Description
PART 1:
1. Download and unzip the ZIP file containing the workspace and sample program from here.
Build and test the sample program that’s already in the src subdirectory of the workspace as
follows:
2. Windows Only: Create a shortcut on your desktop to the file named setup.bat that is located
in your workspace folder. Double-click on the shortcut. This will open a command line window with the root of the workspace folder as the current directory.
3. Linux and OS X Only: Open a terminal window and change the current directory to the root
of the workspace directory.
4. Enter the command “make”. This will run the compiler, the assembler and the linker, producing the final program in a file named output.bin.
5. Connect the single board computer to the desktop computer using a mini-USB cable. A window will open as if you had inserted a USB thumb drive. Into that window, copy the file output.bin from the root directory of your workspace. The board will automatically load and execute your program.
PART 2:
1. Delete any existing files in the src and obj subdirectories of your workspace folder.
2. Download the C main program for Lab1 from here and store it in the src subdirectory of your
workspace folder.
3. Use your favorite text editor (not a word processor) to create a second C source code file in
the src subdirectory that implements the three functions shown below. Do not use filenames
containing spaces or filename extensions with uppercase letters. Each array parameter holds
an 8-bit binary number b7.b6b5b4b3b2b1b0 in the range -1.0 ≤ x < +1.0, where (for example)
bin[7] = b7 and bin[0] = b0. Note that Dec2Bin should produce a rounded result.
void TwosComplement(const int input[8], int output[8]) ;
float Bin2Dec(const int bin[8]) ;
void Dec2Bin(const float x, int bin[8]) ;
4. Repeat steps 2-5 of Part 1 to build and download your program to the board.
RUNNING THE PROGRAM:
When the program runs, it will cycle through a long sequence of tests, displaying various input and output values for each of your functions, and what the correct answer should be. If there is an error in one of your functions, the program will pause, display your output in
white text on a red background, and wait for you to press
the blue pushbutton to proceed. If no errors occur after
going through all the tests, the program will display the
message “No Errors!” at the bottom of the screen.
HINTS:
You may be tempted to use the C library pow() function
in Bin2Dec to compute the terms of a polynomial. Although that certainly works, there is a better way to do it.
Whenever you can avoid calling a library function your
code will occupy less memory (use fewer instructions) and usually run faster – both of which are
important in embedded applications because you should always be trying to squeeze the most
performance you can out of a typically inexpensive processor.
But it’s much worse! It’s not just one function call to pow, it actually becomes four function calls!
The pow function requires its input parameters and the return value to be of type double. But
when converting number representations, you need to compute the integer 2
k
, where 2 and k are
also integers. You can certainly give the pow function integer parameters, and you can certainly
store the value that it returns into an integer variable, but doing so requires the compiler to call
library functions to convert back and forth between floats and ints. For example, the assignment
statement “x2n = pow(x, n);” effectively becomes:
x2n = (int) pow((float) x, (float) n) ;
A BETTER WAY TO CALCULATE THE POLYNOMIAL:
Consider an 8-bit binary signed integer, represented as 𝑏7𝑏6𝑏5𝑏4𝑏3𝑏2𝑏1𝑏0, where the b’s are the
1’s and 0’s of the number. The corresponding polynomial would be:
𝑝𝑜𝑙𝑦𝑛𝑜𝑚𝑖𝑎𝑙 = −2
7𝑏7 + 2
6𝑏6 + 2
5𝑏5 + 2
4𝑏4 + 2
3𝑏3 + 2
2𝑏2 + 2
1𝑏1 + 2
0𝑏0
But note that you can rewrite this as:
𝑝𝑜𝑙𝑦𝑛𝑜𝑚𝑖𝑎𝑙 = 𝑏0 + 2(𝑏1 + 2(𝑏2 + 2(𝑏3 + 2(𝑏4 + 2(𝑏5 + 2(𝑏6 − 2𝑏7))))))
Which can be computed using a simple loop:
𝑝𝑜𝑙𝑦𝑛𝑜𝑚𝑖𝑎𝑙 ← −𝑏7
𝑓𝑜𝑟 𝑖 = 6 𝑑𝑜𝑤𝑛 𝑡𝑜 0: 𝑝𝑜𝑙𝑦𝑛𝑜𝑚𝑖𝑎𝑙 ← 2 × 𝑝𝑜𝑙𝑦𝑛𝑜𝑚𝑖𝑎𝑙 + 𝑏𝑖
Function Bin2Dec: The polynomial gives you an integer in the range -128 to +127; divide it by
128 to get the floating-point value to return. Function Dec2Bin: Start by multiplying the floating-point input parameter by 128 and then convert the resulting integer into an array of bits.