mirror of https://git.ffmpeg.org/ffmpeg.git
x86/tx_float_init: properly specify the supported factors of 15xM FFTs
Only powers of two are currently supported.
This commit is contained in:
parent
cc1df4045e
commit
92100eee5b
|
@ -290,11 +290,11 @@ const FFTXCodelet * const ff_tx_codelet_list_float_x86[] = {
|
|||
TX_DEF(fft_sr_ns, FFT, 64, 131072, 2, 0, 384, b8_i2, avx2, AVX2, AV_TX_INPLACE | FF_TX_PRESHUFFLE,
|
||||
AV_CPU_FLAG_AVXSLOW | AV_CPU_FLAG_SLOW_GATHER),
|
||||
|
||||
TX_DEF(fft_pfa_15xM, FFT, 60, TX_LEN_UNLIMITED, 15, TX_FACTOR_ANY, 320, fft_pfa_init, avx2, AVX2,
|
||||
TX_DEF(fft_pfa_15xM, FFT, 60, TX_LEN_UNLIMITED, 15, 2, 320, fft_pfa_init, avx2, AVX2,
|
||||
AV_TX_INPLACE, AV_CPU_FLAG_AVXSLOW | AV_CPU_FLAG_SLOW_GATHER),
|
||||
TX_DEF(fft_pfa_15xM_asm, FFT, 60, TX_LEN_UNLIMITED, 15, TX_FACTOR_ANY, 384, fft_pfa_init, avx2, AVX2,
|
||||
TX_DEF(fft_pfa_15xM_asm, FFT, 60, TX_LEN_UNLIMITED, 15, 2, 384, fft_pfa_init, avx2, AVX2,
|
||||
AV_TX_INPLACE | FF_TX_PRESHUFFLE | FF_TX_ASM_CALL, AV_CPU_FLAG_AVXSLOW | AV_CPU_FLAG_SLOW_GATHER),
|
||||
TX_DEF(fft_pfa_15xM_ns, FFT, 60, TX_LEN_UNLIMITED, 15, TX_FACTOR_ANY, 384, fft_pfa_init, avx2, AVX2,
|
||||
TX_DEF(fft_pfa_15xM_ns, FFT, 60, TX_LEN_UNLIMITED, 15, 2, 384, fft_pfa_init, avx2, AVX2,
|
||||
AV_TX_INPLACE | FF_TX_PRESHUFFLE, AV_CPU_FLAG_AVXSLOW | AV_CPU_FLAG_SLOW_GATHER),
|
||||
|
||||
TX_DEF(mdct_inv, MDCT, 16, TX_LEN_UNLIMITED, 2, TX_FACTOR_ANY, 384, m_inv_init, avx2, AVX2,
|
||||
|
|
Loading…
Reference in New Issue