Finishing Chapter 2.

• The assembler
  • Consider :

  • .data
    text:	.asciiz "hello world"
    endl:	.asciiz "\n"
    .text
    .globl main
    
    main:
        la	$t1, text
        li  $t2, 0
    top:
        lb	$a0, text($t2)
        beqz $a0, done
        li	$v0, 11
        syscall
    
        la	$a0, endl
        li	$v0, 4
        syscall
    
        addi $t2, $t2, 1
        b top
    
    done:
    	li $v0, 10
    	syscall

• How would you "assemble" this?
  • Instructions?
  • Pseudo instructions?
  • How about labels?
    • The symbol table holds these
    • Usually a two step process: find all labels and associated addresses
    • Then process the code, looking up addresses in the symbol table.
  • Different parts of the code go in different locations in memory
    • This depends on the os.
  • You might also need to translate numbers in different bases.
• The tables on 146-147 are interesting.
  • A comparison of bubble sort vs quick sort.
  • They show the inpact of algorithm and choice of instructions.
  • With different compiler optimizations.