diff options
| author | Lynne <dev@lynne.ee> | 2022-08-29 07:09:03 +0200 |
|---|---|---|
| committer | Lynne <dev@lynne.ee> | 2022-08-29 07:13:43 +0200 |
| commit | f99d15cca0f6ceedb4f2f38d67a52dacc9556fc7 (patch) | |
| tree | 7992f81ab867591101e495719541e4f07a761265 /libavcodec/aarch64/fft_init_aarch64.c | |
| parent | f5c5c04c1452274d223f663bd8ebc9d721acdf0e (diff) | |
| download | ffmpeg-f99d15cca0f6ceedb4f2f38d67a52dacc9556fc7.tar.gz | |
arm/fft: disable NEON optimizations for 131072pt transforms
This has been broken since the start, and it was only discovered
when I started testing my replacement for the FFT.
Disable it, since there's no point in fixing slower code that's about
to be removed anyway.
The vfp version is not affected.
Diffstat (limited to 'libavcodec/aarch64/fft_init_aarch64.c')
| -rw-r--r-- | libavcodec/aarch64/fft_init_aarch64.c | 6 |
1 files changed, 4 insertions, 2 deletions
diff --git a/libavcodec/aarch64/fft_init_aarch64.c b/libavcodec/aarch64/fft_init_aarch64.c index db285205ab..77f5607960 100644 --- a/libavcodec/aarch64/fft_init_aarch64.c +++ b/libavcodec/aarch64/fft_init_aarch64.c @@ -38,8 +38,10 @@ av_cold void ff_fft_init_aarch64(FFTContext *s) int cpu_flags = av_get_cpu_flags(); if (have_neon(cpu_flags)) { - s->fft_permute = ff_fft_permute_neon; - s->fft_calc = ff_fft_calc_neon; + if (s->nbits < 17) { + s->fft_permute = ff_fft_permute_neon; + s->fft_calc = ff_fft_calc_neon; + } #if CONFIG_MDCT s->imdct_calc = ff_imdct_calc_neon; s->imdct_half = ff_imdct_half_neon; |