4.3 Introduction to Assembly Programming

4.3 Introduction to Assembly Programming

4.3.1 Reading and Web Sites

1. Instructions add & sub Read IA-32 Volume 2a [4] and Read IA-32 Volume 2b [5]
2. Instructions idiv & imul Read IA-32 Volume 2 [4] Please note the register restrictions.
3. Carter Chapter 1, Introduction Section 1.3 on. pages 11-26 [1].
4. Read IA-32 Intel Architecture Optimization Reference Manual [7] pages 1-17 & 1-18. This will provide some clues to what Intelis doing inside their processor.

4.3.2 Notes

• Memory hierarchy
  • Registers. Very fastest access times.
  • Cache. Level I, II, III, .... Transparent to the user and the programmer except for processing speed. This may impact the system programmer from time to time.
  • DRAM, Core.
  • Rotating storage both magnetic and optical. Magnetic tends to be faster than optical storage.
  • Remote network storage. One form of off site backup.
  • Removable media storage. Tape and optical drives.
• Program execution. What happens when you run a program? Determination of an executable program. Dynamic linked libraries and shared object files. Transfer of control. Change in execution mode. Interrupts.
• Machine language. What the hardware actually understands & how.
• Machine Architecture. Design tradeoffs. Chip size. Chip power. Pin count. Chip I/O speed. Clock speed. Pipelining. Caching. Number and size of registers. Adders, carry look ahead. Multipliers hardware / microcode / software. Barrel shifters, floating point, branch prediction, speculative execution, dual (multiple) cores, SMP (Symmetric Multi-processing), NUMA (Non-Uniform Memory Access), SIMD (Single Instruction, Multiple Data), Vector processing, CISC (Complex instruction Set Computer), RISC (reduced instruction Set Computer), DSP (Digital Signal Processor) - Multiply/Accumulate instruction, Memory address space, I/O address space, Register address space. Instruction length. Orthogonal instruction set. Harvard Architecture vs Von Neumann Architecture
• IntelPentium Architecture. Historic demands. Clock speed vs. throughput.
• AMD
• 64 bit extensions
• Assembler
  • nasm (dual platform)
  • gas, gcc (ALL platforms)
  • MASM, TASM (Microsoftonly)
• Assembler program format
  • Overview of program structure
    Comments, Data, Storage, Macros, & Code
  • segment .data
    initialized data
  • segment .bss
    uninitialized data
  • segment .text
    program
• Dr. Paul A. Carter's setup.
  • Platform independent
  • Multiple assemblers and compilers supported
  • I/O library
    • print_int. Print an integer
    • print_string. Print a string.
    • print_nl. Prints a new line character.
    • Others see: Carterpage 16
  • Debug macros see: Carterpage 17
  • Files
    • skel.asm (A starting point)
    • Your program
    • asm_io.asm (I/O routines)
    • asm_io.inc (Debug macros)
    • driver.c (Startup routine)

      #include "cdecl.h"
      
      int PRE_CDECL asm_main( void ) POST_CDECL;
      
      int main()
      {
        int ret_status;
        ret_status = asm_main();
        return ret_status;
      }
      

  • Linking and running - gort 
    • curl -o linux-ex.zip http://drpaulcarter.com/pcasm/linux-ex.zip
    • unzip linux-ex.zip
    • make (create initial objects and executables)
    • edit (pico lt1.asm)
    • assemble (nasm -f elf lt1.asm)
    • link (gcc -o lt1 driver.o lt1.o asm_io.o)
    • run (./lt1)
  • Linking and running - Microsoft

4.3.3 Lab Assignment: A little math

1. Install nasm or access system that has nasm [9] installed.
2. Download Dr. Carter's programs
3. Compile/assemble his examples (make)
4. Run some of his examples.
5. Write Program #2.
   The last digit of your student services number will be "N" in this lab. The program needs to do the following:
   1. Print out your name and the date followed by a line with the value of N by itself.
   2. Compute the following using 32 bit two's complement arithmetic. DO NOT concern yourself with detecting overflow conditions. Please note that the program uses no flow-of-control structures. Use the following tables below to determine what your program needs to compute:
   3. Test your program. Do all of the arithmetic manually or with a high level programming language. Compare thees results to what your program produces. Turn this in. Testing is an important part of programming.
   4. Printout A, B, C, D, W, X, Y, and Z with labels.
   5. Printout W+X+Y+Z with a label.
   6. Use the debug macro dump_regs at least once in your program.
   7. Turn in your test data (and program if there is one), an error free listing of your program, and the printed output of your program.

4.3 Introduction to Assembly Programming