[FFADO-svn] r1376 - branches/libffado-2.0/src/libstreaming/amdtp

[<tr...@ff...>]

Author: ppalmers
Date: 2008-10-26 15:01:00 -0700 (Sun, 26 Oct 2008)
New Revision: 1376

Modified:
   branches/libffado-2.0/src/libstreaming/amdtp/AmdtpTransmitStreamProcessor.cpp
Log:
fix clipping and float->int conversion. optimize this a bit. fix SSE code.

Modified: branches/libffado-2.0/src/libstreaming/amdtp/AmdtpTransmitStreamProcessor.cpp
===================================================================
--- branches/libffado-2.0/src/libstreaming/amdtp/AmdtpTransmitStreamProcessor.cpp	2008-10-24 17:58:07 UTC (rev 1375)
+++ branches/libffado-2.0/src/libstreaming/amdtp/AmdtpTransmitStreamProcessor.cpp	2008-10-26 22:01:00 UTC (rev 1376)
@@ -38,8 +38,7 @@
 #include <assert.h>
 #include <cstring>
 
-#define AMDTP_FLOAT_MULTIPLIER 2147483392.0
-
+#define AMDTP_FLOAT_MULTIPLIER (1.0f * ((1<<23) - 1))
 namespace Streaming
 {
 
@@ -515,7 +514,6 @@
 }
 
 #ifdef __SSE2__
-//#if 0
 #include <emmintrin.h>
 #warning SSE2 build
 
@@ -543,6 +541,7 @@
     memset(m_scratch_buffer, 0, nevents * 4);
 
     const __m128i label = _mm_set_epi32 (0x40000000, 0x40000000, 0x40000000, 0x40000000);
+    const __m128i mask = _mm_set_epi32 (0x00FFFFFF, 0x00FFFFFF, 0x00FFFFFF, 0x00FFFFFF);
     const __m128 mult = _mm_set_ps(AMDTP_FLOAT_MULTIPLIER, AMDTP_FLOAT_MULTIPLIER, AMDTP_FLOAT_MULTIPLIER, AMDTP_FLOAT_MULTIPLIER);
 
 #if AMDTP_CLIP_FLOATS
@@ -570,7 +569,6 @@
 
         // the base event for this position
         target_event = (quadlet_t *)(data + i);
-
         // process the events
         for (j=0;j < nevents; j += 1)
         {
@@ -596,8 +594,8 @@
             v_float = _mm_mul_ps(v_float, mult);
             // convert to signed integer
             v_int = _mm_cvttps_epi32( v_float );
-            // shift right 8 bits
-            v_int = _mm_srli_epi32( v_int, 8 );
+            // mask
+            v_int = _mm_and_si128( v_int, mask );
             // label it
             v_int = _mm_or_si128( v_int, label );
 
@@ -606,7 +604,6 @@
             v_int = _mm_or_si128( _mm_slli_epi16( v_int, 8 ), _mm_srli_epi16( v_int, 8 ) );
             // do second swap
             v_int = _mm_or_si128( _mm_slli_epi32( v_int, 16 ), _mm_srli_epi32( v_int, 16 ) );
-
             // store the packed int
             // (target misalignment is assumed since we don't know the m_dimension)
             _mm_storeu_si128 (target, v_int);
@@ -654,13 +651,12 @@
                 v_float = _mm_max_ps(v_float, v_min);
                 v_float = _mm_min_ps(v_float, v_max);
 #endif
-
                 // multiply
                 v_float = _mm_mul_ps(v_float, mult);
                 // convert to signed integer
                 v_int = _mm_cvttps_epi32( v_float );
-                // shift right 8 bits
-                v_int = _mm_srli_epi32( v_int, 8 );
+                // mask
+                v_int = _mm_and_si128( v_int, mask );
                 // label it
                 v_int = _mm_or_si128( v_int, label );
     
@@ -688,13 +684,23 @@
             for(;j < nevents; j += 1) {
                 float *in = (float *)buffer;
 #if AMDTP_CLIP_FLOATS
-                if(*in > 1.0) *in=1.0;
-                if(*in < -1.0) *in=-1.0;
-#endif
+                // clip directly to the value of a maxed event
+                if(*in > 1.0) {
+                    *target_event = CONDSWAPTOBUS32_CONST(0x407FFFFF);
+                } else if(*in < -1.0) {
+                    *target_event = CONDSWAPTOBUS32_CONST(0x40800001);
+                } else {
+                    float v = (*in) * AMDTP_FLOAT_MULTIPLIER;
+                    unsigned int tmp = ((int) v);
+                    tmp = ( tmp & 0x00FFFFFF ) | 0x40000000;
+                    *target_event = CondSwapToBus32((quadlet_t)tmp);
+                }
+#else
                 float v = (*in) * AMDTP_FLOAT_MULTIPLIER;
                 unsigned int tmp = ((int) v);
-                tmp = ( tmp >> 8 ) | 0x40000000;
+                tmp = ( tmp & 0x00FFFFFF ) | 0x40000000;
                 *target_event = CondSwapToBus32((quadlet_t)tmp);
+#endif
                 buffer++;
                 target_event += m_dimension;
             }
@@ -937,14 +943,23 @@
             {
                 float *in = (float *)buffer;
 #if AMDTP_CLIP_FLOATS
-                if(*in > 1.0) *in=1.0;
-                if(*in < -1.0) *in=-1.0;
-#endif
+                // clip directly to the value of a maxed event
+                if(*in > 1.0) {
+                    *target_event = CONDSWAPTOBUS32_CONST(0x407FFFFF);
+                } else if(*in < -1.0) {
+                    *target_event = CONDSWAPTOBUS32_CONST(0x40800001);
+                } else {
+                    float v = (*in) * AMDTP_FLOAT_MULTIPLIER;
+                    unsigned int tmp = ((int) v);
+                    tmp = ( tmp & 0x00FFFFFF ) | 0x40000000;
+                    *target_event = CondSwapToBus32((quadlet_t)tmp);
+                }
+#else
                 float v = (*in) * AMDTP_FLOAT_MULTIPLIER;
-                unsigned int tmp = ((int) lrintf(v));
-
-                tmp = ( tmp >> 8 ) | 0x40000000;
+                unsigned int tmp = ((int) v);
+                tmp = ( tmp & 0x00FFFFFF ) | 0x40000000;
                 *target_event = CondSwapToBus32((quadlet_t)tmp);
+#endif
                 buffer++;
                 target_event += m_dimension;
             }