The IEEE standard provides two main floating-point formats: single-precision (32-bit) and double-precision (64-bit). It is divided into three parts: sign, exponent and decimal. The exponent part determines the range of floating-point numbers. The number of decimal places determines the accuracy. (See below)
In single precision, the exponent length is 8 bits, and the decimal part accounts for 23 bits. Since it is a signed type, the corresponding exponent ranges from -128 to 128. So the range of single precision is -2^128 to 2^128, which is approximately -3.40 * 10^38 to 3.40 * 10^38(or -3.40E38 to +3.40E38). Because the decimal part is 23 places, and 2^23=8388608. The corresponding base 10 is seven digits. So the accuracy range is 6 to 7 bits (i.e. significant digits are usually 6 to 7 bits).
Similarly, in double precision, the exponent length is 11 bits, and the decimal part is 52 bits. Therefore, the range of double precision is -2^1024 to 2^1024, approximately equal to -1.797E308 to 1.797E308. The decimal part is 52 places, and 2^52= 4,503,599,627,370,496. The corresponding base 10 is 16 bits. So the accuracy range is 15 to 16 bits (i.e. significant digits are usually 15 to 16 bits).