Code

small prime FFT based on ulong#2107

Code

Draft

vneiger wants to merge 74 commits intoflintlib:mainflintlib/flint:mainfrom vneiger:introduce_nmod_fftvneiger/flint:introduce_nmod_fftCopy head branch name to clipboard

+2,655-2

Commits

Commits on Sep 16, 2024

• add profile for powmod

  vneiger

  committed
  0bf0127

Commits on Oct 19, 2024

• Merge branch 'main' into introduce_nmod_fft

  vneiger

  committed
  c363adb

• add .h file

  vneiger

  committed
  4a887b4

• fix ifndef

  vneiger

  committed
  aee38b3

• context and init code

  vneiger

  committed
  17faaea

• add profile

  vneiger

  committed
  e738256

• fix include

  vneiger

  committed
  dcaede7

• improve profile init

  vneiger

  committed
  cd50787

Commits on Oct 20, 2024

• rename ctx init

  vneiger

  committed
  afa5ddc

• testing init

  vneiger

  committed
  fd24de2

• fix explanations and complete test for init

  vneiger

  committed
  211ab75

• remove printf

  vneiger

  committed
  6368823

Commits on Oct 22, 2024

• forgot to add main

  vneiger

  committed
  9eeedd6

Commits on Oct 23, 2024

• dft, test passes

  vneiger

  committed
  3fa7944

• add profile

  vneiger

  committed
  ff33533

Commits on Oct 24, 2024

• clean things a bit

  vneiger

  committed
  f4520c9

• introducing dft32 base case

  vneiger

  committed
  e10c29c

• dft32 base case

  vneiger

  committed
  7b605a6

• cleaning things

  vneiger

  committed
  1f236d8

• testing from length 1

  vneiger

  committed
  9bf18c7

• fix

  vneiger

  committed
  fb88c54

• remove useless function argument

  vneiger

  committed
  f6cc96c

• vaguely faster with added lazy14 layer

  vneiger

  committed
  a675b68

• clean explanations

  vneiger

  committed
  28b3276

• finalize lazy14 version

  vneiger

  committed
  b71649d

• small fixes

  vneiger

  committed
  8cd392c

• tentative fix for flint_bits == 32

  vneiger

  committed
  9fa9020

• dft8 is now a macro, code generation was too unpredictable

  vneiger

  committed
  ccd3f71

• putting more args slightly slows down for large lengths...

  vneiger

  committed
  f0587e5

• macro for dft16 helps, let's see for dft32

  vneiger

  committed
  4cf7343

• dft32 macroified does help a bit

  vneiger

  committed
  b96e0a5

• mod4 currently unused

  vneiger

  committed
  82a6c85

Commits on Oct 25, 2024

• some notes / todos

  vneiger

  committed
  0cae527

• test with prime close to announced limitw

  vneiger

  committed
  4818198

• limit is 62 bits

  vneiger

  committed
  c413b63

• fix assert and try struct for args... for circumventing strange behaviour for large arrays

  vneiger

  committed
  174bb2d

Commits on Oct 27, 2024

• fft args and introducing iw

  vneiger

  committed
  c7f6dff

• dft 16 macro pointers

  vneiger

  committed
  5b84260

• dft 16 macro pointers

  vneiger

  committed
  1fb2479

• dft8 simplify a bit

  vneiger

  committed
  4d8071f

• clean

  vneiger

  committed
  6e751ea

• dft16 simplify a bit

  vneiger

  committed
  192a575

• dft16 4 4 now ok as well

  vneiger

  committed
  375a84d

• dft32 expanded as well; this is ugly but will help for versions with strides

  vneiger

  committed
  c9b6575

• starting precomp of tab_inverse(w)

  vneiger

  committed
  c24e7b6

• add tab_iw in fitdepth

  vneiger

  committed
  e0ace33

• unroll a bit is faster

  vneiger

  committed
  b376f7e

• notes about init

  vneiger

  committed
  0e0df84

• wip: use multipoint eval in test

  vneiger

  committed
  40539de

• use multipoint eval in test

  vneiger

  committed
  8c8b08b

• idft_t

  vneiger

  committed
  b1fb674

• idft_t, not tested yet

  vneiger

  committed
  dcf2ae7

• minor changes

  vneiger

  committed
  9d8845d

• progress

  vneiger

  committed
  a872720

• add files

  vneiger

  committed
  c29a5d1

• idft test passes

  vneiger

  committed
  66523a7

• idft test passes

  vneiger

  committed
  453c068

• idft in progress

  vneiger

  committed
  82d5cbf

• fix name

  vneiger

  committed
  07a7dbe

Commits on Oct 28, 2024

• idft in progress

  vneiger

  committed
  f0fed07

• idft in progress

  vneiger

  committed
  f79ba6f

• idft in progress

  vneiger

  committed
  c41bcd4

• idft in progress

  vneiger

  committed
  f09bca0

• idft in progress

  vneiger

  committed
  38a5aba

• minor fix

  vneiger

  committed
  ece93dc

• a bit lazier

  vneiger

  committed
  9ef50f1

• idft becoming good, remains some fine tuning to do

  vneiger

  committed
  3b5669d

• a bit lazier

  vneiger

  committed
  614dacb

• fix name

  vneiger

  committed
  9567b15

Commits on Nov 9, 2024

• Merge branch 'flintlib:main' into introduce_nmod_fft

  vneiger

  authored
  42f6fe6

• minor comment

  vneiger

  committed
  c438c43

• Merge branch 'introduce_nmod_fft' of github.com:vneiger/flint into introduce_nmod_fft

  vneiger

  committed
  eaf6d5c

• remove unnecessary includes

  vneiger

  committed
  4bb083a

• add todo about tests idft any node small depths

  vneiger

  committed
  99284b5