CSE 431. Computer Architecture - University of Engineering and

Download Report

Transcript CSE 431. Computer Architecture - University of Engineering and

1
Engr. Umbreen Sabir
Computer Engineering Department,
University of Engg. & Technology Taxila.
CA&O Lecture 7 by Engr. Umbreen Sabir
Assemblers and
Compilers
20/2/2009
COMPUTER ARCHITECTURE &
ORGANIZATION
PATH FROM PROGRAMS TO BITS
20/2/2009
CA&O Lecture 7 by Engr. Umbreen Sabir
2
COMPILERS
The compiler transforms the C program into an
assembly language program, a symbolic form of
what the machine understands.
 Higher level language programs take many fewer
lines of code than assembly language.
 So, programmer productivity is higher.
20/2/2009

CA&O Lecture 7 by Engr. Umbreen Sabir
3
HOW AN ASSEMBLER WORKS
Three major components of assembly
1) Allocating and initialing data storage
2) Conversion of mnemonics to binary instructions
3) Resolving addresses
 Pseudoinstructions:

20/2/2009
CA&O Lecture 7 by Engr. Umbreen Sabir
Assembler can also treat common variations of machine
level instructions as if they were instructions in their own
right. Such instructions are called Pseudoinstructions. e.g.
Move $t0, $t1 #register $t0 gets register $t1.
 Assembler converts this assembly language instruction
into the machine language equivalent of the following
instruction:
add $t0, $zero, $t1
 It converts blt into two instructions slt and bne.
4
 MIPS assembler also allows 32-bit constants to be loaded
into a register despite the 16-bit limit of immediate
instructions.

HOW AN ASSEMBLER WORKS CONT.
20/2/2009
The assembler turn the assembly language
program into an object file, which is a
combination of machine language instructions,
data and information needed to place
instructions in memory.
 Assembler keeps track of labels used in branches
and data transfer instructions in a symbol table.

CA&O Lecture 7 by Engr. Umbreen Sabir
5
OBJECT FILE COMPONENTS
1.
2.
4.
5.
6.
CA&O Lecture 7 by Engr. Umbreen Sabir
3.
Object file typically contains six distinct pieces.
Object File Header: describes size and position
of other pieces.
Text Segment: Contains machine code.
Data segment: static or dynamic data.
Relocation information: identifies instructions
and data words that depend on absolute
addresses when the program is loaded in
memory.
Symbol Table: contains the remaining labels
such as external references.
Debugging Information
20/2/2009

6
EXAMPLE
20/2/2009
CA&O Lecture 7 by Engr. Umbreen Sabir
7
RESOLVING ADDRESSES- 1ST PASS
“OLD-STYLE” 2-PASS ASSEMBLER APPROACH
20/2/2009
CA&O Lecture 7 by Engr. Umbreen Sabir
8
RESOLVING ADDRESSES – 2ND PASS
20/2/2009
CA&O Lecture 7 by Engr. Umbreen Sabir
9
MODERN WAY – 1-PASS ASSEMBLERS
Modern assemblers keep more information in
their symbol table which allows them to resolve
addresses in a single pass.
 Known addresses (backward references) are
immediately resolved.
 Unknown addresses (forward references) are
“back-filled” once they are resolved.

20/2/2009
CA&O Lecture 7 by Engr. Umbreen Sabir
10
THE ROLE OF A LINKER
Some aspects of address resolution cannot be
handled by the assembler alone.
1) References to data or routines in other object
modules
2)The layout of all segments in memory
3) Support for REUSABLE code modules
4) Support for RELOCATABLE code modules
 This final step of resolution is the job of a
LINKER

20/2/2009
CA&O Lecture 7 by Engr. Umbreen Sabir
11
LINKER
Also called link editor.
 A system program that combines independently
assembled machine language programs and
resolves all undefined labels into an executable
file.
 Three steps of linker

CA&O Lecture 7 by Engr. Umbreen Sabir
The linker uses the relocation information and
symbol table in each object module to resolve all
undefined labels.
 The linker produces an executable file that can
be run on any computer. This file has same
format as object file, except that it contains no
unresolved references.

20/2/2009
Place code and data modules symbolically in memory.
 Determine the addresses of data and instruction
labels.
 Place both the internal and external references.

12
LOADER
A system program that places an object program
in main memory so that it is ready to execute.
 It follows following steps





CA&O Lecture 7 by Engr. Umbreen Sabir

Reads the exe file header to determine size of the text
and data segments.
Creates an address space large enough for text and
data.
Copies the instructions and data from the exe to
memory.
Copies the parameters to the main program onto the
stack.
Initializes the machine registers and sets the stack
pointer to the first free location.
Jumps to a start-up routine that copies the
parameters into the argument registers and calls the
main routine of the program. Terminates program by
exit system call.
20/2/2009

13
STATIC AND DYNAMIC LIBRARIES
LIBRARIES are commonly used routines stored
as a concatenation of “Object files”. A global
symbol table is maintained for the entire library
with entry points for each routine.
 When routines in LIBRARIES are referenced by
assembly modules, the routine’s entry points are
resolved by the LINKER, and the appropriate
code is added to the executable. This sort of
linking is called STATIC linking.
 Many programs use common libraries. It is
wasteful of both memory and disk space to
include the same code in multiple executables.
The modern alternative to STATIC linking is to
allow the LOADER and THE PROGRAM
ITSELF to resolve the addresses of libraries
routines. This form of linking is called DYNAMIC
linking (e.x. .dll).

20/2/2009
CA&O Lecture 7 by Engr. Umbreen Sabir
14
DYNAMICALLY LINKED LIBRARIES
Library routines are not linked and loaded until
the program is run.
 Program and library routines keep extra
information on the location of nonlocal
procedures and their names.
 In Initial version, loader ran a dynamic linker to
linker, using extra information in file to find
appropriate libraries but it still linked all
routines of the library.
 Then came lazy procedure linker version of DLLs,
where each routine is linked after it is called.

20/2/2009
CA&O Lecture 7 by Engr. Umbreen Sabir
15
DYNAMICALLY LINKED LIBRARIES
20/2/2009
CA&O Lecture 7 by Engr. Umbreen Sabir
16
DYNAMICALLY LINKED LIBRARIES
20/2/2009
CA&O Lecture 7 by Engr. Umbreen Sabir
17
DYNAMICALLY LINKED LIBRARIES
20/2/2009
CA&O Lecture 7 by Engr. Umbreen Sabir
18
MODERN LANGUAGES

Intermediate “object code language”




CA&O Lecture 7 by Engr. Umbreen Sabir

Rather than compiled to assembly language, java is
compiled to instructions that are easy to interpret,
the Java Byte Code,
A software interpreter, Java Virtual Machine (JVM)
can execute java bytecodes.
An interpreter is a program that simulates an
instruction set architecture. E.g. MIPS simulator is
an interpreter.
Upside of interpreter is portability.
Downside of interpreter is low performance.
To preserve portability and improve performance,
Just In Time (JIT) compilers were introduced, which
translate while the program is running.
20/2/2009

19
MODERN LANGUAGES

Intermediate “object code language”
20/2/2009
CA&O Lecture 7 by Engr. Umbreen Sabir
20
MODERN LANGUAGES

Intermediate “object code language”
20/2/2009
CA&O Lecture 7 by Engr. Umbreen Sabir
21
COMPILER OPTIMIZATIONS

High Level Optimization

20/2/2009

Transformations that are done at something close to
the source level.
Local & Global Optimization
Local Optimization


Works within a single basic block.
Global Optimization


CA&O Lecture 7 by Engr. Umbreen Sabir

Works across multiple basic blocks.
Global Register Allocation

Allocates variables to registers for regions of the code.
22
COMPILER OPTIMIZATIONS

Common Subexpression Elimination:

Strength Reduction:

Constant Propagation/Constant Folding:


Find constants in code and propagates them,
collapsing constant values whenever possible.
Copy Propagations:


Replaces complex operations by simpler ones.
CA&O Lecture 7 by Engr. Umbreen Sabir

Finds multiple instances of the same expression and
replaces the second one by a reference to the first.
20/2/2009

Propagates values that are simple copies, eliminating
the need to reload values.
Dead Store Elimination:

Finds stores to values that are not used again and
eliminates the stores.
23
COMPILER OPTIMIZATIONS
Global Code Optimization

Code motion

Finds code that is loop invariant: a particular piece of code
computes the same value on every loop iteration and may
be computed once outside the loop.
CA&O Lecture 7 by Engr. Umbreen Sabir

20/2/2009

Induction variable elimination

Combination of transformations that reduce overhead on
indexing arrays, essentially replacing array indexing with
pointer access.
24
COMPILER OPTIMIZATIONS
20/2/2009
CA&O Lecture 7 by Engr. Umbreen Sabir
25
COMPILER OPTIMIZATIONS
20/2/2009
Example “C” Code:

CA&O Lecture 7 by Engr. Umbreen Sabir
26
UNOPTIMIZED ASSEMBLY OUTPUT

With debug flags set:
20/2/2009
CA&O Lecture 7 by Engr. Umbreen Sabir
27
REGISTER ALLOCATION

Assign local variables to registers
20/2/2009
CA&O Lecture 7 by Engr. Umbreen Sabir
28
LOOP-INVARIANT CODE MOTION

20/2/2009
Assign globals to temp registers and moves
assignments outside of loop
CA&O Lecture 7 by Engr. Umbreen Sabir
29
REMOVE UNNECESSARY TESTS

20/2/2009
Since “i” is initially set to “0”, we already know it
is less than “10”, so why test it the first time
through?
CA&O Lecture 7 by Engr. Umbreen Sabir
30
REMOVE UNNECESSARY STORES

20/2/2009
All we care about it the value of total after the
loop, and simplify loop
CA&O Lecture 7 by Engr. Umbreen Sabir
31
UNROLLING LOOP

20/2/2009
Two copies of the inner loop reduce the branching
overhead
CA&O Lecture 7 by Engr. Umbreen Sabir
32