Added SCE SIMD math library in Extras/simdmathlibrary

The upcoming vectormath that will used to speed up the SPU version of Extras/BulletMultiThreaded depends on this.
2007-07-23 04:58:24 +00:00
parent 685138d033
commit 7529cdb3f6
287 changed files with 32064 additions and 0 deletions
--- a/Extras/simdmathlibrary/spu/divi4.c
+++ b/Extras/simdmathlibrary/spu/divi4.c
@@ -0,0 +1,109 @@
+/* divi4 - 
+   Copyright (C) 2006, 2007 Sony Computer Entertainment Inc.
+   All rights reserved.
+
+   Redistribution and use in source and binary forms,
+   with or without modification, are permitted provided that the
+   following conditions are met:
+    * Redistributions of source code must retain the above copyright
+      notice, this list of conditions and the following disclaimer.
+    * Redistributions in binary form must reproduce the above copyright
+      notice, this list of conditions and the following disclaimer in the
+      documentation and/or other materials provided with the distribution.
+    * Neither the name of the Sony Computer Entertainment Inc nor the names
+      of its contributors may be used to endorse or promote products derived
+      from this software without specific prior written permission.
+
+   THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+   AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+   IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+   ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
+   LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+   CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+   SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+   INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+   CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+   ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+   POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include <simdmath.h>
+#include <spu_intrinsics.h>
+
+// divi4 - for each of four integer slots, compute quotient and remainder of numer/denom
+// and store in divi4_t struct.  Divide by zero produces quotient = 0, remainder = numerator.
+
+divi4_t divi4 (vector signed int numer, vector signed int denom)
+{
+   divi4_t res;
+   vec_int4 quot, newQuot, shift;
+   vec_uint4 numerPos, denomPos, quotNeg;
+   vec_uint4 numerAbs, denomAbs;
+   vec_uint4 denomZeros, numerZeros, denomLeft, oneLeft, denomShifted, oneShifted;
+   vec_uint4 newNum, skip, cont;
+   int       anyCont;
+
+   // Determine whether result needs sign change
+
+   numerPos = spu_cmpgt( numer, -1 );
+   denomPos = spu_cmpgt( denom, -1 );
+   quotNeg = spu_xor( numerPos, denomPos );
+    
+   // Use absolute values of numerator, denominator
+
+   numerAbs = (vec_uint4)spu_sel( spu_sub( 0, numer ), numer, numerPos );
+   denomAbs = (vec_uint4)spu_sel( spu_sub( 0, denom ), denom, denomPos );
+
+   // Get difference of leading zeros.
+   // Any possible negative value will be interpreted as a shift > 31
+
+   denomZeros = spu_cntlz( denomAbs );
+   numerZeros = spu_cntlz( numerAbs );
+
+   shift = (vec_int4)spu_sub( denomZeros, numerZeros );
+
+   // Shift denom to align leading one with numerator's
+
+   denomShifted = spu_sl( denomAbs, (vec_uint4)shift );
+   oneShifted = spu_sl( (vec_uint4)spu_splats(1), (vec_uint4)shift );
+   oneShifted = spu_sel( oneShifted, (vec_uint4)spu_splats(0), spu_cmpeq( denom, 0 ) );
+
+   // Shift left all leading zeros.
+
+   denomLeft = spu_sl( denomAbs, denomZeros );
+   oneLeft = spu_sl( (vec_uint4)spu_splats(1), denomZeros );
+
+   quot = spu_splats(0);
+
+   do
+   {
+      cont = spu_cmpgt( oneShifted, 0U );
+      anyCont = spu_extract( spu_gather( cont ), 0 );
+
+      newQuot = spu_or( quot, (vec_int4)oneShifted );
+
+      // Subtract shifted denominator from remaining numerator 
+      // when denominator is not greater.
+
+      skip = spu_cmpgt( denomShifted, numerAbs );
+      newNum = spu_sub( numerAbs, denomShifted );
+
+      // If denominator is greater, next shift is one more, otherwise
+      // next shift is number of leading zeros of remaining numerator.
+
+      numerZeros = spu_sel( spu_cntlz( newNum ), numerZeros, skip );
+      shift = (vec_int4)spu_sub( skip, numerZeros );
+
+      oneShifted = spu_rlmask( oneLeft, shift );
+      denomShifted = spu_rlmask( denomLeft, shift );
+
+      quot = spu_sel( newQuot, quot, skip );
+      numerAbs = spu_sel( newNum, numerAbs, spu_orc(skip,cont) );
+   } 
+   while ( anyCont );
+
+   res.quot = spu_sel( quot, spu_sub( 0, quot ), quotNeg );
+   res.rem = spu_sel( spu_sub( 0, (vec_int4)numerAbs ), (vec_int4)numerAbs, numerPos );
+   return res;
+}
+