CMSC313, Computer Organization & Assembly Language Programming, Spring 2013

Project 3: Print Decimal Subroutine

Due: Tuesday March 5, 2013 11:59pm

Objective

Assignment

For this assignment, you will write an assembly language subroutine that prints out a 32-bit unsigned number to standard output. The 32-bit value will be passed to your subroutine through the EAX register. Your subroutine must preserve all registers (i.e., any registers that you use must be saved on the stack and restored before you return). Your subroutine should do nothing else (not even print a linefeed).

To determine the values of the base 10 digits in the decimal representation of the number, you will use the division method (don't use repeated subtraction). The DIV instruction will divide the 64-bit value stored in EDX:EAX by the single 32-bit operand given. After the DIV instruction is executed, the quotient is stored in the EAX register and the remainder is stored in the EDX register. (See further discussion of the DIV instruction in Implementation Notes below.)

Implement your subroutine in a separate file. The entry point of your subroutine must be called prt_dec. The "main program" will reside in a separate file. A sample main program is provided for you (main3.asm), but you must make your own test program as well. Note: your test program must be in a separate file, otherwise your subroutine will fail to compile with the test program used for grading.

When you assemble your program, you must assemble each source file separately:

Then, you can run the a.out file produced as usual. A sample run with the main3.asm file provided should look like:

Note: The graders will use a different main program to test your subroutine. Your test program should fully exercise your subroutine and check for cases that are not tested in main3.asm

Implementation Notes

• New: The main3.c program for Project 3 is also available for copying directly on the GL file system at: /afs/umbc.edu/users/c/h/chang/pub/cs313/main3.asm

• In the division method for converting a number to base 10, you will get the one's place first (the least significant digit), then the 10's place, then the 100's place, ... This is in the opposite order that the decimal number will be printed. The easiest way to handle this situation is to prepare a string starting at a higher numbered address and loop backwards storing each character at a lower numbered address.

• The number of characters that you will print will depend on the value passed in the EAX register. You will need to determine or keep count of the length of the decimal number to be printed.

• If the value in the EAX register is 0, then your subroutine should print out a single digit 0.

• In order for code in other files to call your subroutine, you must declare the prt_dec label to be global. (If the label is not global, then only code from the same file can use that label.)

  Thus, you must include the following declaration in your prt_dec.asm file:

  global prt_dec The global declarations are typically made just after the SECTION .text declaration.

• In your test main program, you must tell the NASM assembler that the prt_dec label is from a different file. Otherwise, it will only look for prt_dec in the current file (and not find it). To tell NASM this, include the declaration: extern prt_dec The extern declarations are also typically made just after the SECTION .text declaration.

• There are a couple of quirks about the DIV instruction. First, since the dividend is always EDX:EAX, it is not specified in the instruction. For example, to divide EDX:EAX by EBX, we simply say: div ebx

  Another quirk is that the DIV instruction does not support immediate operands. So, you cannot divide by 10 by saying:

  div 10 ; this is WRONG

• Since you are converting a 32-bit value into decimal, the dividend always fits in the lower 32 bits. Thus, you must zero out the EDX register before each DIV instruction.

• We are always dividing by 10 in this program. So, we know that the remainder in EDX will always be less than 10. This value fits in 8 bits, which is quite rather convenient. Why?

• Remember that the bytes that you are printing to the screen will be interpreted as ASCII values.

Turning in your program

Use the UNIX submit command on the GL system to turn in your project. You should submit three files: 1) the assembly language program and 2) the typescript file of sample runs of your program. The class name for submit is cs313. The name of the assignment name is proj3. The UNIX command to do this should look something like: