learning-cuda

Exercise 1: CUDA Basics

Part 1: Convert RGB image to gray-scale image

Part 2: Blur image by convolution with Gaussian filter

Exercise 2: How code is executed in CUDA

Part 1: Array (sum) reduction

Part 2: Calculate sum of two arrays using streams

Exercise 3: Types of memory in CUDA

Task: The same as Part 2 of Exercise 1

kernel 1: using only GMEM

kernel 2: copy the image from GMEM to SMEM, filter is still on GMEM

kernel 3: copy the image from GMEM to SMEM and the filter from GMEM to CMEM

Exercise 4: Parallelizing Radix Sort

Task: Paralellize Radix Sort using CUDA

Parallelized Bits Extraction: simply divide the arrays into blocks and attract bits on those

Parallelized Scanning (Prefix Sum) on Bits: load the extracted bits into SMEM then do scanning

Parallelized Ranking and Result writing: load the scanned array into SMEM then compute rank and write result