - (It's not an image!)
Put it below the code you just entered!

Seriously - you don't want to type all this in!!!
align 2
scr_addr_table:

    defb &00,&00, &00,&08, &00,&10, &00,&18, &00,&20, &00,&28, &00,&30, &00,&38;1
    defb &50,&00, &50,&08, &50,&10, &50,&18, &50,&20, &50,&28, &50,&30, &50,&38;2
    defb &A0,&00, &A0,&08, &A0,&10, &A0,&18, &A0,&20, &A0,&28, &A0,&30, &A0,&38;3
    defb &F0,&00, &F0,&08, &F0,&10, &F0,&18, &F0,&20, &F0,&28, &F0,&30, &F0,&38;4
    defb &40,&01, &40,&09, &40,&11, &40,&19, &40,&21, &40,&29, &40,&31, &40,&39;5
    defb &90,&01, &90,&09, &90,&11, &90,&19, &90,&21, &90,&29, &90,&31, &90,&39;6
    defb &E0,&01, &E0,&09, &E0,&11, &E0,&19, &E0,&21, &E0,&29, &E0,&31, &E0,&39;7
    defb &30,&02, &30,&0A, &30,&12, &30,&1A, &30,&22, &30,&2A, &30,&32, &30,&3A;8
    defb &80,&02, &80,&0A, &80,&12, &80,&1A, &80,&22, &80,&2A, &80,&32, &80,&3A;9
    defb &D0,&02, &D0,&0A, &D0,&12, &D0,&1A, &D0,&22, &D0,&2A, &D0,&32, &D0,&3A;10
    defb &20,&03, &20,&0B, &20,&13, &20,&1B, &20,&23, &20,&2B, &20,&33, &20,&3B;11
    defb &70,&03, &70,&0B, &70,&13, &70,&1B, &70,&23, &70,&2B, &70,&33, &70,&3B;12
    defb &C0,&03, &C0,&0B, &C0,&13, &C0,&1B, &C0,&23, &C0,&2B, &C0,&33, &C0,&3B;13
    defb &10,&04, &10,&0C, &10,&14, &10,&1C, &10,&24, &10,&2C, &10,&34, &10,&3C;14
    defb &60,&04, &60,&0C, &60,&14, &60,&1C, &60,&24, &60,&2C, &60,&34, &60,&3C;15
    defb &B0,&04, &B0,&0C, &B0,&14, &B0,&1C, &B0,&24, &B0,&2C, &B0,&34, &B0,&3C;16
    defb &00,&05, &00,&0D, &00,&15, &00,&1D, &00,&25, &00,&2D, &00,&35, &00,&3D;17
    defb &50,&05, &50,&0D, &50,&15, &50,&1D, &50,&25, &50,&2D, &50,&35, &50,&3D;18
    defb &A0,&05, &A0,&0D, &A0,&15, &A0,&1D, &A0,&25, &A0,&2D, &A0,&35, &A0,&3D;19
    defb &F0,&05, &F0,&0D, &F0,&15, &F0,&1D, &F0,&25, &F0,&2D, &F0,&35, &F0,&3D;20
    defb &40,&06, &40,&0E, &40,&16, &40,&1E, &40,&26, &40,&2E, &40,&36, &40,&3E;21
    defb &90,&06, &90,&0E, &90,&16, &90,&1E, &90,&26, &90,&2E, &90,&36, &90,&3E;22
    defb &E0,&06, &E0,&0E, &E0,&16, &E0,&1E, &E0,&26, &E0,&2E, &E0,&36, &E0,&3E;23
    defb &30,&07, &30,&0F, &30,&17, &30,&1F, &30,&27, &30,&2F, &30,&37, &30,&3F;24
    defb &80,&07, &80,&0F, &80,&17, &80,&1F, &80,&27, &80,&2F, &80,&37, &80,&3F;25


If you want to use this program on the Spectrum or Enterprise, you'll need a different Look Up Table and Get Next Line routine!... and MSX graphics are totally different, so you need to draw in a different way.

Don't worry, we'll get to it once we've learned all the Z80 commands - we're nearly done now!

Reading from a Lookup Table


We can use a lookup table to read data from - in this case, our screen location contains 2 bytes, so we add the 'index' (ypos) twice to the start address, then rad in two byte.

The code you've entered coverts X,Y co-ordinates to screen locations - and calculates the position one pixel line down from the current memory location! this will be used to work out where our sprite will be drawn/read from, and to work through the sprite line by line

Aligned code

Aligned code is code that starts from a certain byte boundary - it is used for speeding things up, or look ups - by knowing where the data will start, we can save time by using INC L rather than INC HL - as we can know H will not need to change... the ALIGN command allows us to make assumptions about the position of the following code, without being as rigid as an ORG command

Lets take a look at  what the ALIGN command does
In this example we have a few bytes of 1's defined with DB 1,1,1
The ALIGN xx command will align to the next xx boundary... in this case, it aligns to a 16 byte boundary
The ALIGN 16 command inserts zeros as required.
The DB 2 inserts a 2 - note it's aligned correctly!

Aligned code allows you to do all kinds of clever things!
A common trick is to define a 256 byte aligned table - with the "Mask" to erase the background when pixels are color 0 for every possible byte a sprite could contain - 
By setting H to the start of the table, and L to the byte - the mask can be applied by LD A,(HL)  AND A
It sounds confusing, but you'll soon think of lots of clever things you can use Lookup tables and Aligned code for!

We can't run the code yet, because we're missing the sprite grabber - but lets take a look how it works!

This is called a Jump block, it's allows us to jump to an unknown location from a known one.

Each JP xxxx command is 3 bytes, so we know "JP Get MemPos" is at &8006 - even though we don't know where "GetMemPos" is, we can use Call &8006 and it will have the same effect as Call GetMemPos - this allows you to write programs that can easily be called from basic, or other programs that were compiled separately!
org &8000
    jp GetSprite
    jp PutSprite
    jp GetMemPos
This command takes an X-pos in B , and a Y-pos in C, and converts them to a screen memory location in HL
note X is measured in BYTES - so there are 80 of them across the width of the  the CPC screen - Y is measured in LINES
GetScreenPos:
    push bc
Load the address of our LookupTable for screen line memory  locations
We don't need the Xpos for now - so we set B to 0
Each memory location is 16 bit (2 bytes) so we add B twice to HL, this means HL points to data containing the 16 bit Memory location of the line we want - we just need to get at it!
        ld b,0
        ld hl,scr_addr_table
        add hl,bc
        add hl,bc
We need to load HL with the data at the memory location IN HL - first we load the low byte into A
We use the faster INC L - not the slower  INC HL  because we used Align 2 - this means the data won't go over a 255 byte boundary (we know we're not on &80FF)  - align will have added 0 bytes to ensure we're not.
Then we load H from (HL) -  don't worry! H doesn't get erased until AFTER (HL) has given us the byte we need!
Then we load L from A - which we got the line before - and we're done!
        ld a,(hl)   
        inc l
        ld h,(hl)
        ld l,a
Get back BC - we want B (the xpos) to be in the low byte - and we need to set the high byte to the position of our memory buffer (&C0 - because the screen is at &C000)
We add it to HL

HL now contains the screen memory position of the X,Y pos we wanted!
    pop bc
    ld c,b
    ld b,&C0
    add hl,bc
ret
Finding the line below the current one is hard on most systems, On the CPC line tend to be &0800 below the last one, so we add &08 to H - which is faster than messing with HL GetNextLine:
    ld a,h
    add &08
    ld h,a
when we go over &FFFF we need to start back again at the top - annoying isn't it?
if the top bit is zero then we've rolled back to &0000 - but otherwise we're done
    bit 7,h
    ret nz   
If we got here, then we rolled over, but we can fix things by adding &C050 - to get the correct position of the line!     ld bc,&c050 
    add hl,bc
    ret

The next part of the program is the sprite grabber!

Type in the program to the right!

You still won't be able to compile it without the 3rd part!

This routine grabs a sprite from the screen with "&8000,MEMDEST,X,Y,W,H" - and returns the next free MEMDEST with "Call &8006,@int"

Undocumented registers IXH, IXL,IYH,IYL 

IX and IY are 16 bit register pairs like HL, they're actually made up of 2 registers we can use for whatever we want!
IX is made up of IXH (IX-High) and IXL (IX-Low)
IY is made up of IYH (IY-High) and IYL (IY-Low)

They aren't as fast as other registers - but there are times they are useful - in this example we use them as loop counters!

These registers are undocumented - That means they weren't originally in the Z80 manual - but all Z80's support them, they even work on the MSX Turbo-R's R800 - so don't be afraid to use then!

Back in the 80's people didn't know about them, so we can use them to give our programs a speed advantage over the old games!


Basic References and 16- bit integers with @ and %

In CPC basic, putting @ before a variable will pass it as a 'reference' - this means we get the address of int, not the just value - and our program can change the value that the basic variable has! This allows us to store the memory location of sprites in integers in basic!

But make sure you specify they are integers by putting % at the end of the variable - otherwise they will be a floating point number,  eg s1%=0 will define s1% as an integer.

You can pass other data types, but they're harder to use, as you need to know how data such as strings and floating point numbers are stored on the CPC

16-bit integers are easiest to work with, so make sure you put the % symbol at the end of the variable name!


Lets take a look at how it works!

Start of the sprite grabber - Check we were given 5 parameters GetSprite:
    cp 5
    ret nz
Parameters come in backwards - and are always 16 bit... so MEMDEST  is at byte position 8 & 9  - this is where we will store the sprite     ld e,(ix+8)
    ld d,(ix+9)
X is at position 6, Y is at position 4 - we only expect 8 bits, so we ignore positions 7 and 5 - These are the screen position to grab from     ld b,(ix+6)
    ld c,(ix+4)
Load the Width from position 2 - store it in IYH, and also in the destination memory
Increase the destination memory pointer
    ld a,(ix+2)
    ld iyh,a
    ld (de),a
    inc de   
Load the width from Position 0 - store it in IYL, and also in the destination memory     ld a,(ix+0)
    ld iyl,a
    ld (de),a
    inc de
Convert B,C to a screen memory pos     Call GetScreenPos
Back up the screen memory co-ordinate
Set BC to the number of bytes we want to copy
use LDIR to copy from the screen to DE
RepeatY:
    push hl
        ld b,0
        ld c,iyh
        ldir
Get back the screen memory pos, and move down one line     pop hl
    call GetNextLine
Decrease the line counter, and repeat if we're not at zero     dec iyl
    jp nz,RepeatY
Store the next empty memory position using self-modifying code     ld (LastSpritePos_Plus2-2),de
ret
Check we were given one parameter GetMemPos:
    cp 1
    ret nz   
Load the reference into HL     ld l,(ix+0)
    ld h,(ix+1)
Load DE with a value - note we're using self-modifying code - so the value will be the end of the last sprite     ld de,&0000    :LastSpritePos_Plus2
Write the 16 bit value in DE to the memory position that HL contains     ld (hl),e
    inc hl
    ld (hl),d
ret


EX DE,HL - When HL just won't do!

HL has more power than anything else, but sometimes we want to use DE for the same job - well, we can't! but we can swap HL and DE quickly to do what we need - it's faster than any kind of PUSH POP options if you just need to swap the two - in this case it allows us to use LDIR to read from HL, and write to DE!

One last bit to go! The PutSprite routine... don't give up, the finish line is in sight!

Type the code to the right!
Compile it, you should get no errors!

Now we can Grab sprites, and Print them to screen!
Here's a little Basic program to try it out!

Type in the program to the right.

It will define 3 sprites,
one of 9 numbers (Stored in S1%)
one of 9 letters  (Stored in S2%)
one is blank (stored in S0%)

Run the program, it will show one of the sprites on screen!

Lets see how our basic program uses the commands

Reset the screen and print some letters to use as sprites 10 MODE 1
20 PRINT"aaa111":PRINT"bbb222":PRINT"ccc333"
Define 3 variables to store the locations of our sprites - the % symbol defines them as 16-bit integers
Our sprites will start at &9000 - we'll work out the locations of the other two later
30 s1%=&9000
40 s2%=0
45 s0%=0
Grab (0,0)-(6,24) and store it at memory location in S1% 50 CALL &8000,s1%,0,0,6,24
Pass S2% by reference (the @ symbol denotes this) - our ASM program will put the next free sprite position into basic variable S2% - so we know where this sprite will start! 60 CALL &8006,@s2%
Grab (6,0)-(12,24) and store it at memory location in S2% 70 CALL &8000,s2%,6,0,6,24 
Get the Next sprite memory location, and store it in S0% 71 CALL &8006,@s0%
Grab (12,0)-(18,24) and store it in position S0% 75 CALL &8000,s0%,12,0,6,24
Show sprite S1% at screen position (50,50) 80 CALL &8003,s1%,50,50

Now you can make and use your own sprites from Basic!

Need more inspiration? try adding the basic code to the right!
This will show a sprite on screen, and erase it - you can move the sprite with keys ZX K and M

Don't let the sprite go off screen though! there's no checking and it may crash the CPC

The basic program is on the Sources.DSK image!

INKEY$ reads a character from the keyboard, it's a quick easy way to read input!
In this example we use it to move a sprite around the screen!

I'm sure you can think of more interesting things to do with the code in todays lesson!

Vector Tables - Lookup tables for Jumps!

We can use the same code we just used in our look-up table for calling subroutines - we can take a 8 bit integer, and use it to decide on a Jump.

"CP X    JP Z,YYYY " uses 5 bytes per command!
A Jump block uses 3 bytes for each command!
A vector table uses only 2 bytes per command!

If memory is tight, a Vector table is the best option... and a jumpblock does not get slower when working with a lot of entries - where as repeated "CP x.. JP Z,yyyy" commands will waste CPU power for each command skipped over to find the matching label

DW - defining 2 byte 16 bit 'Words'

We use DW xxxx to define a 16 bit word - we could do the same with 2 DB commands - but that would make no sense in this case

Let's take a look at a vector table!
Add another jump to the Jump block below JP GetMemPos
Type in the program to the right and compile it

You can run it with Call &8009,x  - where x is 0-2

This will call one of the commands in the VectorTable
if we ALIGN the vector table so all the commands are in the same byte boundary we can make the Vector table lookup simpler!

Lets look at how the Vector Jump works!

Check we got 1 parameter, and load it into A VectorTableTest:
    cp 1
    ret nz
    ld a,(ix+0)
Rotate the bits in A left, effectively doubling it - we do this because each entry in our vector table is 2 bytes     rlca
Load our vector table into HL
Set BC to the byte offset of the command we chose
ADD BC to HL - HL now points to our command
    ld hl, VectorTable
    ld b,0
    ld c,a
    add hl,bc
load HL with the address of a command from the VectorTable     ld a,(hl)
    inc hl
    ld h,(hl)
    ld l,a
Jump to the address in HL jp (hl)
The vector table - each line is a 16 bit address of a command - the first will be called if the parameter is 0 VectorTable:
defw TestA
defw TestB
defw TestC

Lesson 7 - DI EI, RST x, Custom Interrupts, IM1/IM2, HALT, OTI / OTIR, HALT
We're going to learn how to interface with the hardware, and take over the last job of the firmware - so we're going to have to cover a lot of technical content now, so we understand what we're doing!

Both examples are included in the Sources download, just enable,or comment out the "Eg2 equ 1" declaration to toggle between the simple example - and the final one

RSTs and Interrupts in IM1

CALL commands take 3 bytes, but there is a special way the Z80 can do a limited call in just 1 byte! these call to an address in the first 64 bytes of the address space!

These are called with the commands RST 1 to RST 7 ... and they each call a different address - depending on the platform they may be in use, or have a special function. of course, if we don't need the firmware, then all bets are off, and we can do whatever we want with them!

Unfortunately because systems like the Spectrum and TI-83 have Read only Memory in the address range &0000-&3FFF - so we can't write our own RST's
The MSX Rom has predefined RST's, but if we page in RAM, then there are none defined

RST's are only really useful when you need to do the same call or a few commands a lot!

Sometimes you'll see RSTx followed by a byte or two... (EG EXOS - rst6 on the MSX)  The RST function will be looking at the stack, and reading in the 'calling address' to get the location of these bytes as parameters - then modifying the return address to skip over them!


Lets take a look at the RST's on each system, with the firmware use of them
Command Equivalent Call
RST 0 Call &0000 Reset ROM - Reset Rom:? Ram:free reserved for CPM Rom:? Reset
RST 1 Call &0008 Low Jump ROM - Unused Rom:? Ram:free free Rom:? Error Handler
RST 2 Call &0010 Side Call ROM - Print Char Rom:? Ram:free free Rom:? Print char A
RST 3 Call &0018 Far Call ROM - Get Char Rom:? Ram:free free Rom:? get Basic Char
RST 4 Call &0020 Ram Lam ROM - Get Next Char Rom:? Ram:free free Rom:? get Basic Front char
RST 5 Call &0028 Firm Jump ROM - Call Rom1 Routine Rom:? Ram:free free Rom:Bcall Floating point Calc
RST 6 Call &0030 free for user use ROM - Unused Rom:? Ram:free EXOS call Rom:? Usr RST (RST30V)
RST 7 Call &0038 IM1 Interrupt Handler IM1 Interrupt Handler IM1 Interrupt Handler IM1 Interrupt Handler IM1 Interrupt Handler IM1 Interrupt Handler

The one we're really interested in here is RST7 at &0038 - because this one is called automatically by the system hardware when IM1 is enabled.
There will be times we need to do an event with controlled frequency, for example playing music, and to do that we use the 'Interrupts' of the system
Interrupts are when the hardware forces the Z80 to process tasks for it - and up until now we've just let the firmware do them itself, but now we're going to look at how to take them over ourselves

the commands IM0 - IM2 change the interrupt mode of the z80

IM1 is the most useful to us, IM0 is pretty irrelevant to us as programmers, but on the spectrum we can't write our own IM1 interrupt handler, as the rom stops us writing to &0038, but we can use IM2 - but it's more complex... so we'll look at that later!

Not only does writing our own interrupt handler and do 'timed' actions, but if we take over this task from the firmware, then we can be in control of all the Z80 resources - and this includes the 'shadow registers' that' we've not seen until now!

Shadow Registers

Now we've taken total control of the Z80 from the firmware - and we can use all it's registers without worrying about the effect on the firmware (unless we return to basic!)
This gives us access to the Shadow registers - the Z80's 'spare' set of the registers AF BC DE and HL -  these are switched in by the firmware so that the main registers are not altered and the program can resume once the interrupt is done!
  Normal Registers Shadow Registers
Accumulator A
A'
Flags F F'
HighLow Memory Location H L H L
ByteCount B C B C
DEstinaton D E D E
IX IXH IXL

IY IYH IYL


You'll notice that there are no shadow versions if IX or IY

The shadow registers cannot be used in combination their normal counterparts, instead we 'toggle' the shadow versions in or out using two special commands which 'swap' the normal and shadow versions. 
Of course we could just use PUSH and POP to get the same effect - but swapping the shadow registers is faster - which is why they exist, to allow the Interrupt handler it's own registers to quickly use without affecting normal ones!


Command   Effect
EX AF,AF' Swap A and F with the shadow versions
EXX swap BC,DE and HL with the shadow versions

So we can swap just AF... or all the other main registers... because we can's swap just some of the 16 bit registers, if we just need to use one pair such as  DE or HL for a while, PUSH and POP are better...
EXX only works when we need to use ALL the registers...  so times when a separate job needs to be done for a while before carrying on with the main job may be the usage case of shadow registers.
For example In Chibiakumas this is used in the bullet loops, where the main registers are used for calculating bullet position, then the shadow registers are used for player collision detection, before switching back the main registers to continue the loop.

Interrupts

On the systems we're looking at interrupts only occur based on the screen refresh, but the frequency varies depending on the system, some systems need you to 'tell' the hardware the interrupt it's been processed otherwise it'll keep happening immediately again! - lets look at all the Interrupt facts by system so we know what we're dealing with for the systems we're interested in

System Interrupt frequency HZ If interrupt is missed How to clear the interrupt Firmware / Interrupt  register usage
6 times per screen draw
(CPC+ can have line interrupts)
300hz Interrupt happens
as soon as possible
no need All shadow registers used

Shadow register BC' must not be changed or firmware will crash
Once per screen draw 50hz Interrupt never occurs no need Shadow registers are not affected
IY must be preserved for firmware to work properly
(IY Should be &5C3A?)
Once per screen draw 50hz Interrupt happens
as soon as possible
in a,(&99)
(when chosen VDP status
reg is
set to default of 0)
Shadow registers are not affected

No registers need to be preserved
Once per screen draw 50hz Interrupt happens
as soon as possible
ld   a,30h
out  (0b4h),a
Shadow registers are not affected

No registers need to be preserved
Various such as line based ? ? ? Shadow registers are not affected

No registers need to be preserved

HALT, DI, EI

The Interrupts will occur automatically at the appropriate time, but we can wait for one to occur, by using the HALT command... this will cause the Z80 to wait for an interrupt to occur.

There may be times we cannot allow interrupts to occur -   We can stop interrupts from being allowed for a while by using DI - and allow them again using EI...
On most systems if an interrupt is missed it will occur immediately, but on the Spectrum it never happens - see the table above for details.

There are many times you may need to stop interrupts: if you're interrupt handler uses the shadow registers and you need them... if you're communicating with a series of OUTS that interrupts would conflict with (Keyboard and sound chip are on the same CPC ports)...
Disabling Interrupts reduces the complexity of debugging too, so if you don't need them, you may as well disable them.
Finally there's one special reason why interrupts can't run... If we've altered the stack pointer then the RST7 call cannot occur... we'll learn why you may decide to do that next time!

Communicating with hardware - OUT and IN

Interrupts are the only time the hardware takes over the CPU - but there are times we need to give instructions to the hardware, or receive data from the hardware... and we use the commands OUT and IN to send data to the hardware.

If you think of the computer hardware as a telephone system, and each piece of hardware has a phone numbers from 0-255 (called a port)

OUT (xx),yy  will call the hardware at number xx - and give it message yy

IN (xx),yy will call the hardware at number xx - and receive a message which will be stored in yy

Hardware you may access will be things like screen hardware, memory bank switchers, disk systems,joysticks and keyboards - the ports that you'll want to use, and how to use them varies depending on the system.. 

Now! there's a catch!  The Z80 in the  ZX spectrum, Sam Coupe and the CPC are wired oddly... they use 16 bit ports and use BC as the port number - even though the assembly command is Out (C) the command in the assembly code is not the one that effectively occurs when the Z80 runs it....

Even worse, some devices are listening to many ports, so OUT ing to &7F00-&7FFF on the CPC should have the same effect - usually one port is the 'recommended' one, and others just 'may' work (I had color changing code that sometimes worked fine, and other times corrupted disks) 

Let's make this command confusion clear, or this will waste a load of your time!

System Command in ASM code Command the system actually runs
Out (C),A
Out (C),C
Out (n),A
OTIR
Out (BC),A
Out (BC),C
*** Does not work ***
*** Malfunctions ***
Out (C),A
Out (C),C
Out (n),A
OTIR
Out (C),A
Out (C),C
Out (n),A
OTIR

The Z80 has some 'incremental' commands for OUT like OTIR that are functionally similar to LDIR ... because these commands alter B, and B is used as the address of the OUT command these commands will malfunction...

On the Amstrad CPC we can use port &7F to control the screen colors - and because the CPC interrupt occurs 6 times a screen, by changing the color every interrupt, we can get more colors on screen than normally possible - This is how Chibiakumas gets around 8-10 colors on the 4 color mode 1 screen!
We're going to do this ourselves, in the next example

Outs are confusing and hard work! but you won't want to write them very often!
In practice what you'll do is write a function to control the hardware, then you'll forget about how it works and just use that function, so just find a good example online of what you need to do, and just stick to the safe code they use!

Well, that was a lot of theory! but it all fits together into a really neat example!

Type in the program to the right.

Compile it, and run it by using Call &8000

Note, this example does not return to basic, so you will need to reset your emulator
This program will change the background color each interrupt.

You will see the background color change six times during the screen refresh!
(the first one is offscreen!)

How does it work? well lets take a look at the code!

org &8000
Turn off interrupts - We're going to mess with the interrupt handler, and need to make sure an interrupt doesn't occur while we do     di
set &0038 to &C3  &0038 (RST7) is the address called when an interrupt occurs, and &C3 is the bytecode of a JP command         ld a,&C3
        ld (&0038),a
After the jump, we put the address of our code that will handle the interrupt         ld hl,InterruptHandler   
        ld (&0039),hl
Now we've finished setting up our interrupt handler, turn interrupts on!     ei
The HALT command waits for an interrupt, but this loop is just an endless loop.... we can't return to basic, because our interrupt handler will break basic! InfLoop:
    halt
jp InfLoop
When an interrupt occurs the Z80 will end up here... note interrupts are automatically disabled when &0038 is called InterruptHandler:
swap HL,DE and BC with the shadow versions - we can now use these in our interrupt handler without worrying     exx
swap AF with the shadow version     ex af,af'
Load the address of our Rastercolors list - we're going to selfmodify this, so we have a label at the end ld hl,RasterColors :IH_RasterColor_Plus2
We want to check if the screen has finished redrawing - we can do this by accessing the 'PPI' port B... rather strangely this is connected to ports &F500 to &F5FF so we do not need to set C for this     ld      b,&f5
read A from port (BC) (CPC/ZX are 16 bit so use (BC) ... MSX/ENT are 8 bit, so use (C) )     in      a,(c)
Bit 0 is marks 'Vsync' so push the rightmost bit into the carry
if it's zero, just carry on
    rra
    jp nc,InterruptHandlerOk
if we got here, then we're vsyncing, so reset HL to the start of the array     ld hl,RasterColors
Set B to &7F this is the Gate array address
set C to &00 - when we do OUT (BC),C this tells the gate array 'I want to change Color 0'
InterruptHandlerOk:
    ld bc,&7f00     
    out (c),c
Read A from the HL
Effectively performs - Out (BC),A - sends a palette color to the Gate Array to set Color 0 to
    ld a,(hl)
    out (c),a
Increase HL and remember the next address we want to read from using selfmodifying code     inc hl
    ld (IH_RasterColor_Plus2-2),hl
Restore the normal registers, and turn interrupts back on

Return to whatever was happening when
    ex af,af'
    exx
    ei
ret
The 6 colors we will use for the screen - note these are not the usual 'Basic colors' you'll be used to - these are Hardware Colors! RasterColors:
    db &4C,&43,&52,&5C,&5E,&5F
We need to take more care if we need to keep basic and the firmware happy!
You'll need to do return to basic, or use firmware calls for screen operations - or disk access!
On the CPC we need to keep the interrupt handler we replaced, and back up shadow register BC' - the the firmware will be OK

Now lets take a look at how to change all the colors, and this time we'll do some screen changes in between interrupts - and we'll back up everything so we can return to basic!

Type in the example to the right

Compile it, and run with Call &8000

This example will flip the screen colors 10 times, then return to basic
This version will change all 4 screen colors, at all 6 raster points - and returns to basic as if nothing had ever happened!
We can only easily change colors at these 5 points...
Drawing each line of the screen takes 64 ticks/cycles... if you want to change color each line,
or on a specific line, you need to delay the processor with enough NOP's to get to the line you actually want!
So how does this one work? lets take a look!

Turn off interrupts while we're messing with the interrupt handler org &8000
    di
Back up BC' from the Shadow registers - The CPC firmware relies on this being intact
While the CPC firmware uses HL DE and AF in the shadow registers - it doesn't mind us changing them so we don't back those up
        exx
        push bc
        exx
Back up the first 4 bytes of the current interrupt handler so we can remove our changes later         ld hl,(&0038)
        push hl   
        ld hl,(&003A)   
        push hl   
Set up our custom interrupt handler and turn on interrupts         ld a,&C3
        ld (&0038),a
        ld hl,InterruptHandler   
        ld (&0039),hl
ei
Set our pointer HL to the start of the screen
We use D as a loop counter - starting at 10
set C to %11111111 - we use this as our XOR mask
    ld hl,&C000
    ld d,10
    ld c,%11111111
XOR each byte in the screen with the mask in C  - when we finish the screen we jump to PageDone InfLoop:
    ld a,(hl)   
    xor c
    ld (hl),a
    inc hl
    ld a,h
    or a
    jp z,PageDone
jp InfLoop
Reset HL to the &C000 (the start of the screen)
Decrease D and repeat until it's zero
PageDone:
    ld hl,&C000
    dec d
jp nz,InfLoop
Turn off interrupts     di
Restore the Firmware default interrupt handler that we backed up before         pop hl
        ld (&003A),hl
        pop hl   
        ld (&0038),hl
Restore Shadow register BC'         exx   
        pop bc
        exx
Turn Interrupts back on and return!
Basic will never know what happened!!!
    ei
ret
Start of the interrupt handler
Switch to the shadow registers
InterruptHandler:
    exx
    ex af,af'
The same as last time!
Set HL to color array (this will be self-modified)

see if we're at the top of the screen, and if we are, reset HL
ld hl,RasterColors :IH_RasterColor_Plus2
    ld      b,&f5
    in      a,(c)
    rra  
    jp nc,InterruptHandlerOk
    ld hl,RasterColors
Tell the Gate Array at &7Fxx that we want to set color 0 InterruptHandlerOk:
    ld bc,&7f00     
    out (c),c
OUTI will copy one byte from HL, and INC HL
unfortunately it also decreases loop counter B - which will cause problems on the Amstrad - so we do INC B to fix it... even with the 'INC BC'... it's still faster than the separate commands!
    outi
    inc b
Increase C, and do another OUT - this tells the Gate array we want to do the same to Color 1     inc c
    out (c),c
Do the same for Color 1,2 and 3

A loop counter and a jump would take a bit of time - and this is going to run 300 times a second! so actually have the same commands 4 time rather than using a loop to save some speed! ... this is called 'unwrapping a loop'
    outi   
    inc b
    inc c   
    out (c),c
    outi
    inc b
    inc c
    out (c),c
    outi
Remember the next address we want to read from.
Swap the shadow and normal registers back,
Enable interrupts and return.
ld (IH_RasterColor_Plus2-2),hl
    ex af,af'
    exx
    ei
ret

We're totally kicking ass now!
We've now learned how to communicate with the hardware ports, we've made our own interrupt handler, and we've learned how to use the 'Shadow Registers'...
We've now learned almost all the Z80 commands! There's not much left to learn before we know everything!

Lesson 8 - Unwrapped Loops, Stack Misuse for speed & rarer Z80 commands
We've covered the all the most important commands, but there are others you may wish to use time to time...

Also there's some clever tricks you'll want to know, that are often used for graphics to make sprite and fill routines as fast as they can be!

Lets take a look at the remaining commands the Z80 has that you may wish to know!

The last two Registers!

There are two registers left that we've not looked at, they have limits, but they may be useful in some cases!

The first one is R... this is used by the system to 'refresh' the memory to keep the data ok in ram... now you should not change this register, because it could cause damage to the memory, buy you can read from it fine - it will have a value in it from 0-127... and as it constantly changes it may be useful as a random number seed!

The other is I... this is used by IM2 - but unless you're using the ZX Spectrum you'll be using IM1, and it does nothing in IM1 - so I is free for you to do whatever you want...
Unfortunately there are only 2 commands, one to load A into I, and another to load I into A ... but it's faster than PUSH and POP - so if you want, you can use it for temporary storage!

The I register isn't a lot of use as a general register, and you'll need it for interrupts if you end up using the Spectrum.

In Chibiakumas, The I register is used on systems except the spectrum... an assembler 'Macro' replaces LD I,A with a ld (&xxxx),a command on the spectrum - it takes a little extra memory, but means the spectrum version works without the I register for storage.


DAA and Binary Coded Decimal

Converting 16 bit bytes for screen display is hard, and sometimes you may wish to just use 'Binary Coded Decimal'... this is where each byte only stores a number from 0-10... this is how ChibiAkumas stores the player score, there are 10 bytes for each of the 10 digits, this makes showing the score quick and easy.

DAA is a strange command I've never had need to use, it's very complex, and you should look at the Zilog manual if you want to really know about it... the main people actually use it is nothing to do with Binary Coded Decimal, but in fact to convert a byte with a value of 0-15 to a hex char using the sample below
This code will show A onscreen as Hex where A is below 16 daa
add a,&F0
adc a,&40
call PrintChar

there are two other even stranger commands! RRD and RLD, I've never used them, and I can't think why you would, but here they are!
RLD Rotate Left 4 bit nibble at Destination (HL) using bits 0-3 of A as a carry
RRD Rotate Right 4 bit nibble at Destination (HL) using bits 0-3 of A as a carry


Carry Flag
There are some commands which do maths which will also use the Carry flag, this allows mathematical operations to include overflow from previous calculations. There are 8 bit and 16 bit commands, lets take a look at them, Rather strangely the only 16 bit subtract command is with carry - there is no SUB HL,DE!!!

SBC # SBC 4 Subtract a (4+Carry) from A
SBC r SBC B Subtract (B+Carry) from A
SBC rr,rr SBC HL,DE subtract (DE+Carry) from HL
ADC # ADC 4 add (4+Carry) to A
ADC r ADC B add (B+Carry) to A
ADC rr,rr ADC HL,DE add (DE+Carry) to HL

As mentioned there is no way to subtract without the carry, so there may be times we need to set or clear the carry flag... another time will be if we want to use the Carry as a status flag - eg some disk routines set the carry flag if reading worked, and clear it if it failed.

There are two commands
SCF Set the Carry flag (to 1)
CCF Complement Carry Flag - inverts the carry flag
I bet you thought CCF would clear the carry flag, well it doesn't! but there is an easy way to clear the carry flag! just use
OR A reset the carry flag

Alternatively, you can just convert a positive 16 bit number into a negative one, with the following code!
Convert DE into a negative number, then add it to HL

This will have the effect of subtracting DE from HL
    ld a,d
    cpl
    ld d,a
    ld a,e
    cpl
    ld e,a
    inc de
    add hl,de

Return from interrupts with RETI and RETN
There are two 'Special' interrupt return commands RETI and RETN... to my knowledge there is no benefit to using these and they have no practice use on any system I know.

Bulk copy and search
We looked before at LDIR, but there is an alternative, LDDR - this version of the command is for use when HL is pointing to the END of the range you want to copy
RETI Return from an interrupt
RETN Return from nonmaskable interrupt

Rather than copying, there are two 'search' commands, which will scan BC bytes from HL to find byte A ... I have never used them, but maybe I'm missing something!
Lets summarize all those commands
LDDR same as LDIR, but HL goes down not up
CPI Search from HL, increasing HL for BC bytes - try to find A, PO set if found
CPD Search from HL, decreasing HL for BC bytes - try to find A, PO set if found

Special OUT and IN commands you'll probably never need!
There are a few special bulk IN and OUT commands, however as before, because BC is used as the port address on Amstrad and Spectrum, they may not work, and I doubt you'll ever need them, but here they are!

IND In to HL and decrease HL
INDR In to HL and decrease HL and repeat
INI In to hl and Increase HL
INIR In to hl and Increase HL repeat
OUTD Out from hl and Decrease
OTDR Out from hl and Decrease and repeat

The INI command may be useful for reading from something like the Keyboard into a buffer... just remember that it decreases  B each time you use it, and the Amstrad and Spectrum won't like that one bit!
if you use it in a loop, you'll need to reset B, or do an INI, then an INC B and everything should be fine!

Stack Specials!
The stack has a couple of special commands, that may help if you're being tricky, some will use in a moment, but there's a couple that you should know
INC SP increase the stack pointer - do this twice instead of a POP (it's actually slower - but if you don't want to alter your registers)
DEC SP decrease the stack pointer
EX (SP),HL Swap the contents of HL with the item at the top of the stack

Some commands only work with a certain selection of registers, so check the Cheatsheet to see the full range of 'options' available!

Now, lets move on to the lesson!

Unwrapped Loops!

There are times when we need as much speed as possible, especially when we're drawing to screen... When we loop around a command like LDIR a lot of time is wasted with compare and jump commands, If we have enough memory to spare - we can jump less, and copy more data - lets take a look!

First lets create a 'slow' version with a regular LDIR command - this program will copy the whole screen TO and FROM a 128k memory bank

Type the program in to the right.

It gives you 2 commands

Call &8000,xx - will copy the screen TO bank xx
Call &8003,xx - will restore the screen FROM  bank xx

xx should be a CPC bank number from C4-C7 or C0 ... other banks may work if you know how CPC banks work!!!

Note, This program will not work right on a CPC464 - as these banks only exist on 128k systems, all the commands will copy to the same bank!

Most of this should be pretty clear to you by now, but lets have a look at that Bankswitch Command!
This Bankswitch command will send a Bank number to the Gate array

The gate array is at &7Fxx...
Bankswitch:
    LD B,&7F
    OUT (C),A
    ret

on the CPC there are 8 possible bank configurations, &C0 is the default, where the main 64k are in memory....  the other options are shown below:


C0 C1 C2 C3 C4 C5 C6 C7
0000-3FFF RAM_0 RAM_0 RAM_4 RAM_0 RAM_0 RAM_0 RAM_0 RAM_0
4000-7FFF RAM_1 RAM_1 RAM_5 RAM_3 RAM_4 RAM_5 RAM_6 RAM_7
8000-BFFF RAM_2 RAM_2 RAM_6 RAM_2 RAM_2 RAM_2 RAM_2 RAM_2
C000-FFFF RAM_3 RAM_7 RAM_7 RAM_7 RAM_3 RAM_3 RAM_3 RAM_3
If you have more than 128k, you will have options C8,C9 and so on... up to FF if you have 512k!

Note, you cannot use &C2... this swaps out &8000 - which is where your program is, so it would cause a crash!

Speed Up!

Now let's unwrap the loop... We're going to make a few changes, and make it faster!
Lets  try this faster version we use 16 LDI commands - by skipping the R (repeat check) for most of the commands we can increase speed by around 25%!
You only need to make a few modifications to your code for this faster version!

Replace both the LDIR commands with "Call UseLDI"


Add  this function to the bottom of your code!

Try the new commands, you should see they're noticeably faster!

We've sped things up but of course, the code is bigger! We've traded memory for speed!

Was it worth it? well that's depends what you can spare... This often the choice you're faced with when programming in 8 bits!


There's not much to look at here, the only thing to note is that LDI, or LDIR set the PO flag to true when BC=0... Note, BC must be a multiple of 16, or this program will miss the 'return' and overwrite all the memory!

Stack Misuse!

There may be times when we need to read or write data super quickly... and there's a clever trick for this... as mentioned before, the Stack pointer offers the fastest reading and writing the Z80 can offer.

As we know the Stack pointer is designed for PUSH and POP, and for CALL's, but if we don't need to do any of these things, we can temporarily alter the stack pointer, and use it to bulk read, or bulk write a block of data super quick!... of course, Interrupts are a call to &0038 - so we'll have to disable interrupts to be safe! (Depending on what you're doing, you MAY be able to allow interrupts - eg if you're only writing... a few bytes would be temporarily corrupted by the interrupt handler's calls and pushes)

Add "JP ClearScreen" to your jumpblock
Add  the code to the right to the bottom of your program

There are 32 PUSH DE's in total... so get Copy-Pasting!!

We've done some tricky stuff, so lets take a look!

ClearScreen:
Turn off interrupts... We're going to mess with the stack, and an Interrupt Call would use the stack, so lets stop them for safety     di
Back up the stack pointer using self modifying code     ld (SP_Restore_Plus2-2),sp
We're going to overwrite &C000-&FFFF
Load the stackpointer as &FFFF+1 (&0000)... each push decreases the stackpointer by two before pushing the data, so we set the start point to the first byte to write+1
    ld sp,&0000
We're going to push DE, so this is the byte pair we're going to fill the screen with     ld de,&0000
We're going to push 64 bytes each time, so we need to do this 256 times... setting B to 0 means the loop will occur 256 times before B reaches zero again     ld b,0
64 2 byte pushes     push de       
...
    push de
    push de
Loop until B becomes zero djnz ClearScreenAgain
Restore the correct stack pointer that we stored before     ld sp,&0000    :SP_Restore_Plus2
Turn Interrupts back on and return     ei
ret


ChibiAkumas uses Stack in this way to flood fill the background gradient super quick... it fills the screen with a parallax gradient faster than the firmware CLS command!
It also uses POP to quickly read sprite data for transparent sprites.
In fact when a game does things super fast, it's likely PUSH or POP is being misused somewhere!
It's tricky though, so don't worry about it at first, get your game working with the normal stuff,then if you want you can add stack misuse later to up the frame rate!

Congratulations! You've reached the end of this series of lessons...
We've covered all the Z80 commands you're likely to ever need, and learned lots of great stuff!
But stick around! We've moving on to bigger and better things with new series' of lessons!

This ends the Basic series.. but the lessons are just beginning!... Tune in next week for more Z80 fun!


Appendix
Command Alternate form Meaning Example Notes
ADC r ADC A,r Add register r and the carry to A ADC B  
ADC # ADC A,# Add number # and the carry to A ADC 2  
ADC HL,rr   Add 16 bit register rr and the carry to HL ADC HL,BC  
ADD r ADD A,r Adds register r to A ADD B  
ADD # ADD A,# Adds number # to A ADD 5 remember ADD 254 is like -2
ADD HL,rr   Add 16 bit register rr to HL ADD HL,BC remember ADD HL,65534 is like -2
AND r AND A,r bitwise AND register r with A (result in A) AND B  
AND # AND A,# bitwise AND number # with A (result in A) AND %11100000 if A=%00111000... AND %11100000 will result in %00100000
BIT b,r   get bit b from register r (0 is far right bit) BIT 7,a  
CALL ##   Call address ## Call &4000 this is like GOSUB.... Note: CALL pushes PC onto the stack
CALL c,##   Call address ## if condition C is true Call Z,&4000 this is a THEN GOSUB statemenr
CCF   Invert carry flag (compliment carry flag) CCF if you want to clear the carry flag, do OR A ... it will set carry flag to 0 - alternatively SCF, CCF will do the same but is slower
CP r   Compare A to register r CP A this is your IF statement
CP #   Compare A to number # CP 255 Use OR A instead of CP 0 - it has the same effect but is smaller and faster!
CPD   Compare and decrease... CP (HL), DEC HL,DEC BC, PO set if BC=0 CPD  
CPDR   Compare and decrease, repeat ... CP (HL), DEC HL,DEC BC, until BC=0 or found CPDR  
CPI   Compare and increase... CP (HL), INC  HL,DEC BC, PO set if BC=0 CPI  
CPIR   Compare and increase, repeat ... CP (HL), INC HL,DEC BC, until BC=0 or found CPIR  
CPL   invert all bits of A (ones ComPLiment) CPL turn %11001100 to %00110011
DAA   Special command for Binary coded Decimal, update A DAA  
DEC r   decrease value in register r by one DEC B  
DEC rr   decrease value in register rr by one DEC HL  
DI   Disable interrupts DI do this if you need to use shadow registers
djnz #   Decrease B and Jump if NonZero... only jr jump, so must be close.. +-128 bytes away DJNZ label this is how you do a loop
EI   Enable interrupts EI  
EX (SP),HL   Exchange the bytes at (SP) with the value in HL EX (SP),HL  
EX AF,AF'   Exchange A & Flags with the shadow register A' & Flags' EX AF,AF'  
EX DE,HL   Exchange DE with HL EX DE,HL HL can do more than DE, so you may want to quicly swap them
EXX   Exchange BC,DE,HL with the shadow registers  BC',DE',HL' EXX Use this if you need these register values  later, but want to do another job now
HALT   Wait until interrupt occurs HALT force music to play... DI, HALT will force emulator to stop permanently - good for debugging a particular point in the code on MSX or SPECTRUM where defining breakpoints is hard
IM 0   Interrupt mode 0    
IM 1   Interrupt mode 1   on the MSX or CPC this is the only one you need... on spectrum the firmware uses this mode.
on MSX... IM 1 does a CALL &0038 every screen refresh (50hz)
on CPC... IM 1 does a CALL &0038 6 times during screen refresh (300hz)
IM 2   Interrupt mode 2   the only usable interrupt mode for our game on the ZX SPECTRUM
not as easy to use as IM 1... you need to use all the memory between &8000-&8183 and the I register - but the result is the same as  IM1 on the CPC/MSX
IN A,(#)   read from port # into A IN A,(3) Note: the Z80 on the ZX SPECTRUM & CPC are wired strangely  - this function does not work right
IN R,(C)   read from port C into A in A,(C). Note: the Z80 on the ZX SPECTRUM & CPC are wired strangely - they actually do in (B)
IN (C)   read from port C... but do nothing with it in (C). Note: the Z80 on the ZX SPECTRUM & CPC are wired strangely  - they actually do IN (B)
INC r   increase r by one INC B  
INC rr   increase rr by one INC HL  
IND   In (HL),(C)...dec HL... dec B IND Note: the Z80 on the ZX SPECTRUM & CPC are wired strangely - they actually do in (B) which is also the loopcounter, so this will malfunction
INDR   In (HL),(C)...dec HL... dec B until B=0 INDR Note: the Z80 on the ZX SPECTRUM & CPC are wired strangely - they actually do in (B) which is also the loopcounter, so this will malfunction
INI   In (HL),(C).. Inc HL... Dec B INI Note: the Z80 on the ZX SPECTRUM & CPC are wired strangely - they actually do in (B) which is also the loopcounter, so this will malfunction
INIR   In (HL),(C).. Inc HL... Dec B until B=0   Note: the Z80 on the ZX SPECTRUM & CPC are wired strangely - they actually do in (B) which is also the loopcounter, so this will malfunction
JP (HL)   Jump to the address stored in HL JP (HL) if HL=&4000 then this would be the same as JP &4000
JP ##   Jump to ## JP &4000 JP is like GOTO
JP c,##   Jump to ## if condition c is true    this is like a THEN GOTO statement
JR #   jump to # must be +- 128 bytes from current location JR label JR is faster than JP when FALSE (jump didn't occur) but slower when TRUE (Jump occurs)
JR saves one byte over JP - which is a lot on an 8 bit!
JR c,#   jump to # if condition c is true
must be +- 128 bytes from current location
JR z,label JR is faster than JP when FALSE (jump didn't occur) but slower when TRUE (Jump occurs)
JR saves one byte over JP - which is a lot on an 8 bit!
LD (rr),A   Load memory address in (rr) with value in A LD (BC),A  
LD (##),A   Load A into memory address ## LD (&4000),A  
LD (##),rr   load 16 bit regisers rr  into memory ## and ##+1 LD (&4000),BC bytes are loaded into memory in reverse order - though this only matters if you're doing clever things!, eg if you do:
LD BC,&1122... LD (&4000),BC
&4000 = 22    &4001=11
LD (##),SP   Load the Stack Pointer into ## LD (&4000),SP back up the stack pointer - you may want to use an alternate stack for a while or misuse the stack pointer for fast bulk reading or writing.
LD A,(rr)   Load A from memory location (rr) LD A,(BC)  
LD A,(##)   Load A from memory location ##   LD A,(&4000)
LD A,I   Load A from the Interrupt register   The interrupt register only has a few commands, you'll need this to get data out of I
LD A,R   Load A from the Refresh register   R constantly changes - you can use it for random numbers!
LD rr,(##)   load 16 bitr registers rr from memory (##) ld BC,(&4000)  
LD I,A   Set the Interrupt register I to A   This is the only way to get data into I
LD R,A   set Refresh register to A   NEVER USE THIS - I'm told messing with the refresh register can damage your computers memory!
LD SP,(##)   Loadthe stack pointer from memory location ##   Restore the stack pointer.
LD SP,HL   Set the Stack Pointer to HL    
ld r1,r2   Load r2 into r1 LD B,C you can't do LD HL,BC
instead do LD H,B... LD L,C
LD r,#   Load r with value #   note, rather than doing LD A,0 do XOR A... it's faster but the result is the same!
LD r,(ir+-#)   Load r from intirect register ir (IX or IY) +-# bytes LD B,(IY+4) IF IY=&4000
LD B,(IY+4)  would do the same as LD B,(&4004)
LD rr,##   loads 16 bit number ## into rr LD BC,&4000  
LDD   Load (DE),(HL)... DEC HL... DEC DE... DEC BC LDD Copy a range backwards with no repeat - faster to do this 10 times than to use LDDR with B=10
LDDR   Load (DE),(HL)... DEC HL... DEC DE... DEC BC... Repeat until B=0 LDDR Copy a range backwards
LDI   Load (DE),(HL)... INC HL... INC DE... DEC BC   Copy a range with no repeat - faster to do this 10 times than to use LIDR with B=10
LDIR   Load (DE),(HL)... INC HL... INC DE... DEC BC... Repeat until B=0   Copy a range ... can also be used to fill a range by setting DE=HL+1
NEG   NEG   Negates A... Turns 5 into -5 (251)
NOP   Does nothing   Can be used as a placeholder for self modyfying code
OR r OR A,r Bitwise OR r with A OR B  
OR # OR A,# Bitwise OR # with A OR %11100000 if A=%00111000... OR %11100000 will result in %11111000
OTDR   OUT (C),(HL)... DEC HL... DEC B... repeat until B=0   Note: the Z80 on the ZX SPECTRUM,  Sam Coupe &CPC are wired strangely - ZX & CPC actually do OUT (B),(HL)... DEC HL... DEC B.. note B is loop counter and port! this is dangerous!
OTIR   OUT (C),(HL)... INC HL... DEC B... repeat until B=0   Note: the Z80 on the ZX SPECTRUM,  Sam Coupe &  CPC are wired strangely - ZX & CPC actually do OUT (B),(HL)... INC HL... DEC B.. note B is loop counter and port! this is dangerous!
OUT (#),A   OUTput A to port # (Does not work on CPC/XZ.. see notes)   Note: the Z80 on the ZX SPECTRUM,  Sam Coupe & CPC are wired strangely - this command does not work
OUT (C),r   OUTput r to port C (B on CPC/XZ.. see notes) OUT (C),C ZX, Sam Coupe & CPC actually do OUT (B),r... this allows port and val BC to be set in one go by LD BC,nn
OUT (C),0   OUTput 0 to port C (B on CPC/XZ.. see notes) OUT (C),0 Note: the Z80 on the ZX SPECTRUM & CPC are wired strangely - they actually do OUT (B),0
OUTD   OUT (C),(HL)... DEC HL... DEC B   ZX, Sam Coupe & CPC actually do OUT (B),(HL)... DEC HL... DEC B.. note B is loop counter and port!
OUTI   OUT (C),(HL)... INC HL... DEC B   ZX, Sam Coupe & CPC actually do OUT (B),(HL)... INC HL... DEC B.. note B is loop counter and port!
POP rr   Pops 2 bytes off the stack and puts them in rr POP DE Much faster than reading DE using LD DE,(####) or LD D,(HL).. INC HL, LD E,(HL)
PUSH rr   Push 2 bytes from rr into the stack PUSH DE Much faster than writing DE using LD (####),DE or LD (HL),D.. INC HL, LD (HL),E
RES r,b   Reset bit b of register r (sets bit to zero)... Bit 0 is far right RES B,0 using AND # would need loading register into accumulator - this does not
RET   Return RET return... Note: RET pops PC off the stack
RET f   Return if condition f is true RET Z  
RETI   Return from Interrupt RETI You'll probably never need this
RETN   Return from Non maskable interrupt RETN You'll probably never need this
RL r   Rotate bits in r left using carry bit RL B RL A is faster than RL B
RLC r   Rotate Left and Copy bit 7 to bit 0 (wrapping the bits) RLC B doing RLC B on  %10011000 results in %00110001
RLD   Rotate Left 4 bit nibble at Destination (HL) using bits 0-3 of A as a carry RLD  
RR r   Rotate bits in register r Right with carry RR B RR A is faster than RR B
RRC r   Rotate Right  and Copy bit 0 to bit 7 (wrapping the bits) RRC B doing RRC B on  %00001101 results in %10000110
RRD   Rotate Right 4 bit nibble at Destination (HL) using bits 0-3 of A as a carry    
RST 0   Call &0000 RST 0 Rst's are one byte calls, they can save memory! you cant reconfigure them on SPECTRUM
RST 1-5   Call &0008 (1),   &0010 (2),   &0018 (3),   &0020 (4),   &0028 (5) RST 3  
RST 6   Call &0030   CPC defines this as spare for user configuration
RST 7   Call &0038 RST 7 RST 7 is called by the Z80 when an interrupt occurs... put your own interrupt handler here to take over interrupts from the firmware!
SBC r SBC A,r SuBtract register r and the Carry from A SBC B  
SBC # SBC A,# SuBtract # and the Carry from A SBC 3  
SBC HL,rr   Subtract 16 bit register rr from HL with the carry SBC HL,DE Note there is no SUB HL,rr command - so clear the carry and use this if you need to, or use cpl to flip all the bits in each of the two regisers in rr, then do INC rr  then ADD HL,rr
SCF   Set the Carry Flag    
SET b,r   set Bit b to 1 in register r (note bit 0 is at the far right( Set A,0 using OR # would need loading register into accumulator - this does not
SLA r   Shift Left r and Alter bit 0 to 0 SLA B %01111101 becomes %11111010
SLL r   Shift Left and Load bit 0 with 1 SLL B %01111101 becomes %11111011
SRA r   Shift Right r and Alter bit 7 to same as previous bit 7 SRA B %01111101 becomes %00111110
SRL r   Shift Right and Load bit 7 with 0 SRL B %01111101 becomes %10111110
SUB r SUB A,r Subtract register r from A SUB B  
SUB # SUB A,# Subtract number # from A SUB 5  
XOR r XOR A,r XOR (invert bits) in X with register r XOR B XOR A does the same as LD A,0 ... but is smaller and faster!
XOR #   XOR (invert bits) in X with number # (when bit in # is 1) XOR %11110000 if A=%00111000 and you do XOR %11110000 the result is %11001000
Note, as the Accumulator does most mathmatical operations you can just enter ADD 4 ... but ADD A,4 has the same meaning... the shorter form will be used in my guide
Note, On the CPC,Sam Coupe and ZX Spectrum due to the way the Z80 is wired OUT commands do not work as stated... OUT (C) will actually do OUT (B).. making commands like OTIR  that use B as a counter likely to cause hardware problems (It caused random disk writes to me!)... OUT (#) will not work at all... OUT  works normally on he MSX  OUT (C) will do Out (C) and Out (#) works as stated


My Resources for download
Cheatsheet  - A new improved version of the cheatsheet I use for Z80 programming, you'll need to print at high resolution in color - it's intended to be printed on two sides of A4 paper or card and laminated
Sources - The source code with comments for all the lessons on here!
Development Tools - A collection of free software and my scripts for easy building of programs for MSX,CPC,Spectrum,Enterprise and TI-83

Z80 Links
Zilog Z80 manual - The official manual , it's compelx, but if you need a definitive answer you'll find it here so this should be in your toolkit
Learn ASM in 28 days - I learned from this tutorial, it's aimed at the TI-83 calc, but that uses a Z80... if you don't like my tutorial, try this one!
Down to the silicon - Not remotely needed for programming, but this amazing technical breakdown of how the Z80 works is really something
Vasm - The recommended assembler is WinApe, however if you don't want to use windows, the opensource VASM in 'OldStyle' mode will work too.

Main Z80 systems:

Amstrad Links
Winape  - Not just the easiest to use CPC emulator, but the easiest Z80 platform for beginner ASM programmers!
www.cantrell.org.uk
  - A great source of CPC and ASM info.. My cheatsheet ASM list is based on the one from cantrell.org.uk
CRTC - Details of the amstrad CPC CRTC hardware
Amstrad Firmware guide - Pdf documenting the CPC firmware calls
CPC Firmware Guide - Detailed info on how the CPC hardware and firmware
Basic Manual - You'll want to know at least enough basic to do calls and operate the computer
CpcWiki - Web community full of helpful people!

Spectrum Links
Fuse - My Spectrum emulator of choice!
Spectrum 128k and Spectrum 48K reference - Great summary of the hardware - provides much of the info you'll want for ZX dev
Basic Manual - You'll want to know at least enough basic to do calls and operate the computer
Spectrum Computing Forum - Web community full of helpful people!

MSX Links
OpenMSX - The best MSX emulator - you'll need to find your own roms for disk system emulaton (For disk images I reccomend emulating the hb-f1xdj)
Grauw.nl
- great resources and samplecode for MSX development
MSX2 Technical Handbook - full breakdown of the MSX2 hardware - all the info you will need for development purposes
Basic Manual - You'll want to know at least enough basic to do calls and operate the computer
V9938 Programmer’s Guide - Detailed manual on programming the MSX2 V9938 GPU
MSX.Org - Web community full of helpful people!

Other interesting Z80 systems:

Enterprise Links
EP128 - Enterprise 128 Emulator
Ep128.hu - The only site I've found so far with english Enterprise128 manuals for download!
HTML Technical manual - detailed technical documentation that's easy to read
My Enterprise 128 Site - A quick summary of my info on this rare machine!

TI-83 Links
WabbitEmu - Fine Ti-83 emulator - you'll need your own rom files from somewhere!
Learn ASM in 28 days - I learned from this tutorial, it's a great guide that covers the Ti-83

Sam Coupe Links
SimCoupe - Great Sam Coupe Emulator!
Sam.Speccy.cz - Great site with tons of technical info
User Manual - Basic usage guide
Technical Manual - You'll want this for programming
My Sam Coupe site - Some useful stuff I've collected together, that is not widely enough distributed.

Sega Mastersystem & GameGear (Both systems are practically identical)
Fusion - Emulates the Master System, Game Gear and Others!
SMS Power - Sega Master system programming info
My SMS/GG site - A breif overview of the system, I'll be including these systems in my tutorials soon!

Not a Z80, but it's close, and people think it is!


Gameboy & Gameboy Color (Both systems are practically identical)
Visual Boy Advance - Emulates the Gameboy , Gameboy Color, and others!
GameBoy Dev'rs - Gameboy programming website
Color Programming - The only site I can find with examples about Color on the gameboy!
Vasm - You can't easilly compile for the gameboy with a standard Z80 assembler like  WinApe... but VASM in 'OldStyle' mode is pretty similar to WinApe
My Gameboy site - Some useful stuff I've collected together, more to follow!

Learn Assembly for the Greatest Classic Processors:  Z80 - 6502 - 68000
Visit www.ChibiAkumas.com to get my games and their source code! | Support me on patreon