Cisco optimization for x86 & fixed point

1. Only for fixed point on x86 platform (32bit and 64bit, uses SIMD
   intrinsics up to SSE4.2)
2. Use "configure --enable-fixed-point --enable-intrinsics" to enable
   optimization, default is disabled.
3. Official test cases are verified and passed.

Signed-off-by: Timothy B. Terriberry <tterribe@xiph.org>

1 files changed, 16 insertions, 3 deletions

diff --git a/celt/celt_lpc.h b/celt/celt_lpc.h
index dc2a0a3d..dc8967f0 100644
--- a/celt/celt_lpc.h
+++ b/celt/celt_lpc.h
@@ -29,24 +29,37 @@
 #define PLC_H
 
 #include "arch.h"
+#include "cpu_support.h"
+
+#if defined(OPUS_X86_MAY_HAVE_SSE4_1)
+#include "x86/celt_lpc_sse.h"
+#endif
 
 #define LPC_ORDER 24
 
 void _celt_lpc(opus_val16 *_lpc, const opus_val32 *ac, int p);
 
-void celt_fir(const opus_val16 *x,
+void celt_fir_c(
+         const opus_val16 *x,
          const opus_val16 *num,
          opus_val16 *y,
          int N,
          int ord,
-         opus_val16 *mem);
+         opus_val16 *mem,
+         int arch);
+
+#if !defined(OPUS_X86_MAY_HAVE_SSE4_1)
+#define celt_fir(x, num, y, N, ord, mem, arch) \
+    (celt_fir_c(x, num, y, N, ord, mem, arch))
+#endif
 
 void celt_iir(const opus_val32 *x,
          const opus_val16 *den,
          opus_val32 *y,
          int N,
          int ord,
-         opus_val16 *mem);
+         opus_val16 *mem,
+         int arch);
 
 int _celt_autocorr(const opus_val16 *x, opus_val32 *ac,
          const opus_val16 *window, int overlap, int lag, int n, int arch);