Microsoft DirectX 8.0 (C++) |
Provides exponential 2x with full precision to at least 1/220.
exp vDest, vSrc0
SetDestReg(); SetSrcReg(0); float v = m_Source[0].w; m_TmpReg.x = m_TmpReg.y = m_TmpReg.z = m_TmpReg.w = (float)pow(2, v); WriteResult();
This macro takes twelve instruction slots.
This is a scalar operation and takes its input from the .w channel, which must be specified. It always replicates the result into all four output channels.