MCPcopy Create free account
hub / github.com/NVIDIA/TensorRT / fma

Method fma

samples/common/half.h:1749–1756  ·  view source on GitHub ↗

Fused multiply-add implementation. \param x first operand \param y second operand \param z third operand \return \a x * \a y + \a z stored in single-precision

Source from the content-addressed store, hash-verified

1747 /// \param z third operand
1748 /// \return \a x * \a y + \a z stored in single-precision
1749 static expr fma(float x, float y, float z)
1750 {
1751#if HALF_ENABLE_CPP11_CMATH && defined(FP_FAST_FMAF)
1752 return expr(std::fma(x, y, z));
1753#else
1754 return expr(x * y + z);
1755#endif
1756 }
1757
1758 /// Get NaN.
1759 /// \return Half-precision quiet NaN

Callers

nothing calls this directly

Calls 2

exprClass · 0.70
fmaFunction · 0.70

Tested by

no test coverage detected