C++中float与double类型的实际最小/最大值是多少？

Hey there! Let's clear up your confusion about FLT_MIN, FLT_MAX, and the actual range of float/double types—since it's totally normal to scratch your head when comparing them to integer types like short int.

First: What FLT_MAX and FLT_MIN actually mean #

The values you saw in Code::Blocks are correct—they just represent concepts that are different from integer min/max:

• FLT_MAX ≈ 3.40282e+38: This is the largest positive value a float can represent. The scientific notation here means 3.40282 × 10³⁸—an astronomically large number way bigger than any integer range you're used to.
• FLT_MIN ≈ 1.17549e-038: This is not the smallest negative float! It's the smallest positive non-zero value a float can represent (1.17549 × 10⁻³⁸)—a tiny number extremely close to zero.

Why float/double ranges look different from integers #

Integer types like short int are fixed-point numbers: they use all their bits to represent a continuous range of whole numbers (for 16-bit short, that's -32768 to 32767—every integer in that span is exactly representable).

Floats and doubles follow the IEEE 754 standard, which uses a combination of sign bit, exponent bits, and mantissa (fraction) bits. This design lets them:

• Cover a massive range of values (from near-zero up to numbers you can't even fathom)
• Represent decimal fractions with varying precision (higher precision near zero, lower precision for very large numbers)

But this comes with tradeoffs: the range isn't symmetric in the same way integers are, and not every number in the range is exactly representable.

The actual full range of float and double #

If you're looking for the complete span of possible values:

For float: #

• Largest positive value: FLT_MAX ≈ 3.40282347e+38
• Smallest negative value: -FLT_MAX ≈ -3.40282347e+38
• Smallest positive non-zero value: FLT_MIN ≈ 1.17549435e-38

For double: #

• Largest positive value: DBL_MAX ≈ 1.7976931348623157e+308
• Smallest negative value: -DBL_MAX ≈ -1.7976931348623157e+308
• Smallest positive non-zero value: DBL_MIN ≈ 2.2250738585072014e-308

What about those odd values you got (47.249... and -34.804...)? #

Those numbers aren't the actual min/max of float—they're likely the result of a mistake in your code. Maybe you were calculating the range of a specific variable (like a value from sensor input, or a computed result) instead of querying the type's inherent limits. Double-check your code to see where those values came from!

Quick recap #

• Integer types have small, continuous ranges of exact whole numbers.
• Floats/doubles have massive ranges, but MIN refers to the smallest positive non-zero value, not the most negative. To get the most negative float, use -FLT_MAX.

内容的提问来源于stack exchange，提问作者user9318444