(07-17-2013, 11:10 PM)haddockd Wrote: The speed is the same with or without the ZTP speedhack so I am unsure if maybe it was broken at some point.These ideas sound crazy until they turn out not to be, so I checked.
Code is still there. If you enable logging, do you see such a message logged (ZTP map texture found at address)?
Code:
if (g_ActiveConfig.bZTPSpeedHack)
{
if (!mapTexFound)
{
if (bp.address != BPMEM_TEV_COLOR_ENV && bp.address != BPMEM_TEV_ALPHA_ENV)
{
numWrites = 0;
}
else if (++numWrites >= 100) // seem that if 100 consecutive BP writes are called to either of these addresses in ZTP,
{ // then it is safe to assume the map texture address is currently loaded into the BP memory
mapTexAddress = bpmem.tex[0].texImage3[0].hex << 5;
mapTexFound = true;
WARN_LOG(VIDEO, "\nZTP map texture found at address %08x\n", mapTexAddress);
}
FlushPipeline();
}
else if ( (bpmem.tex[0].texImage3[0].hex << 5) != mapTexAddress ||
bpmem.tevorders[0].getEnable(0) == 0 ||
bp.address == BPMEM_TREF)
{
FlushPipeline();
}
} // END ZTP SPEEDUP HACK