CS 179: GPU Computing
Assignment 4

Due: Friday, May 4, 2018 - 3:00 PM

You may submit your final project proposals with this text file. 


Submission: 
------------------

By e-mail to cs179.ta@gmail.com. Package your files in a standard archive format (e.g. zip, tar.gz, tar.bz2). Please also include your name in the title of your archive


Resource usage:
------------------

You'll just be using the CUBLAS API (Feel free to look up general documentation.)

Question 1: Raytracing CUBLAS conversion (coding in comments) (60 pts)
--------------------------------------------------------
--------------------------------------------------------


Introduction:
------------------

Similarily to lab3 this lab will make use of a CUDA library. We will be using cuBLAS to accelerate 
basic linear algebra operations. Due to some unforseen issues regarding cublas interop this section
will involve coding only in the comments, where you will fill in some code in comments that performs
the equivalent linear algebra operation given in one or more cuBLAS calls. Because you won't have an 
easy way to test this we won't be too harsh but please try your best and reference the cuBLAS API documentation
and examples as needed.

To do:
------------------

Convert the BLAS kernels to cublas calls by adding comments for sections marked "TODO" in raytrace.cu .



Notes:
------------------

You'll need to deal with the column major difference during the application
of the scaling and rotation matrices. Since everything else is some form of a
vector of size 3 we don't need to worry about it that much throughout the rest of the code.




Question 2: Matrix multiplication (coding) (40 pts)
--------------------------------------------------------
--------------------------------------------------------


Introduction:
------------------

Since raytracing never actually performs a matrix multiplication we have included
a seperate program to complete in the folder matrixmul. This will be simply performing
the given matrix multiplications using the cublas api and checking against the given cpu code. 

To do:
------------------

Try to complete Question 1 first, but it is not required. 

Then, just fill in the TODOs implementing the matrix multiplications with
the generalized matrix multiplication operation included in cublas. You can test your results
against the given cpu code. 



Notes:
------------------