flatironinstitute/finufft

Issues

• `-fPIC` for static libraries?

  #544 opened 12 days ago by lgarrison
  1

• Python default should be march=native, CI buildwheel should specify x86

  #540 opened 17 days ago by DiamonDinoia
  0

• document editable installation for python

  #513 opened a month ago by DiamonDinoia
  0

• opts.spread_kerevalmeth should be uint16_t throughout

  #476 opened 2 months ago by ahbarnett
  2

• Provide a single-argument kernel evaluator function that finufft:onedim_*_kernel() can use

  #528 opened 25 days ago by ahbarnett
  0

• Use std::exp in examples with c++11

  #537 opened 23 days ago by DiamonDinoia
  0

• require fftw 3.3.9 and use planner_nthreads in finufft.cpp

  #521 opened a month ago by ahbarnett
  2

• Add a MATLAB math test like basicpassfail, to add to CI

  #515 opened a month ago by ahbarnett
  0

• Update building documentation

  #506 opened 2 months ago by DiamonDinoia
  4

• Adding versions to sphinx

  #514 opened a month ago by DiamonDinoia
  1

• cmake does not pick up the correct flags with clang

  #524 opened a month ago by DiamonDinoia
  0

• Towards 2.3

  #490 opened a month ago by DiamonDinoia
  5

• DUCC should be tested in the CI

  #503 opened 2 months ago by DiamonDinoia
  0

• add makefile (and document) DUCC FFT option

  #479 opened a month ago by ahbarnett
  2

• consider whether the big array fwbatch should be malloced at plan or at execute.

  #480 opened 2 months ago by ahbarnett
  5

• python-cmake and python-buildwheel in CI should be merged into one

  #505 opened 2 months ago by DiamonDinoia
  0

• auto-generate .md table of CPU supported platforms from Cmake CI

  #508 opened a month ago by ahbarnett
  0

• please let's not have the cmake build sphinx docs by default

  #512 opened a month ago by ahbarnett
  2

• Revamp CI

  #509 opened a month ago by DiamonDinoia
  0

• Building for Windows and better options to consume finufft via CMake

  #494 opened 2 months ago by turbotage
  1

• DUCC cmake instructions and documentation

  #504 opened a month ago by DiamonDinoia
  0

• cufinufft should be built with pyproject.toml

  #497 opened 2 months ago by DiamonDinoia
  1

• Overflow and admissible range for float 32 input parameters

  #454 opened 2 months ago by ahbarnett
  4

• make spreadinterp.cpp use upsampfac=1.25 .h coeffs code, instead of old .c code.

  #501 opened 2 months ago by ahbarnett
  0

• cufinufft block_gather method can't handle large non-uniform inputs

  #498 opened 2 months ago by blackwer
  3

• pip install cufinufft on windows

  #441 opened 4 months ago by Divisible8737
  8

• upsampfac=1.25 with tol<1e-9 reports warnings for type=1,2 but errors out for type=3

  #500 opened 2 months ago by ahbarnett
  2

• Technical issues found during the switch to templates

  #484 opened 2 months ago by mreineck
  14

• Dispatching FINUFFT_EXECUTE

  #485 opened 2 months ago by DiamonDinoia
  1

• FINUFFT_EXECUTE architecture dispatch

  #486 opened 2 months ago by DiamonDinoia
  1

• Type 3 on GPU

  #489 opened 2 months ago by DiamonDinoia
  1

• constexpr everything

  #487 opened 2 months ago by DiamonDinoia
  2

• Towards 2.4

  #491 opened 2 months ago by DiamonDinoia
  0

• Default CMake build type is slow

  #465 opened 2 months ago by ludvigak
  7

• odd ns kerevalmeth=0 has O(1) errors for spreader only, in master.

  #475 opened 2 months ago by ahbarnett
  10

• Prolate spheroidal wave function

  #462 opened 3 months ago by DiamonDinoia
  7

• doc CPU opts.spread_sort fix

  #474 opened 2 months ago by ahbarnett
  0

• check for availability of `cudaMallocAsync`

  #445 opened 4 months ago by janden
  10

• [cufinuFFT] Different memory allocation for nuFFT-plan via Python and cpp test

  #443 opened 3 months ago by Sebastian-Belkner
  8

• [cufinufft] Python - modeord parameter

  #424 opened 3 months ago by Sebastian-Belkner
  9

• For SAR data of the same size, the time consumed by FINUFFT is much more than that consumed by sinc interpolation

  #434 opened 5 months ago by zengletian1491
  15

• revisit logic so user *can* set more than OMP max threads. But single-threaded compile cannot segfault if opts.nthrs>1

  #437 opened 3 months ago by ahbarnett
  2

• faster zeroing of internal spreadinterp array.

  #438 opened 3 months ago by ahbarnett
  1

• The plan is passed as a pointer or reference but the API is not consistent

  #460 opened 3 months ago by DiamonDinoia
  0

• MATLAB call cuFINUFFT

  #432 opened 5 months ago by zengletian1491
  6

• Potential candidate for "foldrescale"

  #436 opened 4 months ago by mreineck
  15

• Possible threading issues on Macs

  #430 opened 5 months ago by blackwer
  2

• Non-repeatable behavior of finufft (likely caused by a parallelization issue)

  #428 opened 5 months ago by remy-abergel
  4

• Apparent memory leak in cufinufft

  #427 opened 5 months ago by WardBrian
  9

• Python: Can cufinufft automatically figure out `gpu_device_id`?

  #420 opened 7 months ago by WardBrian
  3