IEEE-754 Floating-Point Inspector
See exactly how a number is stored in IEEE-754 floating point. Enter a decimal number to decompose it into sign, exponent, and significand for both single (32-bit) and double (64-bit) precision, with the full binary layout, hex, the nearest representable value, and the rounding error. Or go the other way: paste a hex or binary bit pattern to recover the decimal value. Handles zeros, infinities, NaN, and subnormals. Everything runs locally in your browser.
Single precision (float32, 32 bits)
Double precision (float64, 64 bits)
IEEE-754 stores a float as (−1)sign × 1.mantissa × 2exp−bias. Single precision uses 1 sign bit, 8 exponent bits (bias 127), and 23 mantissa bits; double uses 1, 11 (bias 1023), and 52. The leading 1 of a normal number is implicit (not stored). Exponent field 0 with a nonzero mantissa is a subnormal (no implied leading 1); all-zero is ±0; all-ones exponent with zero mantissa is ±Infinity; all-ones with nonzero mantissa is NaN. Because most decimals (like 0.1) aren't exact powers of two, they're rounded to the nearest representable value — the inspector shows that rounding error. Everything runs locally — nothing leaves your browser.