ShuriZma
/
pcsx2
mirror of https://github.com/PCSX2/pcsx2.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

newVif: last bugfix was crappy. This one actually fixes the bug without introducing another similar bug in the same line of code. :p 

git-svn-id: http://pcsx2.googlecode.com/svn/trunk@2400 96395faa-99c1-11dd-bbfe-3dabce05a288
This commit is contained in:
Jake.Stine 2009-12-25 21:18:36 +00:00
parent b795e7a306
commit b1fc08409b
1 changed files with 6 additions and 8 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

14
pcsx2/x86/newVif_HashBucket.h
View File

@ -65,16 +65,14 @@ public:
u32 d = *((u32*)dataPtr);
const SizeChain<T>& bucket( mBucket[d % hSize] );
for (int i=bucket.Size-1; i>0; --i) {
// This inline version seems about 1-2% faster in tests of games that average 1
// program per bucket. Games that average more should see a bigger improvement --air
int result = _mm_movemask_ps( (cast_m128) _mm_cmpeq_epi32( _mm_load_si128((__m128i*)&bucket.Chain[i]), _mm_load_si128((__m128i*)dataPtr) ) );
if( (result&0x7) == 0x7 ) return &bucket.Chain[i];
const __m128i* endpos = (__m128i*)&bucket.Chain[bucket.Size];
const __m128i data128( _mm_load_si128((__m128i*)dataPtr) );
// Dynamically generated function version, can't be inlined. :(
//if ((((nVifCall)((void*)nVifMemCmp))(&bucket.Chain[i], dataPtr))==7) return &bucket.Chain[i];
for( const __m128i* chainpos = (__m128i*)bucket.Chain; chainpos<endpos; ++chainpos ) {
// This inline SSE code is generally faster than using emitter code, since it inlines nicely. --air
int result = _mm_movemask_ps( (cast_m128) _mm_cmpeq_epi32( data128, _mm_load_si128(chainpos) ) );
if( (result&0x7) == 0x7 ) return (T*)chainpos;
//if (!memcmp(&bucket.Chain[i], dataPtr, sizeof(T)-4)) return &c[i]; // old school version! >_<
}
if( bucket.Size > 3 ) DevCon.Warning( "recVifUnpk: Bucket 0x%04x has %d micro-programs", d % hSize, bucket.Size );
return NULL;