noun
subnormal number; denormal number
Technical term in computing (IEEE 754 floating-point). Refers to numbers too small to be represented in normal form, also called denormalized numbers.
非正規化数は、浮動小数点演算でアンダーフローを緩和するために使われる。
Subnormal numbers are used to mitigate underflow in floating-point arithmetic.
Compound of 非 (non-), 正規化 (normalization), and 数 (number). A direct calque from English 'subnormal number' or 'denormalized number'.