CPU Emulator Tutorial

This program is part of the software suite

that accompanies the book

The Elements of Computing Systems

by Noam Nisan and Shimon Schocken
MIT Press
www.idc.ac.il/tecs
This software was developed by students at the
Efi Arazi School of Computer Science at IDC
Chief Software Architect: Yaron Ukrainitz

CPU Emulator Tutorial, www.idc.ac.il/tecs

Tutorial Index

Slide 1/40

Background
The Elements of Computing Systems evolves around

the construction of a complete computer system,

done in the framework of a 1- or 2-semester course.
In the first part of the book/course, we build the
hardware platform of a simple yet powerful
computer, called Hack. In the second part, we build
the computers software hierarchy, consisting of an
assembler, a virtual machine, a simple Java-like
language called Jack, a compiler for it, and a mini
operating system, written in Jack.
The book/course is completely self-contained,
requiring only programming as a pre-requisite.
The books web site includes some 200 test
programs, test scripts, and all the software
tools necessary for doing all the projects.

CPU Emulator Tutorial, www.idc.ac.il/tecs

Tutorial Index

Slide 2/40

The books software suite

(All the supplied tools are dual-platform: Xxx.bat starts
Xxx in Windows, and Xxx.sh starts it in Unix)

Simulators
(HardwareSimulator, CPUEmulator, VMEmulator):
This tutorial is
about the
CPU emulator.

Used to build hardware platforms and

execute programs;
Supplied by us.

Translators (Assembler, JackCompiler):

Used to translate from high-level to low-level;
Developed by the students, using the books
specs; Executable solutions supplied by us.

Other
Bin: simulators and translators software;
builtIn: executable versions of all the logic

gates and chips mentioned in the book;

OS: executable version of the Jack OS;
TextComparer: a text comparison utility.
CPU Emulator Tutorial, www.idc.ac.il/tecs

Tutorial Index

Slide 3/40

Tutorial Objective

Learn how to use the

CPU Emulator
for simulating the
execution of machine
language programs on
the Hack computer

CPU Emulator Tutorial, www.idc.ac.il/tecs

Tutorial Index

Slide 4/40

The Hack computer

This CPU emulator simulates the
operations of the Hack computer,
built in chapters 1-5 of the book.
Hack -- a 16-bit computer equipped
with a screen and a keyboard -resembles hand-held computers
like game machines, PDAs, and
cellular telephones.
Before such devices are actually
built in hardware, they are planned
and simulated in software.
The CPU emulator is one of
the software tools used for
this purpose.

CPU Emulator Tutorial, www.idc.ac.il/tecs

Tutorial Index

Slide 5/40

CPU Emulator Tutorial

I.

Basic Platform

II. I/O devices

III. Interactive simulation
IV. Script-based simulation
V. Debugging

Relevant reading (from The Elements of Computing Systems):

Chapter 4:

Machine Language

Chapter 5:

Computer Architecture

Appendix B: Test Scripting Language

CPU Emulator Tutorial, www.idc.ac.il/tecs

Tutorial Index

Slide 6/40

CPU Emulator Tutorial

Part I:
Basic Platform

CPU Emulator Tutorial, www.idc.ac.il/tecs

Tutorial Index

Slide 7/40

The Hack Computer Platform (simulated)

Travel Advice:
This tutorial includes some examples
of programs written in the Hack
machine language (chapter 4).
There is no need however to
understand either the language or the
programs in order to learn how to use
the CPU emulator.
Rather, it is only important to grasp the
general logic of these programs,
as explained (when relevant)
in the tutorial.

CPU Emulator Tutorial, www.idc.ac.il/tecs

Tutorial Index

Slide 8/40

The Hack Computer Platform

screen

data
memory
keyboard
enabler

instruction
memory

ALU

registers

CPU Emulator Tutorial, www.idc.ac.il/tecs

Tutorial Index

Slide 9/40

Instruction memory

The loaded code can be viewed

either in binary, or in symbolic
notation (present view)

Instruction memory
(32K): Holds a machine
language program

Next instruction
is highlighted

Program counter (PC) (16-bit):

Selects the next instruction.

CPU Emulator Tutorial, www.idc.ac.il/tecs

Tutorial Index

Slide 10/40

Data memory (RAM)

Data memory (32K RAM), used for:

General-purpose data storage
(variables, arrays, objects, etc.)
Screen memory map
Keyboard memory map

Address (A) register, used to:

Select the current RAM location
OR
Set the Program Counter (PC) for
jumps (relevant only if the current
instruction includes a jump directive).

CPU Emulator Tutorial, www.idc.ac.il/tecs

Tutorial Index

Slide 11/40

Registers
Registers (all 16-bit):
D: Data register
A: Address register
M: Stands for the memory register
whose address is the current
value of the Address register

M (=RAM[A])
D

CPU Emulator Tutorial, www.idc.ac.il/tecs

Tutorial Index

Slide 12/40

Arithmetic/Logic Unit
Arithmetic logic unit (ALU)
The ALU can compute various arithmetic
and logical functions (lets call them f) on
subsets of the three registers {M,A,D}
All ALU instructions are of the form
{M,A,D} = f ({M,A,D})
(e.g. M=M-1, MD=D+A , A=0, etc.)
Current
instruction

M (=RAM[A])

The ALU operation (LHS destination,

function, RHS operands) is specified by
the current instruction.
D

CPU Emulator Tutorial, www.idc.ac.il/tecs

Tutorial Index

Slide 13/40

CPU Emulator Tutorial

Part II:
I/O Devices

CPU Emulator Tutorial, www.idc.ac.il/tecs

Tutorial Index

Slide 14/40

I/O devices: screen and keyboard

Simulated screen: 256 columns by
512 rows, black & white memorymapped device. The pixels are
continuously refreshed from respective
bits in an 8K memory-map, located at
RAM[16384] - RAM[24575].

Simulated keyboard:
One click on this button causes the
CPU emulator to intercept all the
keys subsequently pressed on the
real computers keyboard; another
click disengages the real keyboard
from the emulator.

CPU Emulator Tutorial, www.idc.ac.il/tecs

Tutorial Index

Slide 15/40

Screen action demo

Perspective: Thats how computer
programs put images (text, pictures,
video) on the screen: they write bits into
some display-oriented memory device.
This is rather hard to do in machine
language programming, but quite easy
in high-level languages that write to the
screen indirectly, using OS routines like
printString or drawCircle, as we
will see in chapters 9 and 12.

3. Built-in Refresh action:

The emulator draws the
corresponding pixels on the
screen. In this case, 16 black
pixels, one for each binary 1.

Since all high level programs and OS

routines are eventually translated into
machine language, they all end up
doing something like this example.
1. Select a word in the RAM region
that serves as the screen memory
map, e.g. address 16384 (the first
word in the screen memory map).
2. Enter a value, say 1
(1111111111111111 in binary)

CPU Emulator Tutorial, www.idc.ac.il/tecs

Tutorial Index

Slide 16/40

Keyboard action demo

1. Click the keyboard enabler

2. Press some key on the
real keyboard, say S

3. Watch here:
Keyboard memory
map
(a single 16-bit
memory location)

CPU Emulator Tutorial, www.idc.ac.il/tecs

Tutorial Index

Slide 17/40

Keyboard action demo

Perspective: Thats how computer
programs read from the keyboard: they
peek some keyboard-oriented memory
device, one character at a time.
Visual echo
AsGUI
long a key
(convenient
pressed,
effect, not is
part
of the
hardware platform)

This is rather tedious in machine

language programming, but quite easy in
high-level languages that handle the
keyboard indirectly, using OS routines
like readLine or readInt, as we will see
in Chapters 9 and 12.
Since all high level programs and OS
routines are eventually translated into
machine language, they all end up doing
something like this example.

Keyboard memory
map
(a single 16-bit
memory location)

CPU Emulator Tutorial, www.idc.ac.il/tecs

The emulator displays

Its character code in the
keyboard memory map

Tutorial Index

Slide 18/40

CPU Emulator Tutorial

Part III:
Interactive
Simulation

CPU Emulator Tutorial, www.idc.ac.il/tecs

Tutorial Index

Slide 19/40

Loading a program

Navigate to a
directory and select
a .hack or .asm file.

CPU Emulator Tutorial, www.idc.ac.il/tecs

Tutorial Index

Slide 20/40

Loading a program

Can switch
from binary to
symbolic
representation

CPU Emulator Tutorial, www.idc.ac.il/tecs

Tutorial Index

Slide 21/40

Running a program

4. Watch
here
2. Click the
run button.

1. Enter a
number,
say 50.

3. To speed up
execution,
use the speed
control slider

Programs description: Draws a rectangle at the

top left corner of the screen. The rectangles width
is 16 pixels, and its length is determined by the
current contents of RAM[0].
Note: There is no need to understand the programs
code in order to understand whats going on.

CPU Emulator Tutorial, www.idc.ac.il/tecs

Tutorial Index

Slide 22/40

Hack programming at a glance (optional)

Next instruction is M=-1.
Since presently A=17536, the next ALU
instruction will effect RAM[17536] =
1111111111111111. The 17536 address,
which falls in the screen memory map,
corresponds to the row just below the
rectangles current bottom. In the next screen
refresh, a new row of 16 black pixels will be
drawn there.

Program action:
Since RAM[0] happens to be 50,
the program draws a 16X50
rectangle. In this example the
user paused execution when
there are 14 more rows to draw.

Programs description: Draws a rectangle at the

top left corner of the screen. The rectangles width
is 16 pixels, and its length is determined by the
current contents of RAM[0].
Note: There is no need to understand the programs
code in order to understand whats going on.

CPU Emulator Tutorial, www.idc.ac.il/tecs

Tutorial Index

Slide 23/40

Animation options

Controls execution
(and animation)
speed.

Animation control:
Program flow (default): highlights the
current instruction in the instruction memory
and the currently selected RAM location
Program & data flow: animates all
program and data flow in the computer
No animation: disables all animation
The simulator can
animate both program
flow and data flow

CPU Emulator Tutorial, www.idc.ac.il/tecs

Usage tip: To execute any non-trivial program

quickly, select no animation.

Tutorial Index

Slide 24/40

CPU Emulator Tutorial

Part IV:
Script-Based
Simulation

CPU Emulator Tutorial, www.idc.ac.il/tecs

Tutorial Index

Slide 25/40

Interactive VS Script-Based Simulation

A program can be executed and debugged:
Interactively, by ad-hoc playing with the emulators GUI
(as we have done so far in this tutorial)

Batch-ly, by running a pre-planned set of tests, specified in a script.

Script-based simulation enables planning and using tests that are:
Pro-active
Documented
Replicable
Complete (as much as possible)
Test scripts:
Are written in a Test Description Language (described in Appendix B)
Can cause the emulator to do anything that can be done interactively,
and quite a few things that cannot be done interactively.
CPU Emulator Tutorial, www.idc.ac.il/tecs

Tutorial Index

Slide 26/40

The basic setting

test script
tested program

CPU Emulator Tutorial, www.idc.ac.il/tecs

Tutorial Index

Slide 27/40

Example: Max.asm
Note: For now, it is not necessary to understand either the Hack
machine language or the Max program. It is only important to
grasp the programs logic. But if youre interested, we give a
language overview on the right.

Hack language at a glance:

(label) // defines a label

//
// Computes
Computes M[2]=max(M[0],M[1])
M[2]=max(M[0],M[1]) where
where MM stands
stands for
for RAM
RAM
@0
@0
D=M
//
D=M
// DD == M[0]
M[0]
@1
@1
D=D-M
//
D=D-M
// DD == DD -- M[1]
M[1]
@FIRST_IS_GREATER
@FIRST_IS_GREATER
D;JGT
//
D;JGT
// If
If D>0
D>0 goto
goto FIRST_IS_GREATER
FIRST_IS_GREATER
@1
@1
D=M
//
D=M
// DD == M[1]
M[1]
@SECOND_IS_GREATER
@SECOND_IS_GREATER
0;JMP
//
0;JMP
// Goto
Goto SECOND_IS_GREATER
SECOND_IS_GREATER
(FIRST_IS_GREATER)
(FIRST_IS_GREATER)
@0
@0
D=M
//
D=M
// D=first
D=first number
number
(SECOND_IS_GREATER)
(SECOND_IS_GREATER)
@2
@2
M=D
//
M=D
// M[2]=D
M[2]=D (greater
(greater number)
number)
(INFINITE_LOOP)
(INFINITE_LOOP)
@INFINITE_LOOP
@INFINITE_LOOP //
// Infinite
Infinite loop
loop (our
(our standard
standard
0;JMP
//
way
to
terminate
programs).
0;JMP
// way to terminate programs).
CPU Emulator Tutorial, www.idc.ac.il/tecs

Tutorial Index

@xxx

// sets the A register

// to xxxs value

The other commands are selfexplanatory; Jump directives

like JGT and JMP mean Jump
to the address currently stored
in the A register
Before any command involving
a RAM location (M), the A
register must be set to the
desired RAM address
(@address)
Before any command involving
a jump, the A register must be
set to the desired ROM
address (@label).
Slide 28/40

Sample test script: Max.tst

//
// Load
Load the
the program
program and
and set
set up:
up:
load
load Max.asm,
Max.asm,
output-file
output-file Max.out,
Max.out,
compare-to
Max.cmp,
compare-to Max.cmp,
output-list
output-list RAM[0]%D2.6.2
RAM[0]%D2.6.2
RAM[1]%D2.6.2
RAM[1]%D2.6.2
RAM[2]%D2.6.2;
RAM[2]%D2.6.2;
//
// Test
Test 1:
1: max(15,32)
max(15,32)
set
RAM[0]
15,
set RAM[0] 15,
set
set RAM[1]
RAM[1] 32;
32;
repeat
14
{
repeat 14 {
ticktock;
ticktock;
}}
output;
output; //
// to
to the
the Max.out
Max.out file
file
//
// Test
Test 2:
2: max(47,22)
max(47,22)
set
set PC
PC 0,
0, //
// Reset
Reset prog.
prog. counter
counter
set
RAM[0]
47,
set RAM[0] 47,
set
set RAM[1]
RAM[1] 22;
22;
repeat
14
{
repeat 14 {
ticktock;
ticktock;
}}
output;
output;

The scripting language

has commands for:

Loading programs
Setting up output and compare files
Writing values into RAM locations
Writing values into registers
Executing the next command (ticktack)
Looping (repeat)
And more (see Appendix B).

Notes:

As it turns out, the Max program requires 14

cycles to complete its execution

All relevant files (.asm,.tst,.cmp) must

be present in the same directory.
Output
|| RAM[0]
RAM[0] || RAM[1]
RAM[1] || RAM[2]
RAM[2] ||
||
15
32
32
15 ||
32 ||
32 ||
||
47
22
47
47 ||
22 ||
47 ||

//
// test
test 3:
3: max(12,12)
max(12,12)
//
Etc.
// Etc.
CPU Emulator Tutorial, www.idc.ac.il/tecs

Tutorial Index

Slide 29/40

Using test scripts

Speed
control

Load a
script
Script = a series of
simulation steps, each
ending with a semicolon;
Reset
Important
point: Whenever an assembly
the script
program (.asm file) is loaded into the
emulator, the program is assembled on the
Pause
fly into the
machine language code, and this is
simulation
the code that actually gets loaded. In the
process, all comments and white space are
Execute step
after from the code, and all symbols
removed
step repeatedly
resolve to the numbers that they stand for.
Execute the next
simulation step

CPU Emulator Tutorial, www.idc.ac.il/tecs

Tutorial Index

Slide 30/40

The default script

(and a deeper understanding of the CPU emulator logic)

If you load a program file without first

loading a script file, the emulator
loads a default script (always). The
default script consists of a loop that
runs the computer clock infinitely.

Note that these run/stop buttons dont

control the program. They control the
script, which controls the computers
clock, which causes the computer
hardware to fetch and execute the
programs instructions, one instruction
per clock cycle.

CPU Emulator Tutorial, www.idc.ac.il/tecs

Tutorial Index

Slide 31/40

CPU Emulator Tutorial

Part V:
Debugging

CPU Emulator Tutorial, www.idc.ac.il/tecs

Tutorial Index

Slide 32/40

Breakpoints: a powerful debugging tool

The CPU emulator continuously keeps track of:

A:

value of the A register

D:

value of the D register

PC:

value of the Program Counter

RAM[i]: value of any RAM location

time:

number of elapsed machine cycles

Breakpoints:

A breakpoint is a pair <variable, value> where variable is one of

{A, D, PC, RAM[i], time} and i is between 0 and 32K.

Breakpoints can be declared either interactively, or via script commands.

For each declared breakpoint, when the variable reaches the value, the
emulator pauses the programs execution with a proper message.

CPU Emulator Tutorial, www.idc.ac.il/tecs

Tutorial Index

Slide 33/40

Breakpoints declaration

2. Previouslydeclared
breakpoints
1. Open the
breakpoints
panel

3. Add, delete,
or update
breakpoints

CPU Emulator Tutorial, www.idc.ac.il/tecs

Tutorial Index

Slide 34/40

Breakpoints declaration

1. Select the system

variable on which you
want to break

2. Enter the value

at which the break
should occur

CPU Emulator Tutorial, www.idc.ac.il/tecs

Tutorial Index

Slide 35/40

Breakpoints usage

1. New
breakpoint
2. Run the
program

3. When the A register will be 2, or

RAM[20] will be 5, or 12 time units
(cycles) will elapse, or RAM[21] will
be 200, the emulator will pause the
programs execution with an
appropriate message.
A powerful debugging tool!

CPU Emulator Tutorial, www.idc.ac.il/tecs

Tutorial Index

Slide 36/40

Postscript: Maurice Wilkes (computer pioneer) discovers debugging:

As soon as we started programming, we found to our
surprise that it wasn't as easy to get programs right
as we had thought. Debugging had to be discovered. I
can remember the exact instant when I realized that
a large part of my life from then on was going to be
spent in finding mistakes in my own programs.
(Maurice Wilkes, 1949).

CPU Emulator Tutorial, www.idc.ac.il/tecs

Tutorial Index

Slide 37/40