From e1958604fdbf67aa8c300d3700c8805db039919d Mon Sep 17 00:00:00 2001 From: Fabrice Bellard Date: Tue, 7 Jan 2003 17:41:43 +0000 Subject: [PATCH] added define for builtins use - inverse fix by Romain Dolbeau Originally committed as revision 1410 to svn://svn.ffmpeg.org/ffmpeg/trunk --- libavcodec/i386/fft_sse.c | 12 ++++++++++++ 1 file changed, 12 insertions(+) diff --git a/libavcodec/i386/fft_sse.c b/libavcodec/i386/fft_sse.c index 8e8e36b0f7..175cea506c 100644 --- a/libavcodec/i386/fft_sse.c +++ b/libavcodec/i386/fft_sse.c @@ -19,11 +19,16 @@ #include "../dsputil.h" #include +#ifdef HAVE_BUILTIN_VECTOR + #include static const float p1p1p1m1[4] __attribute__((aligned(16))) = { 1.0, 1.0, 1.0, -1.0 }; +static const float p1p1m1p1[4] __attribute__((aligned(16))) = + { 1.0, 1.0, -1.0, 1.0 }; + static const float p1p1m1m1[4] __attribute__((aligned(16))) = { 1.0, 1.0, -1.0, -1.0 }; @@ -54,6 +59,11 @@ void fft_calc_sse(FFTContext *s, FFTComplex *z) r = (__m128 *)&z[0]; c1 = *(__m128 *)p1p1m1m1; c2 = *(__m128 *)p1p1p1m1; + if (s->inverse) + c2 = *(__m128 *)p1p1m1p1; + else + c2 = *(__m128 *)p1p1p1m1; + j = (np >> 2); do { a = r[0]; @@ -126,3 +136,5 @@ void fft_calc_sse(FFTContext *s, FFTComplex *z) nloops = nloops << 1; } while (nblocks != 0); } + +#endif