Add FFT related operators and APIs#35665

Merged

XiaoguangHu01 merged 112 commits intoPaddlePaddle:developfrom

iclementine:project_fft

Sep 18, 2021

Commits on Aug 10, 2021

1. add interface for fft;
chenfeiyu
committed

Commits on Aug 12, 2021

add fft c2c cufft kernel
jeff41404
committed

Commits on Aug 13, 2021

implement argument checking & op calling parts for fft_c2c and fftn_c2c
chenfeiyu
committed
add operator and opmaker definitions
chenfeiyu
committed
only register float and double for cpu.
chenfeiyu
committed
Merge pull request #1 from iclementine/fft
jeff41404
authored
Merge pull request #2 from jeff41404/fft_c2c_cufft
Feiyu Chan
authored

Commits on Aug 17, 2021

add common code for implementing FFT, add pocketfft as a dependency
chenfeiyu
committed
Merge pull request #3 from iclementine/fft
jeff41404
authored
add fft c2c cufft kernel function
jeff41404
committed
fix bugs in python interface
chenfeiyu
committed
Merge pull request #4 from jeff41404/fft_c2c_cufft
Feiyu Chan
authored

Commits on Aug 18, 2021

add support for c2r, r2c operators, op makers, kernels and kernel functors.
chenfeiyu
committed
test and fix bugs
chenfeiyu
committed
1. fft_c2c function: add support for onesided=False;
chenfeiyu
committed
Merge pull request #5 from iclementine/fft
jeff41404
authored
1. fft: fix python api bugs;
chenfeiyu
committed

Commits on Aug 19, 2021

Commits on Aug 20, 2021

Commits on Aug 26, 2021

complete fft c2c backward in ND
root
committed
Merge pull request #12 from jeff41404/fft_c2c_cufft
Feiyu Chan
authored

Commits on Aug 30, 2021

Commits on Sep 2, 2021

Commits on Sep 4, 2021

Merge pull request #16 from lijiaqi0612/fft_c2r
Feiyu Chan
authored
fix C2R and R2C for cufft, becase the direction is not an option in these cases.
chenfeiyu
committed
Merge pull request #17 from iclementine/fft
Feiyu Chan
authored
add fft r2c onesided with cpu(pocketfft/mkl) and gpu
root
committed
fix bugs in python APIs
chenfeiyu
committed
fix fft_c2r grad kernal
lijiaqi0612
committed
build fft_r2c oneided pass with cpu and gpu
root
committed
Merge pull request #18 from cxxly/fft_r2c
Feiyu Chan
authored
Merge branch 'develop' into project_fft
Feiyu Chan
authored

Commits on Sep 5, 2021

fix bugs in python APIs
chenfeiyu
committed

Commits on Sep 6, 2021

Merge branch 'project_fft' of https://github.com/iclementine/Paddle into fft_c2r
lijiaqi0612
committed
add cuda fft c2r grad kernal functor
lijiaqi0612
committed
clean code
lijiaqi0612
committed
Merge pull request #20 from lijiaqi0612/fft_c2r
Feiyu Chan
authored
Merge pull request #21 from iclementine/fft
Feiyu Chan
authored
Merge branch 'project_fft' of github.com:iclementine/Paddle into fft
chenfeiyu
committed
fix fft_c2r python API
chenfeiyu
committed
Merge pull request #22 from iclementine/fft
Feiyu Chan
authored

Commits on Sep 7, 2021

fill fft r2c result with conjugate symmetry (#19)
cxxly
authored

Commits on Sep 8, 2021

add placeholder for unittests (#24)
Feiyu Chan
authored

Commits on Sep 9, 2021

simple parameterize test function by auto generate test case from parm list (#25)
cxxly
authored
miscellaneous fixes for python APIs (#26)
Feiyu Chan
authored
fix typos in axes checking (#27)
Feiyu Chan
authored
fix argument checking (#28)
Feiyu Chan
authored

Commits on Sep 10, 2021

Commits on Sep 11, 2021

fix typos and clean code (#39)
Feiyu Chan
authored

Commits on Sep 14, 2021

Commits on Sep 15, 2021

remove inclusion of thrust, add __all__ list for fft (#42)
Feiyu Chan
authored
Add api doc and update unittest. (#43)
KPatr1ck
authored
fix MKL-based FFT implementation (#44)
Feiyu Chan
authored
remove code for debug (#45)
Feiyu Chan
authored
use dynload for cufft (#46)
Feiyu Chan
authored

Commits on Sep 16, 2021

Update doc and unittest. (#47)
KPatr1ck
authored
use dynload for cufft (#48)
Feiyu Chan
authored
fix conflicts and merge upstream (#49)
Feiyu Chan
authored
Merge branch 'develop' into project_fft
Feiyu Chan
authored
fix compile error: only link dyload_cuda when cuda is available (#50)
Feiyu Chan
authored
fix dynload for cufft on windows (#51)
Feiyu Chan
authored
add NOMINMAX to compile on windows (#52)
Feiyu Chan
authored
explicitly specify capture mode for lambdas (#55)
Feiyu Chan
authored
fix fft sample (#53)
lijiaqi0612
authored
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into project_fft
chenfeiyu
committed
update scipy and numpy version for unittests of fft (#56)
Feiyu Chan
authored

Commits on Sep 17, 2021