cu_arithmetic Namespace Reference

perform arithmetics for __half and __nv_bfloat16 types in round-to-nearest-even mode. More...

Detailed Description

perform arithmetics for __half and __nv_bfloat16 types in round-to-nearest-even mode.

This namespace is a templated unifying API for CUDA's arithmetic operations for __half and __nv_bfloat16 types, including addition and multiplication of two float numbers.

See also

cu_arithmetics::add, cu_arithmetics::is_equal cu_arithmetics::cast

This namespace is a templated unifying API for CUDA's arithmetic operations for __half and __nv_bfloat16 types, including addition and multiplication of two float numbers.

See also

cu_arithmetics::sub, cu_arithmetics::is_equal cu_arithmetics::cast

This namespace is a templated unifying API for CUDA's arithmetic operations for __half and __nv_bfloat16 types, including addition and divtiplication of two float numbers.

See also

cu_arithmetics::mul, cu_arithmetics::is_equal cu_arithmetics::cast

This namespace is a templated unifying API for CUDA's arithmetic operations for __half and __nv_bfloat16 types, including epsilonition and multiplication of two float numbers.

See also

cu_arithmetics::add, cu_arithmetics::is_equal cu_arithmetics::cast

This namespace is a templated unifying API for CUDA's arithmetic operations for __half and __nv_bfloat16 types, including addition and multiplication of two float numbers.

See also

cu_arithmetics::div, cu_arithmetics::is_equal cu_arithmetics::cast

This namespace is a templated unifying API for CUDA's arithmetic operations for __half and __nv_bfloat16 types, including subition and multiplication of two float numbers.

See also

cu_arithmetics::add, cu_arithmetics::is_equal cu_arithmetics::cast