Programming in CUDA
This repo takes you through my journey in CUDA programming DEaling with Concurrencies, harnessing Parallelism and a lot more. Join me on this thrilling journey,
Here is a guide on navigating through this repo:
The Root contains many folders each consisting of various CUDA programs. help.cu in the root file is used in all the programs.
An easy way to execute programs is thus to move the program and its dependancies to the root directory, update the makefile to include any extra dependancies and run the makefile.
If proceeding as above dont forget to name the source file as main.cu
To do list
- Shared memory Optimization and Coalescing a.Tiled Matrix Multiplication b.Image convolution
- Synchronization and Atomic operations a.Histogram Calculation b.Concurrent Hash Map c.Parallel Breadth First Search
- Parallel Algorithms and Data Structures a. Parallel Reduction- Array addition, Tree based Reduction b.Bitonic Sort or Merge Sort Concepts o Cover:
- Warp Divergence
- Shared Memory access