MCPcopy Create free account
hub / github.com/OpenNMT/CTranslate2 / pole

Function pole

include/half_float/half.hpp:755–761  ·  view source on GitHub ↗

Raise pole error and return infinity. \param sign half-precision value with sign bit only \return half-precision infinity with sign of \a sign \exception FE_DIVBYZERO

Source from the content-addressed store, hash-verified

753 /// \return half-precision infinity with sign of \a sign
754 /// \exception FE_DIVBYZERO
755 inline HALF_CONSTEXPR_NOERR unsigned int pole(unsigned int sign = 0)
756 {
757 #if HALF_ERRHANDLING
758 raise(FE_DIVBYZERO);
759 #endif
760 return sign | 0x7C00;
761 }
762
763 /// Check value for underflow.
764 /// \param arg non-zero half-precision value to check

Callers 11

operator/Function · 0.85
logFunction · 0.85
log10Function · 0.85
log2Function · 0.85
log1pFunction · 0.85
rsqrtFunction · 0.85
powFunction · 0.85
atanhFunction · 0.85
lgammaFunction · 0.85
tgammaFunction · 0.85
logbFunction · 0.85

Calls 1

raiseFunction · 0.85

Tested by

no test coverage detected