CUDA vs CPU Performance

Fri Jul 03 2020

High-performance parallel computing is all the buzz right now, and new technologies such as CUDA is making it more accessible to do GPU computing. However, it is vital to know in what scenarios GPU/CPU processing is faster. This post explores several variables that affect CUDA vs. CPU performance. The full Jupyter notebook for this blog post is posted on my GitHub.

For reference, I am using a Nvidia GTX 1060 running CUDA version 10.2 on Linux.

!nvidia-smi
    Wed Jul  1 11:16:12 2020       
    +-----------------------------------------------------------------------------+
    | NVIDIA-SMI 440.82       Driver Version: 440.82       CUDA Version: 10.2     |
    |-------------------------------+----------------------+----------------------+
    | GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
    | Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
    |===============================+======================+======================|
    |   0  GeForce GTX 1060..  Off  | 00000000:01:00.0  On |                  N/A |
    |  0%   49C    P2    26W / 120W |   2808MiB /  3016MiB |      2%      Default |
    +-------------------------------+----------------------+----------------------+
                                                                                   
    +-----------------------------------------------------------------------------+
    | Processes:                                                       GPU Memory |
    |  GPU       PID   Type   Process name                             Usage      |
    |=============================================================================|
    |    0      1972      G   /usr/libexec/Xorg                             59MiB |
    |    0      2361      G   /usr/libexec/Xorg                            280MiB |
    |    0      2485      G   /usr/bin/gnome-shell                         231MiB |
    |    0      5777      G   /usr/lib64/firefox/firefox                     2MiB |
    |    0     33033      G   /usr/lib64/firefox/firefox                     4MiB |
    |    0     37575      G   /usr/lib64/firefox/firefox                   167MiB |
    |    0     37626      G   /usr/lib64/firefox/firefox                     2MiB |
    |    0     90844      C   /home/jeff/Documents/python/ml/bin/python   1881MiB |
    +-----------------------------------------------------------------------------+

Read More »



Creating An Audio Switch

Wed Jul 01 2020

This post covers a bit of an old project, but I wanted to write about how I made my 3.5mm audio switch. The goal of this project was to make an audio switch that had multiple inputs and multiple outputs. At the time, I could not find a commercial product that did this, and similar DIY projects only had multiple inputs and a single output rather than numerous inputs and outputs.

I started this project by making a wiring schematic.

Wiring diagram on white board
Wiring diagram on white board

Read More »



Creating Pixel Art With Open CV

Mon Jun 29 2020

Let's jump right into the fun and start making pixel art with Open CV. Before you read this article, consider checkout out these articles:

Like most CV projects, we need to start by importing some libraries and loading an image.

# Open cv library
import cv2

# matplotlib for displaying the images 
from matplotlib import pyplot as plt

img = cv2.imread('dolphin.jpg')

Read More »



GANs in PyTorch

Sun Jun 21 2020

Generative adversarial networks (GAN) are all the buzz in AI right now due to their fantastic ability to create new content. Last semester, my final Computer Vision (CSCI-431) research project was on comparing the results of three different GAN architectures using the NMIST dataset. I'm writing this post to go over some of the PyTorch code used because PyTorch makes it easy to write GANs.

1 GAN Background

Read More »



PyTorch Headfirst

Sun Jun 14 2020

This post dives headfirst into PyTorch: a powerful open-source ML library for Python. Google's flagship ML library is TensorFlow, whereas Facebook developed PyTorch. Researchers are gravitating towards PyTorch due to its flexibility and efficiency. This tutorial goes over the basics of PyTorch, including tensors and a simple perceptron. This tutorial requires knowledge of Python, Numpy, and neural networks. If you don't know anything about neural networks, I suggest that you watch this amazing video by 3Blue1Brown:

Read More »