Mercurial > hg > sv-dependency-builds
annotate src/fftw-3.3.3/TODO @ 169:223a55898ab9 tip default
Add null config files
author | Chris Cannam <cannam@all-day-breakfast.com> |
---|---|
date | Mon, 02 Mar 2020 14:03:47 +0000 |
parents | 89f5e221ed7b |
children |
rev | line source |
---|---|
cannam@95 | 1 TODO before FFTW-$2\pi$: |
cannam@95 | 2 |
cannam@95 | 3 * Wisdom: make it clear that it is specific to the exact fftw version |
cannam@95 | 4 and configuration. Report error codes when reading wisdom. Maybe |
cannam@95 | 5 have multiple system wisdom files, one per version? |
cannam@95 | 6 |
cannam@95 | 7 * DCT/DST codelets? which kinds? |
cannam@95 | 8 |
cannam@95 | 9 * investigate the addition-chain trig computation |
cannam@95 | 10 |
cannam@95 | 11 * I can't believe that there isn't a closed form for the omega |
cannam@95 | 12 array in Rader. |
cannam@95 | 13 |
cannam@95 | 14 * convolution problem type(s) |
cannam@95 | 15 |
cannam@95 | 16 * Explore the idea of having n < 0 in tensors, possibly to mean |
cannam@95 | 17 inverse DFT. |
cannam@95 | 18 |
cannam@95 | 19 * better estimator: possibly, let "other" cost be coef * n, where |
cannam@95 | 20 coef is a per-solver constant determined via some big numerical |
cannam@95 | 21 optimization/fit. |
cannam@95 | 22 |
cannam@95 | 23 * vector radix, multidimensional codelets |
cannam@95 | 24 |
cannam@95 | 25 * it may be a good idea to unify all those little loops that do |
cannam@95 | 26 copying, (X[i], X[n-i]) <- (X[i] + X[n-i], X[i] - X[n-i]), |
cannam@95 | 27 and multiplication of vectors by twiddle factors. |
cannam@95 | 28 |
cannam@95 | 29 * Pruned FFTs (basically, a vecloop that skips zeros). |
cannam@95 | 30 |
cannam@95 | 31 * Try FFTPACK-style back-and-forth (Stockham) FFT. (We tried this a |
cannam@95 | 32 few years ago and it was slower, but perhaps matters have changed.) |
cannam@95 | 33 |
cannam@95 | 34 * Generate assembly directly for more processors, or maybe fork gcc. =) |
cannam@95 | 35 |
cannam@95 | 36 * ensure that threaded solvers generate (block_size % 4 == 0) |
cannam@95 | 37 to allow SIMD to be used. |
cannam@95 | 38 |
cannam@95 | 39 * memoize triggen. |