Hello I have got this course project in Computer Architectur

Hello,

I have got this course project in Computer Architecture course.If someone could explain me what is asked in parts iii and iv,I\'ll be very happy.What is a benchmark program? am I suppose to write it in memory or cpu? Thank you.

a. Reference system modify the simple CPU example given in class to include a memory module with the following characteristics: i. Size should be 4kbytes of 16 bits words (address width of 12 bits) with data ready output (name it \"dry\'). See appendix for a suggested approach. ii. Use a PLL module from the IP library to generate a clock that run the memory at of the system clock speed. Instantiate the PLL at the top level of the simple CPU architecture to generate the clock of the memory. Modify the memory module iii. Modify the simple CPU example to work with the mew memory mechanism iv. Write a benchmark program for integer matrix multiplication that makes the multiplication by adding the multiplicand as many times as specified by the multipler. Matrix size should be at least 5x5.

Solution

In the 3rd question , it is mentioned like you have to implement the CPU example with the specification mentioned in 1st and 2nd questions. In 1st & 2nd questions the word size, address size , clock speed are given. Use this data to model the CPU example given as reference in your course/class work.

4th question:

Benchmark matrix multiplication means instead of using the multiplication operation use the addition operation to achieve multiplication.

Ex: 4*5=20;

Instead if you add 4 five times to it you will also get the same result. (4+4+4+4+4=20). This is what you should perform instead of using multiplication.

Note: My suggestion is if you are comfirtable with the above approach of matrix multiplication, you can use it. Else you can use Strassen\'s Matrix Multiplication to multiply two matrices, if you are familiar.