Skip to content

Latest commit

 

History

History
89 lines (57 loc) · 3.6 KB

README.md

File metadata and controls

89 lines (57 loc) · 3.6 KB
Raw

hipFFT

hipFFT is an FFT marshalling library. Currently, hipFFT supports either rocFFT or cuFFT as backends.

hipFFT exports an interface that does not require the client to change, regardless of the chosen backend. It sits between the application and the backend FFT library, marshalling inputs into the backend and results back to the application.

Documentation

Run the steps below to build documentation locally.

cd docs

pip3 install -r .sphinx/requirements.txt

python3 -m sphinx -T -E -b html -d _build/doctrees -D language=en . _build/html

Installing pre-built packages

Download pre-built packages either from ROCm's package servers.

  • On Ubuntu: sudo apt update && sudo apt install hipfft

Transitioning from rocFFT

If you are transitioning from the hipFFT version included in rocFFT to this standalone hipFFT version; please modify your build following this example:

  • previously: hipcc hipfft_1d_z2z.cpp -L/opt/rocm/lib -lrocfft
  • during transition: hipcc -I/opt/rocm/hipfft/include hipfft_1d_z2z.cpp -L/opt/rocm/lib -lhipfft -lrocfft

Building from source

Library build dependencies

To build the hipFFT library:

  • hipFFT depends on rocFFT on AMD platforms;
  • hipFFT depends on cuFFT on Nvidia platforms.

Client build dependencies

  • The clients (samples, tests etc) included with the hipFFT source depend on FFTW, gtest, and boost program-options.

  • The rider and test clients also require the rocFFT source tree to build:

    git submodule update --init

Building hipFFT

To show all build options:

mkdir build && cd build
cmake -LH ..

Here are some CMake build examples:

Hardware target Case Build command line
AMD GPU Build a project using HIP language APIs + hipFFT with standard host compiler cmake -DCMAKE_CXX_COMPILER=g++ -DCMAKE_BUILD_TYPE=Release -L ..
AMD GPU Build a project using HIP language APIs + hipFFT + device kernels with HIP-clang cmake -DCMAKE_CXX_COMPILER=hipcc -DCMAKE_BUILD_TYPE=Release -DBUILD_CLIENTS=ON -L ..
NVIDIA GPU Build a project using HIP language APIs + hipFFT with standard host compiler cmake -DCMAKE_CXX_COMPILER=g++ -DCMAKE_BUILD_TYPE=Release -DBUILD_WITH_LIB=CUDA -L ..
NVIDIA GPU Build a project using HIP language APIs + hipFFT + device kernels with HIP-nvcc HIP_PLATFORM=nvidia cmake -DCMAKE_CXX_COMPILER=hipcc -DCMAKE_BUILD_TYPE=Release -DBUILD_CLIENTS=ON -L ..

Note that the option -DBUILD_CLIENTS=ON is only allowed for the hipcc compiler.

Quick CUDA porting guide

If you have existing CUDA code and want to transition to HIP:

  • HIPIFY your code and fix all unsupported CUDA features or user-defined macros.
  • Build with HIP-nvcc to run on an Nvidia device.
  • Build with HIP-clang to run on an AMD device.

More information about porting to HIP is available on the HIP porting guide.