2024 FLINT development workshop in Bordeaux


workshop1 workshop2 workshop3


Edgar Costa:

Hartmut Monien:

Marc Mezzarobba:

Albin Ahlbäck:

Jean Kieffer:

My main focus was to improve the performance of acb_theta functions by avoiding unnecessary computations as much as possible (I figured it would be a good idea to clean this up before thinking about thresholds with acb_modular_theta):

Ricardo Buring:

Vincent Neiger:

Hello, small report (on my remote participation): besides a bug fix on n_primitive_root, I have been working on a small prime FFT (prime up to ~60 bits, not using floating point conversions). Initially this was with nmod_poly_mat multiplication in mind, but in the end I thought it was good to understand better the FFT for nmod_poly first, and then move to the matrix case. I still need some more time (e.g. only forward FFT is done until now, and with no TFT) and then I should also do some more testing. But my current code is relatively concise and matches the performance of NTL’s forward FFT. It also seems competitive with the current fft_small, if not faster, which surprises me: I would very much welcome any help on a good benchmarking file for fft_small’s forward DFT for various sizes (not counting the time for precomputing root tables).

Bill Allombert:

Kartik Venkatram

Oscar Benjamin

Fredrik Johansson

When and where




Planned projects



To register for the workshop, please contact Fredrik. We expect to have funds to cover local expenses and/or travel for some participants, but we are still waiting for confirmation about the budget. Remote participation will be possible.