The main branch keeps the non-optimized version of code. For optimized versions with different methods, just switch to respective branches by git checkout <the branch>

• 1D Test Dataset(EEG Data)
• 2D Test Dataset(CIFAR10)

flowchart TD
	N[Rewrite the Algorithm] --> A
	A --> |Def NO| N
	subgraph code [KNOW UR CODE]
	N
	A
	end
	A{is Parallelizable?}
	Y[["Distribution of Work
	Memory Access Pattern
	Occupancy/Device Usage"]]
	A -->|Independency-YES| Y
	A --> |Dependencies-Maybe NO?| M{{"Requires
	 Synchronizations
	 Barriers"}}
	subgraph machine [KNOW UR MACHINE]
	Y
	M
	end

    style Y fill:#f6f,stroke:#333,stroke-width:4px
	style M fill:#f6f,stroke:#333,stroke-width:4px
	style A fill:#99d,stroke:#f66,stroke-width:2px

Each folder has (and will have) its respective readme describing what type of modules should be placed inside

Make sure you have the following folders

• Data Files should be placed in ./FFT/Data/
• Output results by methods from /FFT/eval/ will be output to /FFT/Data/Results

• FFTW C/C++ library for validating the results (not necessary)
  • MacOS installation: brew install fftw

Change the FFTW library path as needed

In the ./FFT/ folder use the bash command make

If a fresh build is desired, simply run make clean before running make or make all

./It_CT.out or the correponding SLURM script if present

• Consider larger dataset

• Consider larger dataset

• Keep Codes in Separate Branches
• makefile for each
  • OpenMP
  • MPI
  • CUDA
• SLURM script for each
  • OpenMP
  • MPI
  • CUDA