PerlinNoise

Perlin noise generation algorithm with NVidia CUDA.

List of targets

CUDA

• 1D Perlin noise

  • Noise generation
  • Putting octaves on noise
    • Naïve implementation (Removed as unnecessary)
    • Shared memory* implementation
      • Naïve implementation
      • Implementation optimized for any noise size
    • Layering different noises on top of each other

• 2D Perlin noise

  • Noise generation
  • Putting octaves on noise
    • Naïve implementation
    • Shared memory implementation
    • Layering different noises on top of each other

• 3D Perlin noise

*shared memory is the GPU equivalent of the L1-cache in the CPU.

Benchmarking

• Add code to measure the performance of CUDA cores
• Display the results in the form of a table

planned table view:

#	kernelName	occupancy[%]	perform.[%]	time[ms]	boost[%]
┌1	kernel 1	66	100	0.02199	100
└2	kernel 2 (compare with 1)	100	124.93	0.01718	+17.74
┌3	kernel 3	100	100	0.0228	100
├4	kernel 4 (compare with 3)	100	106.743	0.02136	6.32
└5	kernel 5 (compare with 3)	100	93.097	0.0245	-14.66

C++

• Structure for storing benchmarking results
• Structure for storing the data of the function under test
• Wrap functions for Perlin noise generation in a class
• Create an enum with errors provided by the program

Project design

• Add code documentation
• Translate all comments into English (some of them are written in the native language of the developer - Russian)

Global targets

• Make it into a library
• Generate 2D noise using CUDA directly into a texture video card, explore the possibility of directly mapping textures using OpenGL