On 04/09/2013 06:57 PM, Christophe Gisquet wrote:
> This patch can be controversial, by assuming floats are IEEE-754 and
> particular behaviour of the FPU will get in the way.
> Timing on Arrandale and Win32 (thus, x87 FPU is used in the reference).
>
> sbr_qmf_pre_shuffle_c: 115 to 76
> sbr_neg_odd_64_c: 84 to 55
> sbr_qmf_post_shuffle_c: 112 to 83
> ---
> libavcodec/sbrdsp.c | 50 +++++++++++++++++++++++++++++++++-----------------
> 1 file changed, 33 insertions(+), 17 deletions(-)
[...]
I'm generally ok with the patch. I'm pretty sure Libav won't work on
non-IEEE-754 due to assumptions being made in other places as well. It
seems safe enough to me. Maybe others disagree though?
> @@ -135,11 +151,11 @@ static void sbr_autocorrelate_c(const float x[40][2],
> float phi[3][2][2])
> float real_sum1 = 0.f, imag_sum1 = 0.f, real_sum0 = 0.0f;
> int i;
> for (i = 1; i < 38; i++) {
> - real_sum0 += x[i][0] * x[i ][0] + x[i][1] * x[i ][1];
> - real_sum1 += x[i][0] * x[i+1][0] + x[i][1] * x[i+1][1];
> - imag_sum1 += x[i][0] * x[i+1][1] - x[i][1] * x[i+1][0];
> - real_sum2 += x[i][0] * x[i+2][0] + x[i][1] * x[i+2][1];
> - imag_sum2 += x[i][0] * x[i+2][1] - x[i][1] * x[i+2][0];
> + real_sum0 += x[i][0] * x[i ][0] + x[i][1] * x[i ][1];
> + real_sum1 += x[i][0] * x[i + 1][0] + x[i][1] * x[i + 1][1];
> + imag_sum1 += x[i][0] * x[i + 1][1] - x[i][1] * x[i + 1][0];
> + real_sum2 += x[i][0] * x[i + 2][0] + x[i][1] * x[i + 2][1];
> + imag_sum2 += x[i][0] * x[i + 2][1] - x[i][1] * x[i + 2][0];
> }
> phi[2-2][1][0] = real_sum2;
> phi[2-2][1][1] = imag_sum2;
But that hunk looks like unrelated cosmetics.
-Justin
_______________________________________________
libav-devel mailing list
[email protected]
https://lists.libav.org/mailman/listinfo/libav-devel
