Project 2: Rony Edde

[reddeupenn]

Repo Link
redde

*Part 1

• Cpu scan
• Compact without scan
• Compact with scan
• Part 2
  • Naive GPU scan implementation
• Part 3
  • Work efficient scan implementation
  • GPU stream compaction using scan
• Part 4
  • Thrust implementation using cuda's thrust function
• Part 5
  • Radix sort implementation using the CPU
  • Radix sort using naive and thrust implementation

Added CPU benchmarks and GPU function calls benchmarks.

NOTE: CMakelists was also modified to include radix.h and radix.cu

...
Feedback on the project itself, if any.
This was significantly harder than the first project. A lot of time was spent debugging and deciphering
the GPU gems book. I also found it a bit unclear what every implementation needed. I had to spend quite a few allnighters to get this far.