Download as pdf or txt
Download as pdf or txt
You are on page 1of 1

EXERCISE I (30Marks)

1. CUDA is a ……………………………… based programming language (5Marks)


a) python b) C/C++ c) C/C# d) java / R

2. The NVIDIA GPU processor cores are also known as …………………(5Marks)

a) Thread b) Multi-core c) Core d) CUDA cores

EXERCISE II (30Marks)

3. The historical evolution of the GPU architecture can be divided up into three major phases
or eras. Which?(5Marks)

4. Correct this table giving CUDA Terms to GPU’s Hardware Components Equivalence
Mapping (5Marks)

CUDA Term Definition Equivalent GPU Hardware Component


Kernel A group of thread assigned to a Not applicable
particular SM
Grid The GPU GPU/CUDA processor core
Thread Parallel code in the form of a CUDA Multiprocessor (SM)
function to be run on GPU
Block And instance of a kernel on the GPU
GPU

EXERCISE III (40Marks)

5. It is important for the programmer to understand the nuances of the various GPU
memories, particularly the sizes available for each memory type, their relative access
times, and accessibility limitations, to enable correct and efficient code development using
CUDA.
(a) What is CUDA ?(15Marks)
(b) CUDA C Program can be divided into three general sections. Which?
(c). Propose a CUDA representation of a GPU’s Basic Architecture. It will
correspond to the example of GPU having two SMs and two CUDA cores per
SM.

You might also like