Skip to content

Latest commit

 

History

History
38 lines (29 loc) · 866 Bytes

File metadata and controls

38 lines (29 loc) · 866 Bytes

ForRF

Time iterative deconvolution in Fortran with CUDA acceleration.

Requirements

  • Fortran Compiler (gfortran/ifort)
  • CUDA Toolkit (for GPU support)
  • CMake >= 3.10

Build

mkdir build && cd build
cmake .. && make

Usage

Run the test program with an optional argument to select the device (default is GPU).

cd src
# Run on GPU (default)
./forrf_test gpu

# Run on CPU
./forrf_test cpu

Comparison with Seispy

Comparison between CPU and GPU

Performance comparison on a test dataset:

  • CPU: ~3.07s
  • GPU: ~0.23s (~13x speedup)

Defect

The amplitude at peaks may not be exactly consistent with Seispy.