From 62b40158921b6d800bac70156bdd4a76804334be Mon Sep 17 00:00:00 2001 From: mru Date: Mon, 11 May 2009 02:41:50 +0000 Subject: [PATCH] PPC: 32-bit asm for MAC64 and MLS64 GCC makes a mess of these operations, so give it a hand. 55% faster MP3 decoding on G4. git-svn-id: file:///var/local/repositories/ffmpeg/trunk@18794 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b --- libavcodec/ppc/mathops.h | 30 ++++++++++++++++++++++++++++++ 1 file changed, 30 insertions(+) diff --git a/libavcodec/ppc/mathops.h b/libavcodec/ppc/mathops.h index 2b5a5c4eb..a74fcfa79 100644 --- a/libavcodec/ppc/mathops.h +++ b/libavcodec/ppc/mathops.h @@ -44,4 +44,34 @@ static inline av_const int MULH(int a, int b){ return r; } +#if !HAVE_PPC64 +static inline av_const int64_t MAC64(int64_t d, int a, int b) +{ + union { uint64_t x; unsigned hl[2]; } x = { d }; + int h, l; + __asm__ ("mullw %3, %4, %5 \n\t" + "mulhw %2, %4, %5 \n\t" + "addc %1, %1, %3 \n\t" + "adde %0, %0, %2 \n\t" + : "+r"(x.hl[0]), "+r"(x.hl[1]), "=&r"(h), "=&r"(l) + : "r"(a), "r"(b)); + return x.x; +} +#define MAC64(d, a, b) ((d) = MAC64(d, a, b)) + +static inline av_const int64_t MLS64(int64_t d, int a, int b) +{ + union { uint64_t x; unsigned hl[2]; } x = { d }; + int h, l; + __asm__ ("mullw %3, %4, %5 \n\t" + "mulhw %2, %4, %5 \n\t" + "subfc %1, %3, %1 \n\t" + "subfe %0, %2, %0 \n\t" + : "+r"(x.hl[0]), "+r"(x.hl[1]), "=&r"(h), "=&r"(l) + : "r"(a), "r"(b)); + return x.x; +} +#define MLS64(d, a, b) ((d) = MLS64(d, a, b)) +#endif + #endif /* AVCODEC_PPC_MATHOPS_H */ -- 2.39.2