OpenCores
URL https://opencores.org/ocsvn/oc-h264-encoder/oc-h264-encoder/trunk

Subversion Repositories oc-h264-encoder

[/] [oc-h264-encoder/] [trunk/] [x264/] [patches/] [x264-e381f6d-or32-or1ksim-with-fp-1.0.patch] - Rev 49

Compare with Previous | Blame | View Log

diff --exclude=.git --exclude=.gitignore -Naur ../x264/common/common.c x264/common/common.c
--- ../x264/common/common.c	2009-10-25 17:41:22.000000000 +0100
+++ x264/common/common.c	2009-10-27 16:45:55.000000000 +0100
@@ -764,8 +764,10 @@
     align_buf = malloc( i_size );
 #elif defined( HAVE_MALLOC_H )
     align_buf = memalign( 16, i_size );
+    //fprintf(stderr, "memalign, result addr: 0x%.8x\n", (unsigned int) align_buf);
 #else
     uint8_t *buf = malloc( i_size + 15 + sizeof(void **) + sizeof(int) );
+    //fprintf(stderr, "malloc, result addr: 0x%.8x\n", (unsigned int) buf);
     if( buf )
     {
         align_buf = buf + 15 + sizeof(void **) + sizeof(int);
diff --exclude=.git --exclude=.gitignore -Naur ../x264/common/common.h x264/common/common.h
--- ../x264/common/common.h	2009-10-25 17:41:22.000000000 +0100
+++ x264/common/common.h	2009-10-27 16:45:55.000000000 +0100
@@ -51,7 +51,8 @@
 #define X264_BFRAME_MAX 16
 #define X264_THREAD_MAX 128
 #define X264_PCM_COST (386*8)
-#define X264_LOOKAHEAD_MAX 250
+//#define X264_LOOKAHEAD_MAX 250
+#define X264_LOOKAHEAD_MAX 140
 // arbitrary, but low because SATD scores are 1/4 normal
 #define X264_LOOKAHEAD_QP 12
 
diff --exclude=.git --exclude=.gitignore -Naur ../x264/common/frame.c x264/common/frame.c
--- ../x264/common/frame.c	2009-10-25 17:41:10.000000000 +0100
+++ x264/common/frame.c	2009-10-27 16:45:55.000000000 +0100
@@ -31,6 +31,8 @@
     x264_frame_t *frame;
     int i, j;
 
+    //V(fprintf(stderr, "x264_frame_new\n"));
+
     int i_mb_count = h->mb.i_mb_count;
     int i_stride, i_width, i_lines;
     int i_padv = PADV << h->param.b_interlaced;
@@ -981,11 +983,14 @@
 
 x264_frame_t *x264_frame_pop_unused( x264_t *h, int b_fdec )
 {
+  //V(fprintf(stderr, "x264_frame_pop_unused\n"));
     x264_frame_t *frame;
     if( h->frames.unused[b_fdec][0] )
         frame = x264_frame_pop( h->frames.unused[b_fdec] );
     else
         frame = x264_frame_new( h, b_fdec );
+
+    //V(fprintf(stderr, "x264_frame_pop_unused: frame ptr = 0x%.8x\n", (unsigned long) frame));
     if( !frame )
         return NULL;
     frame->b_last_minigop_bframe = 0;
diff --exclude=.git --exclude=.gitignore -Naur ../x264/common/mdate.c x264/common/mdate.c
--- ../x264/common/mdate.c	2009-10-25 17:41:10.000000000 +0100
+++ x264/common/mdate.c	2009-10-27 16:45:55.000000000 +0100
@@ -32,9 +32,12 @@
 int64_t x264_mdate( void )
 {
 #ifndef __MINGW32__
+  /* --jb
     struct timeval tv_date;
     gettimeofday( &tv_date, NULL );
     return( (int64_t) tv_date.tv_sec * 1000000 + (int64_t) tv_date.tv_usec );
+  */
+  return 0;
 #else
     struct _timeb tb;
     _ftime(&tb);
diff --exclude=.git --exclude=.gitignore -Naur ../x264/common/pixel.c x264/common/pixel.c
--- ../x264/common/pixel.c	2009-10-25 17:41:10.000000000 +0100
+++ x264/common/pixel.c	2009-10-27 16:45:55.000000000 +0100
@@ -22,6 +22,7 @@
  *****************************************************************************/
 
 #include "common.h"
+#include "x264.h"
 
 #ifdef HAVE_MMX
 #   include "x86/pixel.h"
@@ -40,6 +41,7 @@
 /****************************************************************************
  * pixel_sad_WxH
  ****************************************************************************/
+/*  V(fprintf(stderr, "pixel_sad %d x %d, *pix1: 0x%.8x *pix2: 0x%.8x i_stride_pix1: %d i_stride_pix2: %d\n", lx, ly, (unsigned long) pix1, (unsigned long) pix2, i_stride_pix1, i_stride_pix2)); \*/
 #define PIXEL_SAD_C( name, lx, ly ) \
 static int name( uint8_t *pix1, int i_stride_pix1,  \
                  uint8_t *pix2, int i_stride_pix2 ) \
@@ -387,6 +389,7 @@
 
 /****************************************************************************
  * pixel_sad_x4
+ * Debug printf :   V(fprintf(stderr, "x264_pixel_sadx4 size : *fenc: 0x%.8x *pix0: 0x%.8x *pix1: 0x%.8x *pix2: 0x%.8x *pix3: 0x%.8x i_stride: %d  *scores: 0x%.8x\n", (unsigned long) fenc, (unsigned long) pix0, (unsigned long) pix1, (unsigned long)pix2, (unsigned long)pix3, i_stride, (unsigned long) scores)); \
  ****************************************************************************/
 #define SAD_X( size ) \
 static void x264_pixel_sad_x3_##size( uint8_t *fenc, uint8_t *pix0, uint8_t *pix1, uint8_t *pix2, int i_stride, int scores[3] )\
diff --exclude=.git --exclude=.gitignore -Naur ../x264/configure x264/configure
--- ../x264/configure	2009-10-25 17:41:10.000000000 +0100
+++ x264/configure	2009-10-27 16:45:55.000000000 +0100
@@ -29,7 +29,7 @@
     rm -f conftest.c
     [ -n "$1" ] && echo "#include <$1>" > conftest.c
     echo "int main () { $3 return 0; }" >> conftest.c
-    $CC conftest.c $CFLAGS $LDFLAGS $2 -o conftest 2>$DEVNULL
+    $CC conftest.c $CFLAGS $LDFLAGS $2 -o conftest #2>$DEVNULL
 }
 
 as_check() {
diff --exclude=.git --exclude=.gitignore -Naur ../x264/encoder/analyse.c x264/encoder/analyse.c
--- ../x264/encoder/analyse.c	2009-10-25 17:41:10.000000000 +0100
+++ x264/encoder/analyse.c	2009-10-27 16:45:55.000000000 +0100
@@ -27,6 +27,7 @@
 #include <unistd.h>
 
 #include "common/common.h"
+#include "x264.h"
 #include "common/cpu.h"
 #include "macroblock.h"
 #include "me.h"
@@ -34,6 +35,7 @@
 #include "analyse.h"
 #include "rdo.c"
 
+
 typedef struct
 {
     /* 16x16 */
@@ -233,40 +235,76 @@
     5, 3, 3, 1
 };
 
+extern const float x264_analyse_init_log2_array[]; //jb
+
 static void x264_analyse_update_cache( x264_t *h, x264_mb_analysis_t *a );
 
 static uint16_t x264_cost_ref[92][3][33];
 static x264_pthread_mutex_t cost_ref_mutex = X264_PTHREAD_MUTEX_INITIALIZER;
 
+/*
+float log2f( float n )
+{
+// log(n)/log(2) is log2.
+//return log10( n ) / log10( 2 );
+  return (float) (log( n ) / _M_LOG2_E);
+} 
+*/
 int x264_analyse_init_costs( x264_t *h, int qp )
 {
     int i, j;
     int lambda = x264_lambda_tab[qp];
+    float tmp_f,lambda_f = lambda;
+
     if( h->cost_mv[lambda] )
         return 0;
+
+    fprintf( stderr, 
+	     "x264_analyse_init_costs: lambda (= x264_lambda_tab[%d])  = %d\n",
+	     qp, lambda);
+
     /* factor of 4 from qpel, 2 from sign, and 2 because mv can be opposite from mvp */
     CHECKED_MALLOC( h->cost_mv[lambda], (4*4*2048 + 1) * sizeof(uint16_t) );
+    
     h->cost_mv[lambda] += 2*4*2048;
+
+    /* Computationally expensive thing: logs, floating point math, much of -- jb*/
+    //fprintf(stderr, "x264_analyse_init_costs: beginning loop of 2*4*2048");
+
     for( i = 0; i <= 2*4*2048; i++ )
-    {
-        h->cost_mv[lambda][-i] =
-        h->cost_mv[lambda][i]  = lambda * (log2f(i+1)*2 + 0.718f + !!i) + .5f;
-    }
+      {
+	
+	  //h->cost_mv[lambda][i]  = lambda * (log2f(i+1)*2 + 0.718f + !!i) + .5f;
+	  //h->cost_mv[lambda][i]  = lambda * (log2(i+1)*2 + 0.718f + !!i) + .5f;
+
+	// precalulated array, doing the above few calculations for us
+	tmp_f = lambda_f * x264_analyse_init_log2_array[i];
+	tmp_f += .5f;
+	h->cost_mv[lambda][-i] =
+	  h->cost_mv[lambda][i]  = tmp_f;
+	
+	
+      }
     x264_pthread_mutex_lock( &cost_ref_mutex );
+    //fprintf(stderr, "\tfinished\n");
+    /* -- for now jb
     for( i = 0; i < 3; i++ )
-        for( j = 0; j < 33; j++ )
-            x264_cost_ref[lambda][i][j] = i ? lambda * bs_size_te( i, j ) : 0;
+      for( j = 0; j < 33; j++ )
+	x264_cost_ref[lambda][i][j] = i ? lambda * bs_size_te( i, j ) : 0;
+    
     x264_pthread_mutex_unlock( &cost_ref_mutex );
     if( h->param.analyse.i_me_method >= X264_ME_ESA && !h->cost_mv_fpel[lambda][0] )
-    {
+      {
+	printf("analyse_init_costs: h->param.analyse.i_me_method >= X264_ME_ESA && !h->cost_mv_fpel[lambda][0]\n");
         for( j=0; j<4; j++ )
-        {
+	  {
             CHECKED_MALLOC( h->cost_mv_fpel[lambda][j], (4*2048 + 1) * sizeof(uint16_t) );
             h->cost_mv_fpel[lambda][j] += 2*2048;
             for( i = -2*2048; i < 2*2048; i++ )
-                h->cost_mv_fpel[lambda][j][i] = h->cost_mv[lambda][i*4+j];
+	      h->cost_mv_fpel[lambda][j][i] = h->cost_mv[lambda][i*4+j];
         }
     }
+    */
     return 0;
 fail:
     return -1;
@@ -288,6 +326,7 @@
 /* initialize an array of lambda*nbits for all possible mvs */
 static void x264_mb_analyse_load_costs( x264_t *h, x264_mb_analysis_t *a )
 {
+  //V(fprintf(stderr, "x264_mb_analyse_load_costs() called\n"));
     a->p_cost_mv = h->cost_mv[a->i_lambda];
     a->p_cost_ref0 = x264_cost_ref[a->i_lambda][x264_clip3(h->sh.i_num_ref_idx_l0_active-1,0,2)];
     a->p_cost_ref1 = x264_cost_ref[a->i_lambda][x264_clip3(h->sh.i_num_ref_idx_l1_active-1,0,2)];
diff --exclude=.git --exclude=.gitignore -Naur ../x264/encoder/analyse_gen_init_array.sh x264/encoder/analyse_gen_init_array.sh
--- ../x264/encoder/analyse_gen_init_array.sh	1970-01-01 01:00:00.000000000 +0100
+++ x264/encoder/analyse_gen_init_array.sh	2009-10-27 16:45:55.000000000 +0100
@@ -0,0 +1,32 @@
+#!/bin/bash
+# Pre-generate an array of numbers according to the algorithm from analyse_init_costs()
+# to speedup calculation. Here we calculate the stuff in brackets:
+# lambda * (log2f(i+1)*2 + 0.718f + !!i) + .5f; 
+outfile=analyse_init_log2.c
+
+echo; echo "Generating analyse_init_log2_array[] to $outfile" 
+echo "This may take a minute or two..."
+
+#Size of the array: 4*2*2048
+numvals=16384
+
+rm  -f $outfile
+
+echo "const float x264_analyse_init_log2_array[] = {" >> $outfile
+
+n=0
+
+    for num in `seq 1 $numvals`; 
+    do 
+# Use bc to do the calculation, scale down to the output to one decimal place
+	newnum=`echo "a=((l($num)/l(2))*2)+1.718; scale=1; a=(a*2)/2; a" | bc -l`
+	echo -n $newnum >> $outfile
+	if [ $num -lt $numvals ]; then echo -n ", " >> $outfile; fi;
+	# Newline every 16 values
+	let "n = $num % 16"
+	if [ $n -eq 0 ]; then echo >> $outfile; fi;
+    done
+echo "};" >> $outfile
+
+echo "finished"
+echo
diff --exclude=.git --exclude=.gitignore -Naur ../x264/encoder/encoder.c x264/encoder/encoder.c
--- ../x264/encoder/encoder.c	2009-10-25 17:41:22.000000000 +0100
+++ x264/encoder/encoder.c	2009-10-27 16:45:55.000000000 +0100
@@ -62,6 +62,9 @@
 
 static void x264_frame_dump( x264_t *h )
 {
+  //printf("frame_dump()\n");
+  return;
+  
     FILE *f = fopen( h->param.psz_dump_yuv, "r+b" );
     int i, y;
     if( !f )
@@ -301,6 +304,7 @@
     if( ( h->param.b_cabac && (h->cabac.p_end - h->cabac.p < 2500) )
      || ( h->out.bs.p_end - h->out.bs.p < 2500 ) )
     {
+      //V(fprintf(stderr, "x264_bitstream_check_buffer: h->out.bs.p = 0x%.8x, h->out.bs.p_end = 0x%.8x\n",h->out.bs.p_end, h->out.bs.p ));
         intptr_t delta;
         int i;
 
@@ -695,6 +699,8 @@
 static void mbcmp_init( x264_t *h )
 {
     int satd = !h->mb.b_lossless && h->param.analyse.i_subpel_refine > 1;
+    //V(if (satd)fprintf(stderr, "macroblock compare functions. satd enabled\n"));
+    //V(if (!satd)fprintf(stderr, "macroblock compare functions. satd not enabled\n"));
     memcpy( h->pixf.mbcmp, satd ? h->pixf.satd : h->pixf.sad_aligned, sizeof(h->pixf.mbcmp) );
     memcpy( h->pixf.mbcmp_unaligned, satd ? h->pixf.satd : h->pixf.sad, sizeof(h->pixf.mbcmp_unaligned) );
     h->pixf.intra_mbcmp_x3_16x16 = satd ? h->pixf.intra_satd_x3_16x16 : h->pixf.intra_sad_x3_16x16;
@@ -750,15 +756,14 @@
     x264_t *h;
     char buf[1000], *p;
     int i, qp, i_slicetype_length;
-
+    
     CHECKED_MALLOCZERO( h, sizeof(x264_t) );
-
+    
     /* Create a copy of param */
     memcpy( &h->param, param, sizeof(x264_param_t) );
-
     if( param->param_free )
         param->param_free( param );
-
+    
     if( x264_validate_parameters( h ) < 0 )
         goto fail;
 
@@ -772,7 +777,7 @@
         h->param.rc.psz_stat_in = strdup( h->param.rc.psz_stat_in );
 
     x264_set_aspect_ratio( h, param, 1 );
-
+    
     x264_reduce_fraction( &h->param.i_fps_num, &h->param.i_fps_den );
 
     /* Init x264_t */
@@ -831,8 +836,8 @@
     h->i_ref0 = 0;
     h->i_ref1 = 0;
 
-    x264_rdo_init();
-
+    x264_rdo_init(); 
+    
     /* init CPU functions */
     x264_predict_16x16_init( h->param.cpu, h->predict_16x16 );
     x264_predict_8x8c_init( h->param.cpu, h->predict_8x8c );
@@ -870,17 +875,25 @@
         p += sprintf( p, " none!" );
     x264_log( h, X264_LOG_INFO, "%s\n", buf );
 
+    /* // should skip during cycle accurate -- jb    */
+    // Check the different quantisation points and their costs
     for( qp = h->param.rc.i_qp_min; qp <= h->param.rc.i_qp_max; qp++ )
+      {
+	//printf("%d\n", qp);
         if( x264_analyse_init_costs( h, qp ) )
-            goto fail;
+	  goto fail;
+      }
     if( x264_analyse_init_costs( h, X264_LOOKAHEAD_QP ) )
         goto fail;
     if( h->cost_mv[1][2013] != 24 )
-    {
+      {
+	//fprintf(stderr, "h->cost_mv[1][2013] = %d\n",h->cost_mv[1][2013]);
         x264_log( h, X264_LOG_ERROR, "MV cost test failed: x264 has been miscompiled!\n" );
         goto fail;
     }
 
+    //exit(0); // jb
+    
     h->out.i_nal = 0;
     h->out.i_bitstream = X264_MAX( 1000000, h->param.i_width * h->param.i_height * 4
         * ( h->param.rc.i_rc_method == X264_RC_ABR ? pow( 0.95, h->param.rc.i_qp_min )
@@ -908,16 +921,17 @@
         if( x264_macroblock_cache_init( h->thread[i] ) < 0 )
             goto fail;
     }
-
+    
     if( x264_lookahead_init( h, i_slicetype_length ) )
         goto fail;
 
     if( x264_ratecontrol_new( h ) < 0 )
         goto fail;
-
+    
+    /* // skip this for now, it's a fopen() call we won't really do, plus it's an option we'll never need --jb
     if( h->param.psz_dump_yuv )
     {
-        /* create or truncate the reconstructed video file */
+    //create or truncate the reconstructed video file
         FILE *f = fopen( h->param.psz_dump_yuv, "w" );
         if( f )
             fclose( f );
@@ -927,6 +941,8 @@
             goto fail;
         }
     }
+    */
+
 
     x264_log( h, X264_LOG_INFO, "profile %s, level %d.%d\n",
         h->sps->i_profile_idc == PROFILE_BASELINE ? "Baseline" :
@@ -1307,6 +1323,8 @@
     int slice_max_size = h->param.i_slice_max_size > 0 ? (h->param.i_slice_max_size-3-NALU_OVERHEAD)*8 : INT_MAX;
     int starting_bits = bs_pos(&h->out.bs);
 
+    //V(fprintf(stderr, "x264_slice_write\n"));
+    
     /* Slice */
     x264_nal_start( h, h->i_nal_type, h->i_nal_ref_idc );
 
@@ -1320,6 +1338,7 @@
         /* init cabac */
         x264_cabac_context_init( &h->cabac, h->sh.i_type, h->sh.i_qp, h->sh.i_cabac_init_idc );
         x264_cabac_encode_init ( &h->cabac, h->out.bs.p, h->out.bs.p_end );
+	//V(fprintf(stderr, "\tcabac inited\n"));
     }
     h->mb.i_last_qp = h->sh.i_qp;
     h->mb.i_last_dqp = 0;
@@ -1362,7 +1381,10 @@
         x264_macroblock_encode( h );
 
         if( x264_bitstream_check_buffer( h ) )
+	  {
+	    //V(fprintf(stderr, "\tx264_bitstream_check_buffer failed\n"));
             return -1;
+	  }
 
         if( h->param.b_cabac )
         {
@@ -1522,6 +1544,8 @@
         x264_fdec_filter_row( h, h->sps->i_mb_height );
     }
 
+    //V(fprintf(stderr, "x264_slice_write ok\n"));
+
     return 0;
 }
 
@@ -1625,8 +1649,8 @@
     int     i_nal_type, i;
     int     i_nal_ref_idc;
 
-    int   i_global_qp;
 
+    int   i_global_qp;
     if( h->param.i_threads > 1)
     {
         int i = ++h->i_thread_phase;
@@ -1653,6 +1677,7 @@
     /* no data out */
     *pi_nal = 0;
     *pp_nal = NULL;
+    
 
     /* ------------------- Setup new frame from picture -------------------- */
     if( pic_in != NULL )
@@ -1663,7 +1688,10 @@
             return -1;
 
         if( x264_frame_copy_picture( h, fenc, pic_in ) < 0 )
+	  {
+	    //V(fprintf(stderr, "x264_frame_copy_picture returned >0\n"));
             return -1;
+	  }
 
         if( h->param.i_width != 16 * h->sps->i_mb_width ||
             h->param.i_height != 16 * h->sps->i_mb_height )
@@ -1684,13 +1712,15 @@
 
         /* 2: Place the frame into the queue for its slice type decision */
         x264_lookahead_put_frame( h, fenc );
-
+	
+	//V(fprintf(stderr, "x264_encoder_encode setup new frame from picture\n"));
+	
         if( h->frames.i_input <= h->frames.i_delay + 1 - h->param.i_threads )
         {
             /* Nothing yet to encode, waiting for filling of buffers */
             pic_out->i_type = X264_TYPE_AUTO;
             return 0;
-        }
+        }	
     }
     else
     {
@@ -1873,7 +1903,10 @@
     }
     else
         if( (intptr_t)x264_slices_write( h ) )
+	  {
+	    //V(fprintf(stderr, "x264_encoder_encode(): x264_slices_write() failed\n"));
             return -1;
+	  }
 
     return x264_encoder_frame_end( thread_oldest, thread_current, pp_nal, pi_nal, pic_out );
 }
diff --exclude=.git --exclude=.gitignore -Naur ../x264/encoder/me.c x264/encoder/me.c
--- ../x264/encoder/me.c	2009-10-25 17:41:10.000000000 +0100
+++ x264/encoder/me.c	2009-10-27 16:45:55.000000000 +0100
@@ -182,6 +182,7 @@
     int omx, omy, pmx, pmy;
     uint8_t *p_fenc = m->p_fenc[0];
     uint8_t *p_fref = m->p_fref[0];
+    //V(fprintf(stderr, "s264_me_search_ref *p_fenc = 0x%.8x *p_fref = 0x%.8x\n", (unsigned long) p_fenc, (unsigned long) p_fref));
     ALIGNED_ARRAY_16( uint8_t, pix,[16*16] );
 
     int i, j;
@@ -197,7 +198,7 @@
 
     const uint16_t *p_cost_mvx = m->p_cost_mv - m->mvp[0];
     const uint16_t *p_cost_mvy = m->p_cost_mv - m->mvp[1];
-
+    //V(fprintf(stderr, "p_cost_mvx: 0x%.8x, ( m->p_cost_mv: 0x%.8x m->mvp[0]: 0x%.8x )p_cost_mvy: 0x%.8x (m->p_cost_mv: 0x%.8x  m->mvp[1]: 0x%.8x)\n", (unsigned long) p_cost_mvx, (unsigned long) m->p_cost_mv, (unsigned long) m->mvp[0], (unsigned long) p_cost_mvy, (unsigned long) m->p_cost_mv, (unsigned long) m->mvp[1]));
     bmx = x264_clip3( m->mvp[0], mv_x_min*4, mv_x_max*4 );
     bmy = x264_clip3( m->mvp[1], mv_y_min*4, mv_y_max*4 );
     pmx = ( bmx + 2 ) >> 2;
@@ -251,6 +252,7 @@
     {
     case X264_ME_DIA:
         /* diamond search, radius 1 */
+      //V(fprintf(stderr, "x264_me_search_ref; diamond search. *p_fenc = 0x%.8x *p_fref = 0x%.8x\n", (unsigned long) p_fenc, (unsigned long) p_fref));
         i = 0;
         bcost <<= 4;
         do
diff --exclude=.git --exclude=.gitignore -Naur ../x264/encoder/slicetype.c x264/encoder/slicetype.c
--- ../x264/encoder/slicetype.c	2009-10-25 17:41:22.000000000 +0100
+++ x264/encoder/slicetype.c	2009-10-27 16:45:55.000000000 +0100
@@ -909,7 +909,9 @@
 
     /* Restore frametypes for all frames that haven't actually been decided yet. */
     for( j = reset_start; j <= num_frames; j++ )
-        frames[j]->i_type = X264_TYPE_AUTO;
+      frames[j]->i_type = X264_TYPE_AUTO;
+    
+    return;
 }
 
 void x264_slicetype_decide( x264_t *h )
diff --exclude=.git --exclude=.gitignore -Naur ../x264/link.ld x264/link.ld
--- ../x264/link.ld	1970-01-01 01:00:00.000000000 +0100
+++ x264/link.ld	2009-10-27 16:45:55.000000000 +0100
@@ -0,0 +1,100 @@
+/*
+MEMORY
+        {
+        vectors : ORIGIN = 0x00000000, LENGTH = 0x00002000
+        flash   : ORIGIN = 0x04000000, LENGTH = 0x00200000
+        ram     : ORIGIN = 0x00002000, LENGTH = 0x001fe000
+        icm     : ORIGIN = 0x00800000, LENGTH = 0x00004000
+        }
+ */
+
+MEMORY
+	{
+/*
+	reset	 : ORIGIN = 0x00000000, LENGTH = 0x00000200
+	vectors  : ORIGIN = 0x00000200, LENGTH = 0x00001E00
+        text     : ORIGIN = 0x00002000, LENGTH = 0x000fe000
+	data     : ORIGIN = 0x00100000, LENGTH = 0x00fe0000
+	stack    : ORIGIN = 0x001fe000, LENGTH = 0x00010000
+*/
+	yuv_data : ORIGIN = 25M, LENGTH = 7M
+	}
+
+/*
+MEMORY
+        {
+        reset   : ORIGIN = 0xc0000000, LENGTH = 0x00000200
+        vectors : ORIGIN = 0xc0000200, LENGTH = 0x00001000
+        ram     : ORIGIN = 0xc0001200, LENGTH = 0x00FFED00
+        }
+*/
+SECTIONS
+{
+/*
+	.reset :
+        {
+        *(.reset)
+        } > reset
+
+
+ 
+	.vectors :
+        {
+        _vec_start = .;
+        *(.vectors)
+        _vec_end = .;
+        } > vectors
+
+        .text :
+        {
+        *(.text)
+        } > text
+
+      .rodata :
+        {
+        *(.rodata)
+        *(.rodata.*)
+        } > text
+ 
+     .icm :
+        {
+        _icm_start = .;
+        *(.icm)
+        _icm_end = .;
+        } > data
+
+     .data :
+        {
+        _dst_beg = .;
+        *(.data)
+        _dst_end = .;
+        } > data
+ 
+      .bss :
+        {
+        *(.bss)
+        } > data
+	
+       .heap :
+	{
+	_heap_start = .;
+	*(.heap)
+ 	_heap_end = .;
+	} > data
+
+      .stack (NOLOAD) :
+        {
+        *(.stack)
+        _src_addr = .;
+        } > data
+*/
+	.yuv_data :
+	{
+	_yuv_data_start = .;
+	__yuv_data_start = .;
+	*(.yuv_data)	
+	_yuv_data_end = .;
+	__yuv_data_end = .;
+	} > yuv_data
+
+}
diff --exclude=.git --exclude=.gitignore -Naur ../x264/Makefile x264/Makefile
--- ../x264/Makefile	2009-10-25 17:41:22.000000000 +0100
+++ x264/Makefile	2009-10-27 17:30:09.000000000 +0100
@@ -91,7 +91,54 @@
 $(SONAME): .depend $(OBJS) $(OBJASM)
 	$(CC) -shared -o $@ $(OBJS) $(OBJASM) $(SOFLAGS) $(LDFLAGS)
 
-x264$(EXE): $(OBJCLI) libx264.a 
+
+## OR32-specific build rules
+##
+## Essentially we create some sample YUV video data
+## and embed it in an ELF file, in a section called yuv_data.
+## In the linker script we then link this section to a place
+## in ram and tell x264 to look here when it wants data.
+## 
+## You can download sample h264 encoded CIF sequences
+## here: http://www.tkn.tu-berlin.de/research/evalvid/cif.html
+## 
+##
+GEN_FILES=encoder/analyse_init_log2.c
+VIDEO_DATA_FILE=yuv_data.elf
+H264_VIDEO_FILE ?= ../test-sequences/football_cif.264
+YUV_VIDEO_FILE ?= video_data.yuv
+# We will use a limit of about 5MB of video data to encode
+# For CIF resultion of 4:2:0 this is about 30 frames. It's
+# possible to convert a set amount of data with ffmpeg but
+# it's probably not wise to attempt to encode partial frames
+NUM_FRAMES=30
+
+$(H264_VIDEO_FILE):
+	@echo; echo;
+	@echo "\tNo sample video file to embed! Please edit the Makefile"
+	@echo "\tand set the variable H264_VIDEO_FILE to the location of "
+	@echo "\ta raw h264 bitstream file, or set this variable on the"
+	@echo "\tcommand line when calling make, eg.:"
+	@echo "\t\tH264_VIDEO_FILE=../football_cif.264 make"; echo
+	exit 1
+$(YUV_VIDEO_FILE): $(H264_VIDEO_FILE)
+# First convert it to raw YUV, only about 5MB large, so for cif take 30 frames
+	ffmpeg -i $(H264_VIDEO_FILE) -vframes $(NUM_FRAMES) $(YUV_VIDEO_FILE)
+$(VIDEO_DATA_FILE): $(YUV_VIDEO_FILE)
+	or32-elf-ld -r -b binary -o yuv_data.o $(YUV_VIDEO_FILE)
+	or32-elf-objcopy --rename-section .data=.yuv_data yuv_data.o
+	mv yuv_data.o $(VIDEO_DATA_FILE)
+encoder/analyse_init_log2.c: encoder/analyse_gen_init_array.sh
+	cd encoder && chmod a+x analyse_gen_init_array.sh && ./analyse_gen_init_array.sh
+#OR32_DEPS= reset.o except.o uart.o syscalls.o $(VIDEO_DATA_FILE)
+
+OR32_DEPS= $(VIDEO_DATA_FILE) $(GEN_FILES)
+
+sim: x264$(EXE)
+	or32-elf-sim -f or1ksim_x264.cfg $<
+
+
+x264$(EXE): $(OR32_DEPS) $(OBJCLI) libx264.a
 	$(CC) -o $@ $+ $(LDFLAGS)
 
 checkasm: tools/checkasm.o libx264.a
@@ -147,12 +194,13 @@
 endif
 
 clean:
-	rm -f $(OBJS) $(OBJASM) $(OBJCLI) $(SONAME) *.a x264 x264.exe .depend TAGS
+	rm -f $(OBJS) $(OBJASM) $(OBJCLI) $(SONAME) *.a x264 x264.exe .depend TAGS $(YUV_VIDEO_FILE)
 	rm -f checkasm checkasm.exe tools/checkasm.o tools/checkasm-a.o
 	rm -f $(SRC2:%.c=%.gcda) $(SRC2:%.c=%.gcno)
 	- sed -e 's/ *-fprofile-\(generate\|use\)//g' config.mak > config.mak2 && mv config.mak2 config.mak
 
 distclean: clean
+	rm -f $(OR32_DEPS) uart0* $(GEN_FILES)
 	rm -f config.mak config.h x264.pc
 	rm -rf test/
 
diff --exclude=.git --exclude=.gitignore -Naur ../x264/muxers.c x264/muxers.c
--- ../x264/muxers.c	2009-10-25 17:41:22.000000000 +0100
+++ x264/muxers.c	2009-10-27 16:45:55.000000000 +0100
@@ -52,15 +52,32 @@
 
 typedef struct
 {
-    FILE *fh;
-    int width, height;
-    int next_frame;
+  //FILE *fh;
+  char *fh;
+  
+  int width, height;
+  int next_frame;
+  
 } yuv_input_t;
 
+void init_yuv_dataspace(char* yuv_dat_addr, hnd_t *p_handle, x264_param_t *p_param)
+{
+  yuv_input_t *h = malloc( sizeof(yuv_input_t) );
+  V(fprintf( stderr, "init_yuv_dataspace: h addr = 0x%.8x\n", h));
+  h->fh = yuv_dat_addr;
+  V(fprintf( stderr, "init_yuv_dataspace: h->fh 0x%.8x\n",h->fh));
+  h->width = p_param->i_width;
+  V(fprintf( stderr, "init_yuv_dataspace: h->width %d\n",h->width));
+  h->height = p_param->i_height;
+  V(fprintf( stderr, "init_yuv_dataspace: h->height %d\n",h->height));
+  h->next_frame = 0;
+  *p_handle = (hnd_t *)h;
+}
+
 /* raw 420 yuv file operation */
 int open_file_yuv( char *psz_filename, hnd_t *p_handle, x264_param_t *p_param )
 {
-    yuv_input_t *h = malloc( sizeof(yuv_input_t) );
+  yuv_input_t *h = malloc( sizeof(yuv_input_t) );
     if( !h )
         return -1;
     h->width = p_param->i_width;
@@ -80,7 +97,10 @@
 
 int get_frame_total_yuv( hnd_t handle )
 {
-    yuv_input_t *h = handle;
+      yuv_input_t *h = handle;
+
+      V(fprintf(stderr, "get_frame_total_yuv args: handle: 0x%.8x, h: 0x%.8x\n", handle, h));
+      /*
     int i_frame_total = 0;
 
     if( !fseek( h->fh, 0, SEEK_END ) )
@@ -88,24 +108,45 @@
         uint64_t i_size = ftell( h->fh );
         fseek( h->fh, 0, SEEK_SET );
         i_frame_total = (int)(i_size / ( h->width * h->height * 3 / 2 ));
-    }
+	}*/
+  
+#ifdef      USE_HARDCODED_FRAME_NUM
+      return (int)HARDCODED_FRAME_NUM;
+#else
+      fprintf( stderr, "get_frame_total_yuv: %d %d %d %d\n",
+		 h->width, h->height,
+		 YUV_DATA_SIZE,
+		 (int)(YUV_DATA_SIZE / ( h->width * h->height * 3 / 2 )));
+      return (int)(YUV_DATA_SIZE / ( h->width * h->height * 3 / 2 ));//i_frame_total;
+
+#endif
 
-    return i_frame_total;
 }
 
 int read_frame_yuv( x264_picture_t *p_pic, hnd_t handle, int i_frame )
 {
     yuv_input_t *h = handle;
-
+    int i;
+    
     if( i_frame != h->next_frame )
-        if( fseek( h->fh, (uint64_t)i_frame * h->width * h->height * 3 / 2, SEEK_SET ) )
-            return -1;
-
-    if( fread( p_pic->img.plane[0], 1, h->width * h->height, h->fh ) <= 0
-     || fread( p_pic->img.plane[1], 1, h->width * h->height / 4, h->fh ) <= 0
-     || fread( p_pic->img.plane[2], 1, h->width * h->height / 4, h->fh ) <= 0 )
-        return -1;
+      //if( fseek( h->fh, (uint64_t)i_frame * h->width * h->height * 3 / 2, SEEK_SET ) )
+      // Progress pointer to beginning of this frame
+      h->fh = (void*) YUV_DATA_ADDR + i_frame * h->width * h->height * 3 / 2;
+      //return -1;
+
+    //if( fread( p_pic->img.plane[0], 1, h->width * h->height, h->fh ) <= 0
+    for (i=0;i<h->width * h->height;i++) p_pic->img.plane[0][i] = h->fh[i];
+    h->fh += h->width * h->height;
+    //|| fread( p_pic->img.plane[1], 1, h->width * h->height / 4, h->fh ) <= 0
+    for (i=0;i<(h->width * h->height) / 4;i++) p_pic->img.plane[1][i] = h->fh[i];
+    h->fh += (h->width * h->height) / 4;
+    
+    //|| fread( p_pic->img.plane[2], 1, h->width * h->height / 4, h->fh ) <= 0 )
+    for (i=0;i<(h->width * h->height) / 4;i++) p_pic->img.plane[2][i] = h->fh[i];
+    h->fh += (h->width * h->height) / 4;
 
+    //return -1;
+    
     h->next_frame = i_frame+1;
 
     return 0;
@@ -113,12 +154,15 @@
 
 int close_file_yuv(hnd_t handle)
 {
+  return 0;
+  /* // -- jb
     yuv_input_t *h = handle;
     if( !h || !h->fh )
         return 0;
     fclose( h->fh );
     free( h );
     return 0;
+  */
 }
 
 /* YUV4MPEG2 raw 420 yuv file operation */
@@ -530,11 +574,21 @@
 }
 #endif
 
+#ifdef ENC_OUT_ADDR
+char * enc_out_pos;
+#endif
 
 int open_file_bsf( char *psz_filename, hnd_t *p_handle )
 {
+#ifdef ENC_OUT_ADDR
+  V(fprintf(stderr, "open_file_bsf\n"));
+  // Use a hardcoded memory location for outputting encoded data to
+  p_handle = (void *) ENC_OUT_ADDR;
+  enc_out_pos = (char *) ENC_OUT_ADDR;
+#else
     if( !(*p_handle = fopen(psz_filename, "w+b")) )
         return -1;
+#endif
 
     return 0;
 }
@@ -546,9 +600,16 @@
 
 int write_nalu_bsf( hnd_t handle, uint8_t *p_nalu, int i_size )
 {
+#ifdef ENC_OUT_ADDR
+  V(fprintf(stderr, "write_nalu_bsf: writing %d bytes from 0x%.8x\n", i_size, (unsigned long) enc_out_pos));
+  // Just write to the spot in memory and increment the pointer
+  int i; for (i=0;i<i_size;i++) *enc_out_pos++ = p_nalu[i];
+  return i_size;
+#else
     if( fwrite( p_nalu, i_size, 1, (FILE*)handle ) > 0 )
         return i_size;
     return -1;
+#endif
 }
 
 int set_eop_bsf( hnd_t handle,  x264_picture_t *p_picture )
@@ -558,6 +619,9 @@
 
 int close_file_bsf( hnd_t handle )
 {
+#ifdef ENC_OUT_ADDR
+  return 0;
+#endif
     if( !handle || handle == stdout )
         return 0;
 
diff --exclude=.git --exclude=.gitignore -Naur ../x264/muxers.h x264/muxers.h
--- ../x264/muxers.h	2009-10-25 17:41:22.000000000 +0100
+++ x264/muxers.h	2009-10-27 16:45:55.000000000 +0100
@@ -27,6 +27,7 @@
 typedef void *hnd_t;
 
 int open_file_yuv( char *psz_filename, hnd_t *p_handle, x264_param_t *p_param );
+void init_yuv_dataspace(char* yuv_dat_addr, hnd_t *p_handle, x264_param_t *p_param);
 int get_frame_total_yuv( hnd_t handle );
 int read_frame_yuv( x264_picture_t *p_pic, hnd_t handle, int i_frame );
 int close_file_yuv( hnd_t handle );
diff --exclude=.git --exclude=.gitignore -Naur ../x264/or1ksim_x264.cfg x264/or1ksim_x264.cfg
--- ../x264/or1ksim_x264.cfg	1970-01-01 01:00:00.000000000 +0100
+++ x264/or1ksim_x264.cfg	2009-10-27 16:45:55.000000000 +0100
@@ -0,0 +1,886 @@
+/* sim.cfg -- Simulator configuration script file
+   Copyright (C) 2001-2002, Marko Mlinar, markom@opencores.org
+
+This file is part of OpenRISC 1000 Architectural Simulator.
+It contains the default configuration and help about configuring
+the simulator.
+
+This program is free software; you can redistribute it and/or modify
+it under the terms of the GNU General Public License as published by
+the Free Software Foundation; either version 2 of the License, or
+(at your option) any later version.
+
+This program is distributed in the hope that it will be useful,
+but WITHOUT ANY WARRANTY; without even the implied warranty of
+MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+GNU General Public License for more details.
+
+You should have received a copy of the GNU General Public License
+along with this program; if not, write to the Free Software
+Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. */
+
+
+/* INTRODUCTION
+
+   The ork1sim has various parameters, that are set in configuration files
+   like this one. The user can switch between configurations at startup by
+   specifying the required configuration file with the -f <filename.cfg> option.
+   If no configuration file is specified or1ksim searches for the default
+   configuration file sim.cfg. First it searches for './sim.cfg'. If this
+   file is not found, it searches for '~/or1k/sim.cfg'. If this file is
+   not found too, it reverts to the built-in default configuration.
+   
+   NOTE: Users should not rely on the built-in configuration, since the
+         default configuration may differ between version.
+         Rather create a configuration file that sets all critical values.
+
+   This file may contain (standard C) comments only - no // support.
+   
+   Configure files may be be included, using:
+   include "file_name_to_include"
+
+   Like normal configuration files, the included file is divided into
+   sections. Each section is described in detail also.
+
+   Some section have subsections. One example of such a subsection is:
+
+   device <index>
+     instance specific parameters...
+   enddevice
+
+   which creates a device instance.
+*/
+
+
+/* MEMORY SECTION
+
+   This section specifies how the memory is generated and the blocks
+   it consists of.
+
+   type = random/unknown/pattern
+      Specifies the initial memory values.
+      'random' generates random memory using seed 'random_seed'.
+      'pattern' fills memory with 'pattern'.
+      'unknown' does not specify how memory should be generated,
+      leaving the memory in a undefined state. This is the fastest
+      option.
+
+   random_seed = <value>
+      random seed for randomizer, used if type = 'random'.
+
+   pattern = <value>
+      pattern to fill memory, used if type = 'pattern'.
+
+   nmemories = <value>
+      number of memory instances connected
+
+   baseaddr = <hex_value>
+      memory start address
+
+   size = <hex_value>
+      memory size
+
+   name = "<string>"
+      memory block name
+
+   ce = <value>
+      chip enable index of the memory instance
+
+   mc = <value>
+      memory controller this memory is connected to
+
+   delayr = <value>
+      cycles, required for read access, -1 if instance does not support reading
+
+   delayw = <value>
+      cycles, required for write access, -1 if instance does not support writing
+
+   log = "<filename>"
+      filename, where to log memory accesses to, no log, if log command is not specified
+*/
+
+
+section memory
+  /*random_seed = 12345
+  type = random*/
+  pattern = 0x00
+  type = unknown /* Fastest */
+
+  name = "FLASH"
+  ce = 0
+  mc = 0
+  baseaddr = 0xf0000000
+  size = 0x01000000
+  delayr = 10
+  delayw = -1
+end
+
+section memory
+  /*random_seed = 12345
+  type = random*/
+  pattern = 0x00
+  type = unknown /* Fastest */
+
+  name = "RAM"
+  ce = 1
+  mc = 0
+  baseaddr = 0x00000000
+  size = 0x02000000
+  delayr = 20
+  delayw = 25
+end
+
+section memory
+  /*random_seed = 12345
+  type = random*/
+  pattern = 0x00
+  type = unknown /* Fastest */
+
+  name = "SRAM"
+  mc = 0
+  ce = 2
+  baseaddr = 0xa4000000
+  size = 0x00100000
+  delayr = 1
+  delayw = 2
+end
+
+
+/* IMMU SECTION
+
+    This section configures the Instruction Memory Manangement Unit
+
+    enabled = 0/1
+       '0': disabled
+       '1': enabled
+       (NOTE: UPR bit is set)
+
+    nsets = <value>
+       number of ITLB sets; must be power of two
+
+    nways = <value>
+       number of ITLB ways
+
+    pagesize = <value>
+       instruction page size; must be power of two
+
+    entrysize = <value>
+       instruction entry size in bytes
+
+    ustates = <value>
+       number of ITLB usage states (2, 3, 4 etc., max is 4)
+       
+    hitdelay = <value>
+       number of cycles immu hit costs
+    
+    missdelay = <value>
+       number of cycles immu miss costs
+*/
+
+section immu
+  enabled = 1
+  nsets = 64
+  nways = 1
+  pagesize = 8192
+  hitdelay = 0
+  missdelay = 0
+end
+
+
+/* DMMU SECTION
+
+    This section configures the Data Memory Manangement Unit
+
+    enabled = 0/1
+       '0': disabled
+       '1': enabled
+       (NOTE: UPR bit is set)
+
+    nsets = <value>
+       number of DTLB sets; must be power of two
+
+    nways = <value>
+       number of DTLB ways
+
+    pagesize = <value>
+       data page size; must be power of two
+
+    entrysize = <value>
+       data entry size in bytes
+
+    ustates = <value>
+       number of DTLB usage states (2, 3, 4 etc., max is 4)
+
+    hitdelay = <value>
+       number of cycles dmmu hit costs
+
+    missdelay = <value>
+       number of cycles dmmu miss costs
+*/
+
+section dmmu
+  enabled = 1
+  nsets = 64
+  nways = 1
+  pagesize = 8192
+  hitdelay = 0
+  missdelay = 0
+end
+
+
+/* IC SECTION
+
+   This section configures the Instruction Cache
+
+   enabled = 0/1
+       '0': disabled
+       '1': enabled
+      (NOTE: UPR bit is set)
+
+   nsets = <value>
+      number of IC sets; must be power of two
+
+   nways = <value>
+      number of IC ways
+
+   blocksize = <value>
+      IC block size in bytes; must be power of two
+
+   ustates = <value>
+      number of IC usage states (2, 3, 4 etc., max is 4)
+
+   hitdelay = <value>
+      number of cycles ic hit costs
+    
+    missdelay = <value>
+      number of cycles ic miss costs
+*/
+
+section ic
+  enabled = 0
+  nsets = 512
+  nways = 1
+  blocksize = 16
+  hitdelay = 20
+  missdelay = 20
+end
+
+
+/* DC SECTION
+
+   This section configures the Data Cache
+
+   enabled = 0/1
+       '0': disabled
+       '1': enabled
+      (NOTE: UPR bit is set)
+
+   nsets = <value>
+      number of DC sets; must be power of two
+
+   nways = <value>
+      number of DC ways
+
+   blocksize = <value>
+      DC block size in bytes; must be power of two
+
+   ustates = <value>
+      number of DC usage states (2, 3, 4 etc., max is 4)
+
+   load_hitdelay = <value>
+      number of cycles dc load hit costs
+   
+   load_missdelay = <value>
+      number of cycles dc load miss costs
+       
+   store_hitdelay = <value>
+      number of cycles dc load hit costs
+    
+   store_missdelay = <value>
+      number of cycles dc load miss costs
+*/
+
+section dc
+  enabled = 0
+  nsets = 512
+  nways = 1
+  blocksize = 16
+  load_hitdelay = 20
+  load_missdelay = 20
+  store_hitdelay = 20
+  store_missdelay = 20
+end
+
+
+/* SIM SECTION
+
+  This section specifies how or1ksim should behave.
+
+  verbose = 0/1
+       '0': don't print extra messages
+       '1': print extra messages
+
+  debug = 0-9
+      0  : no debug messages
+      1-9: debug message level.
+           higher numbers produce more messages
+
+  profile = 0/1
+      '0': don't generate profiling file 'sim.profile'
+      '1': don't generate profiling file 'sim.profile'
+
+  prof_fn = "<filename>"
+      optional filename for the profiling file.
+      valid only if 'profile' is set
+      
+  mprofile = 0/1
+      '0': don't generate memory profiling file 'sim.mprofile'
+      '1': generate memory profiling file 'sim.mprofile'
+
+  mprof_fn = "<filename>"
+      optional filename for the memory profiling file.
+      valid only if 'mprofile' is set
+
+  history = 0/1
+      '0': don't track execution flow
+      '1': track execution flow
+      Execution flow can be tracked for the simulator's
+      'hist' command. Useful for back-trace debugging.
+
+  iprompt = 0/1
+     '0': start in <not interactive prompt> (so what do we start in ???)
+     '1': start in interactive prompt.
+
+  exe_log = 0/1
+      '0': don't generate execution log.
+      '1': generate execution log.
+      
+  exe_log = default/hardware/simple/software
+      type of execution log, default is used when not specified
+      
+  exe_log_start = <value>
+      index of first instruction to start logging, default = 0
+        
+  exe_log_end = <value>
+      index of last instruction to end logging; not limited, if omitted
+  
+  exe_log_marker = <value>
+      <value> specifies number of instructions before horizontal marker is
+      printed; if zero, markers are disabled (default)
+
+  exe_log_fn = "<filename>"
+      filename for the exection log file.
+      valid only if 'exe_log' is set
+
+  clkcycle = <value>[ps|ns|us|ms]
+      specifies time measurement for one cycle
+*/
+
+section sim
+  verbose = 1
+  debug = 0
+  profile = 0
+  history = 0
+
+  clkcycle = 10ns
+end
+
+
+/* SECTION VAPI
+
+    This section configures the Verification API, used for Advanced
+    Core Verification.
+
+    enabled = 0/1
+        '0': disbable VAPI server
+        '1': enable/start VAPI server
+
+    server_port = <value>
+        TCP/IP port to start VAPI server on
+
+    log_enabled = 0/1
+       '0': disable VAPI requests logging
+       '1': enable VAPI requests logging
+
+    hide_device_id = 0/1
+       '0': don't log device id (for compatability with old version)
+       '1': log device id
+       
+
+    vapi_fn = <filename>
+       filename for the log file.
+       valid only if log_enabled is set
+*/
+
+section VAPI
+  enabled = 0
+  server_port = 9998
+  log_enabled = 0
+  vapi_log_fn = "vapi.log"
+end
+
+
+/* CPU SECTION
+
+   This section specifies various CPU parameters.
+
+   ver = <value>
+   rev = <value>
+      specifies version and revision of the CPU used
+
+   upr = <value>
+      changes the upr register
+      
+   sr = <value>
+      sets the initial Supervision Register value
+
+   superscalar = 0/1
+      '0': CPU is scalar
+      '1': CPU is superscalar
+      (modify cpu/or32/execute.c to tune superscalar model)
+
+   hazards = 0/1
+      '0': don't track data hazards in superscalar CPU
+      '1': track data hazards in superscalar CPU
+      If tracked, data hazards can be displayed using the
+      simulator's 'r' command.
+
+   dependstats = 0/1
+      '0': don't calculate inter-instruction dependencies.
+      '1': calculate inter-instruction dependencies.
+      If calculated, inter-instruction dependencies can be
+      displayed using the simulator's 'stat' command.
+
+   sbuf_len = <value>
+      length of store buffer (<= 256), 0 = disabled
+*/
+
+section cpu
+  ver = 0x12
+  cfg = 0x00
+  rev = 0x01
+  /* upr = */
+  superscalar = 0
+  hazards = 0
+  dependstats = 0
+  sbuf_len = 0
+  hardfloat = 1
+end
+
+
+/* PM SECTION
+
+   This section specifies Power Management parameters
+
+   enabled = 0/1
+      '0': disable power management
+      '1': enable power management
+*/
+
+section pm
+  enabled = 0
+end
+
+
+/* BPB SECTION
+
+   This section specifies how branch prediction should behave.
+   
+   enabled = 0/1
+     '0': disable branch prediction
+     '1': enable branch prediction
+      
+   btic = 0/1
+     '0': disable branch target instruction cache model
+     '1': enable branch target instruction cache model
+
+   sbp_bf_fwd = 0/1
+     Static branch prediction for 'l.bf'
+     '0': don't use forward prediction
+     '1': use forward prediction 
+   
+   sbp_bnf_fwd = 0/1
+     Static branch prediction for 'l.bnf'
+     '0': don't use forward prediction
+     '1': use forward prediction 
+
+   hitdelay = <value>
+       number of cycles bpb hit costs
+    
+   missdelay = <value>
+       number of cycles bpb miss costs
+*/
+
+section bpb
+  enabled = 0
+  btic = 0
+  sbp_bf_fwd = 0
+  sbp_bnf_fwd = 0
+  hitdelay = 0
+  missdelay = 0
+end
+
+
+/* DEBUG SECTION
+
+   This sections specifies how the debug unit should behave.
+
+   enabled = 0/1
+      '0': disable debug unit
+      '1': enable debug unit
+
+   gdb_enabled = 0/1
+      '0': don't start gdb server
+      '1': start gdb server at port 'server_port'
+
+   server_port = <value>
+      TCP/IP port to start gdb server on
+      valid only if gdb_enabled is set
+
+   vapi_id = <hex_value>
+      Used to create "fake" vapi log file containing the JTAG proxy messages.
+*/
+section debug
+  enabled = 0
+  rsp_enabled = 0
+  rsp_port = 5554
+  /*server_port = 9999*/
+end
+
+
+/* MC SECTION
+
+   This section configures the memory controller
+
+   enabled = 0/1
+     '0': disable memory controller
+     '1': enable memory controller
+
+   baseaddr = <hex_value>
+      address of first MC register
+
+   POC = <hex_value>
+      Power On Configuration register
+
+   index = <value>
+      Index of this memory controller amongst all the memory controllers
+*/
+
+section mc
+  enabled = 0
+  baseaddr = 0x93000000
+  POC = 0x00000008                 /* Power on configuration register */
+  index = 0
+end
+
+
+/* UART SECTION
+
+   This section configures the UARTs
+
+     enabled = <0|1>
+        Enable/disable the peripheral.  By default if it is enabled.
+
+     baseaddr = <hex_value>
+        address of first UART register for this device
+
+     
+     channel = <channeltype>:<args>
+     
+        The channel parameter indicates the source of received UART characters
+        and the sink for transmitted UART characters.
+
+        The <channeltype> can be either "file", "xterm", "tcp", "fd", or "tty"
+        (without quotes).
+
+          A) To send/receive characters from a pair of files, use a file
+             channel:
+
+               channel=file:<rxfile>,<txfile>
+
+	  B) To create an interactive terminal window, use an xterm channel:
+
+               channel=xterm:[<xterm_arg>]*
+
+	  C) To create a bidirectional tcp socket which one could, for example,
+             access via telnet, use a tcp channel:
+
+               channel=tcp:<port number>
+
+	  D) To cause the UART to read/write from existing numeric file
+             descriptors, use an fd channel:
+
+               channel=fd:<rx file descriptor num>,<tx file descriptor num>
+
+          E) To connect the UART to a physical serial port, create a tty
+             channel:
+
+	       channel=tty:device=/dev/ttyS0,baud=9600
+
+     irq = <value>
+        irq number for this device
+
+     16550 = 0/1
+        '0': this device is a UART16450
+        '1': this device is a UART16550
+
+     jitter = <value>
+        in msecs... time to block, -1 to disable it
+
+     vapi_id = <hex_value>
+        VAPI id of this instance
+*/
+
+section uart
+  enabled = 1
+  baseaddr = 0x90000000
+  irq = 2
+  channel = "file:uart0.rx,uart0.tx"
+  /* channel = "tcp:10084" */
+  /* channel = "xterm:" */
+  jitter = -1                     /* async behaviour */
+  16550 = 1
+end
+
+
+/* DMA SECTION
+
+   This section configures the DMAs
+
+     enabled = <0|1>
+        Enable/disable the peripheral.  By default if it is enabled.
+
+     baseaddr = <hex_value>
+        address of first DMA register for this device
+
+     irq = <value>
+        irq number for this device
+
+     vapi_id = <hex_value>
+        VAPI id of this instance
+*/
+
+section dma
+  enabled = 1
+  baseaddr = 0x9a000000
+  irq = 11
+end
+
+
+/* ETHERNET SECTION
+
+   This section configures the ETHERNETs
+
+     enabled = <0|1>
+        Enable/disable the peripheral.  By default if it is enabled.
+
+     baseaddr = <hex_value>
+        address of first ethernet register for this device
+
+     dma = <value>
+        which controller is this ethernet "connected" to
+
+     irq = <value>
+        ethernet mac IRQ level
+
+     rtx_type = <value>
+        use 0 - file interface, 1 - socket interface
+
+     rx_channel = <value>
+        DMA channel used for RX
+
+     tx_channel = <value>
+        DMA channel used for TX
+
+     rxfile = "<filename>"
+        filename, where to read data from
+
+     txfile = "<filename>"
+        filename, where to write data to
+
+     sockif = "<ifacename>"
+        interface name of ethernet socket
+
+     vapi_id = <hex_value>
+        VAPI id of this instance
+*/
+
+section ethernet
+  enabled = 0
+  baseaddr = 0x92000000
+  /* dma = 0 */
+  irq = 4
+  rtx_type = 0
+  /* tx_channel = 0 */
+  /* rx_channel = 1 */
+  /*rxfile = "eth0.rx"*/
+  txfile = "eth0.tx"
+  sockif = "eth0"
+end
+
+
+/* GPIO SECTION
+
+   This section configures the GPIOs
+
+     enabled = <0|1>
+        Enable/disable the peripheral.  By default if it is enabled.
+
+     baseaddr = <hex_value>
+        address of first GPIO register for this device
+
+     irq = <value>
+        irq number for this device
+
+     base_vapi_id = <hex_value>
+        first VAPI id of this instance
+	GPIO uses 8 consecutive VAPI IDs
+*/
+
+section gpio
+  enabled = 0
+  baseaddr = 0x91000000
+  irq = 3
+  base_vapi_id = 0x0200
+end
+
+/* VGA SECTION
+    
+    This section configures the VGA/LCD controller
+    
+      enabled = <0|1>
+        Enable/disable the peripheral.  By default if it is enabled.
+
+      baseaddr = <hex_value>
+        address of first VGA register
+        
+      irq = <value>
+        irq number for this device
+        
+      refresh_rate = <value>
+        number of cycles between screen dumps
+        
+      filename = "<filename>"
+        template name for generated names (e.g. "primary" produces "primary0023.bmp")
+*/
+
+section vga
+  enabled = 1
+  baseaddr = 0x97100000
+  irq = 8
+  refresh_rate = 100000
+  filename = "primary"
+end
+
+
+/* TICK TIMER SECTION
+
+    This section configures tick timer
+
+    enabled = 0/1
+      whether tick timer is enabled
+*/
+
+section pic
+  enabled = 1
+  edge_trigger = 1
+end
+
+/* FB SECTION
+    
+    This section configures the frame buffer
+    
+    enabled = <0|1>
+      Enable/disable the peripheral.  By default if it is enabled.
+
+    baseaddr = <hex_value>
+      base address of frame buffer
+        
+    paladdr = <hex_value>
+      base address of first palette entry
+        
+    refresh_rate = <value>
+      number of cycles between screen dumps
+        
+    filename = "<filename>"
+      template name for generated names (e.g. "primary" produces "primary0023.bmp")
+*/
+
+section fb
+  enabled = 1
+  baseaddr = 0x97000000
+  refresh_rate = 1000000
+  filename = "primary"
+end
+
+
+/* KBD SECTION
+
+    This section configures the PS/2 compatible keyboard
+    
+    baseaddr = <hex_value>
+      base address of the keyboard device
+      
+    rxfile = "<filename>"
+      filename, where to read data from
+*/
+
+section kbd
+  enabled = 1
+  irq = 5
+  baseaddr = 0x94000000
+  rxfile = "kbd.rx"
+end
+
+
+/* ATA SECTION
+    
+    This section configures the ATA/ATAPI host controller
+    
+      baseaddr = <hex_value>
+        address of first ATA register
+        
+      enabled = <0|1>
+        Enable/disable the peripheral.  By default if it is enabled.
+
+      irq = <value>
+        irq number for this device
+
+      debug = <value>
+        debug level for ata models.
+	0: no debug messages
+	1: verbose messages
+	3: normal messages (more messages than verbose)
+        5: debug messages (normal debug messages)
+	7: flow control messages (debug statemachine flows)
+	9: low priority message (display everything the code does)
+
+      dev_type0/1 = <value>
+        ata device 0 type
+        0: NO_CONNeCT: none (not connected)
+	1: FILE      : simulated harddisk
+	2: LOCAL     : local system harddisk
+
+      dev_file0/1 = "<filename>"
+        filename for simulated ATA device
+	valid only if dev_type0 == 1
+
+      dev_size0/1 = <value>
+        size of simulated hard-disk (in MBytes)
+	valid only if dev_type0 == 1
+
+      dev_packet0/1 = <value>
+        0: simulated ATA device does NOT implement PACKET command feature set
+	1: simulated ATA device does implement PACKET command feature set
+
+   FIXME: irq number
+*/
+
+section ata
+  enabled = 0
+  baseaddr = 0x9e000000
+  irq = 15
+
+end
+
+
diff --exclude=.git --exclude=.gitignore -Naur ../x264/x264.c x264/x264.c
--- ../x264/x264.c	2009-10-25 17:41:22.000000000 +0100
+++ x264/x264.c	2009-10-27 16:45:55.000000000 +0100
@@ -40,6 +40,8 @@
 #define SetConsoleTitle(t)
 #endif
 
+//#include "uart.h"
+
 uint8_t *mux_buffer = NULL;
 int mux_buffer_size = 0;
 
@@ -54,11 +56,11 @@
 }
 
 typedef struct {
-    int b_progress;
-    int i_seek;
-    hnd_t hin;
-    hnd_t hout;
-    FILE *qpfile;
+  int b_progress;
+  int i_seek;
+  hnd_t hin; /* hnd_t is a void* */
+  hnd_t hout;
+  FILE *qpfile;
 } cli_opt_t;
 
 /* input file operation function pointers */
@@ -78,15 +80,28 @@
 static int  Parse( int argc, char **argv, x264_param_t *param, cli_opt_t *opt );
 static int  Encode( x264_param_t *param, cli_opt_t *opt );
 
-
+int yuv_data_size;
 /****************************************************************************
  * main:
  ****************************************************************************/
 int main( int argc, char **argv )
 {
-    x264_param_t param;
-    cli_opt_t opt;
-    int ret;
+  char* new_argv ="./x264 --profile=baseline";
+  argc = 2;
+
+  x264_param_t param;
+  cli_opt_t opt;
+  int ret;
+
+  //uart_init()  ;
+
+  extern int _end;   // start of free memory
+  extern int _stack; // end of free memory
+
+  V(fprintf( stderr, "start of heap at 0x%.8x, size of heap: %d bytes\n",
+	     (unsigned int) &_end, (unsigned int) ((unsigned int) &_stack) - ((unsigned int) &_end)));
+
+  //fprintf(stderr,"start\n");
 
 #ifdef PTW32_STATIC_LIB
     pthread_win32_process_attach_np();
@@ -100,14 +115,74 @@
 
     x264_param_default( &param );
 
-    /* Parse command line */
-    if( Parse( argc, argv, &param, &opt ) < 0 )
-        return -1;
+    // Baseline parameters
+     param.analyse.b_transform_8x8 = 0;
+     param.b_cabac = 0;
+     param.i_cqm_preset = X264_CQM_FLAT;
+     param.i_bframe = 0;
+
+     // Ultrafast preset
+     param.i_frame_reference = 1;
+     param.i_scenecut_threshold = 0;
+     param.b_deblocking_filter = 0;
+     param.b_cabac = 0;
+     param.i_bframe = 0;
+     param.analyse.intra = 0;
+     param.analyse.inter = 0;
+     param.analyse.b_transform_8x8 = 0;
+     param.analyse.i_me_method = X264_ME_DIA;
+     param.analyse.i_subpel_refine = 0;
+     param.rc.i_aq_mode = 0;
+     param.analyse.b_mixed_references = 0;
+     param.analyse.i_trellis = 0;
+     param.i_bframe_adaptive = X264_B_ADAPT_NONE;
+     param.rc.b_mb_tree = 0;
+
+     // Verbose (info per frame encoded)
+     param.i_log_level = X264_LOG_DEBUG;
+
+     // Set a bitrate
+     param.rc.i_bitrate = 1000; // kbps
+     param.rc.i_rc_method = X264_RC_ABR;
+
+     // A smaller range of QP (qp_max - qp_min) means less startup time for computing motion vectors costs (not sure exactly how this is done) --jb
+     // Defaults here were i_qp_min = 10, i_qp_max = 51
+     param.rc.i_qp_min = 10;
+     param.rc.i_qp_max = 51;
+     param.rc.i_qp_step = 4;
+
+    
+     // Picture height and width
+     // CIF = 352x288
+     param.i_width = 352;
+     param.i_height = 288;
+     V(fprintf( stderr, "x264 [info]: %dx%d @ %.2f fps\n", 
+		param.i_width, param.i_height, 
+		(float)param.i_fps_num / (float)param.i_fps_den));
+     
+     V(printf("video file in memory from 0x%.8x, size %d bytes\n", YUV_DATA_ADDR, YUV_DATA_SIZE));
+     /* Must define a few things for use of static CIF data
+	YUV_DATA_ADDR - address where the data starts
+	YUV_DATA_SIZE - size of data in bytes
+	ENC_OUT_ADDR  - where we'll store the encoded data
+     */
+
+     /* Parse command line */
+     Parse( argc, &new_argv, &param, &opt );
+     //if( Parse( argc, &new_argv, &param, &opt ) < 0 )
+     //return -1;
+
+     // We specify a place in memory of some YUV CIF data instad of opening a file
+     init_yuv_dataspace((char *)YUV_DATA_ADDR, (hnd_t *)&opt.hin, &param);
+#ifdef ENC_OUT_ADDR
+     p_open_outfile( 0, &opt.hout ); // Setup the out "file" which is really just a spot in memory
+#endif
+     //opt.hout = (void*) ENC_OUT_ADDR; // We specify a place in memory of where we'll dump the encoded data
 
-    /* Control-C handler */
-    signal( SIGINT, SigIntHandler );
+     /* Control-C handler */
+     //signal( SIGINT, SigIntHandler );
 
-    ret = Encode( &param, &opt );
+   ret = Encode( &param, &opt );
 
 #ifdef PTW32_STATIC_LIB
     pthread_win32_thread_detach_np();
@@ -555,10 +630,12 @@
     /* Default output file driver */
     p_open_outfile = open_file_bsf;
     p_set_outfile_param = set_param_bsf;
-    p_write_nalu = write_nalu_bsf;
+    p_write_nalu = write_nalu_bsf; //--jb
     p_set_eop = set_eop_bsf;
     p_close_outfile = close_file_bsf;
 
+    return 0; // -- added jb
+
     /* Presets are applied before all other options. */
     for( optind = 0;; )
     {
@@ -978,9 +1055,11 @@
         }
         else
         {
-            sscanf( argv[optind++], "%ux%u", &param->i_width, &param->i_height );
-            if( param->i_log_level >= X264_LOG_INFO )
-                fprintf( stderr, "x264 [info]: %dx%d @ %.2f fps\n", param->i_width, param->i_height, (double)param->i_fps_num / (double)param->i_fps_den);
+	  sscanf( argv[optind++], "%ux%u", &param->i_width, &param->i_height );
+	  if( param->i_log_level >= X264_LOG_INFO )
+	    fprintf( stderr, "x264 [info]: %dx%d @ %.2f fps\n", 
+		     param->i_width, param->i_height, 
+		     (double)param->i_fps_num / (double)param->i_fps_den);
         }
     }
 
@@ -1127,7 +1206,10 @@
     {
         i_nalu_size = p_write_nalu( hout, nal[i].p_payload, nal[i].i_payload );
         if( i_nalu_size < 0 )
+	  {
+	    fprintf(stderr, "Encode_frame: p_write_nalu() returned i_nalu_size < 0: %d\n", i_nalu_size);
             return -1;
+	  }
         i_file += i_nalu_size;
     }
     if (i_nal)
@@ -1172,12 +1254,15 @@
     opt->b_progress &= param->i_log_level < X264_LOG_DEBUG;
     i_frame_total = p_get_frame_total( opt->hin );
     i_frame_total -= opt->i_seek;
+    
     if( ( i_frame_total == 0 || param->i_frame_total < i_frame_total )
         && param->i_frame_total > 0 )
         i_frame_total = param->i_frame_total;
+    
     param->i_frame_total = i_frame_total;
+    V(fprintf(stderr, "Encode: i_frame_total:%d\n", param->i_frame_total));
     i_update_interval = i_frame_total ? x264_clip3( i_frame_total / 1000, 1, 10 ) : 10;
-
+    V(fprintf(stderr, "Encode: i_update_interval:%d\n", i_update_interval));
     if( ( h = x264_encoder_open( param ) ) == NULL )
     {
         fprintf( stderr, "x264 [error]: x264_encoder_open failed\n" );
@@ -1202,6 +1287,7 @@
 
     i_start = x264_mdate();
 
+    V(fprintf(stderr, "Starting Encoder loop: i_frame_total %d\n", i_frame_total));
     /* Encode frames */
     for( i_frame = 0, i_file = 0, i_frame_output = 0; b_ctrl_c == 0 && (i_frame < i_frame_total || i_frame_total == 0); )
     {
diff --exclude=.git --exclude=.gitignore -Naur ../x264/x264.h x264/x264.h
--- ../x264/x264.h	2009-10-25 17:41:22.000000000 +0100
+++ x264/x264.h	2009-10-27 16:45:55.000000000 +0100
@@ -24,6 +24,29 @@
 #ifndef X264_X264_H
 #define X264_X264_H
 
+extern int yuv_data_start;
+extern int yuv_data_end;
+//#define YUV_DATA_ADDR 0x01000000
+#define YUV_DATA_ADDR &yuv_data_start
+#define YUV_DATA_SIZE ((int)(&yuv_data_end) - (int)(YUV_DATA_ADDR))
+#define ENC_OUT_ADDR &yuv_data_end
+
+//#define USE_HARDCODED_FRAME_NUM
+#define HARDCODED_FRAME_NUM 10
+
+
+
+/* Enable for verbose output during runtime */
+//#define X264_VERBOSE
+#ifdef X264_VERBOSE
+#define V(x) x
+#else
+#define V(x)
+#endif
+
+
+extern int yuv_data_size;
+
 #if !defined(_STDINT_H) && !defined(_STDINT_H_) && \
     !defined(_INTTYPES_H) && !defined(_INTTYPES_H_)
 # ifdef _MSC_VER
 

Compare with Previous | Blame | View Log

powered by: WebSVN 2.1.0

© copyright 1999-2024 OpenCores.org, equivalent to Oliscience, all rights reserved. OpenCores®, registered trademark.