Sebastian John Howard
Reputation: 890
Is it possible for a platform to have "partial" IEEE754 floating point types
I'm wondering, is it theoretically possible for a platform to implement the IEEE754 specification for a double, but not for a float? (And the converse case.)
Upvotes: 0
Views: 61
Answers (1)
Mikhail
Reputation: 8038
Yes, here is a real world example of the opposite thing:
Take a look at this slide about CUDA numeric from the following presentation.
Upvotes: 1
Related Questions
- Are there any modern platforms with non-IEEE C/C++ float formats?
- How to detect non IEEE-754 float, and how to use them?
- C/C++: Are IEEE 754 float addition/multiplication/... and int-to-float conversion standardized?
- Floating-point and ieee-754
- Floating point types representation
- Are there any commonly used floating point formats besides IEEE754?
- Does the C++ standard allow this floating-point behaviour?
- Does the C++ standard specify anything on the representation of floating point numbers?
- Float in c++ smaller range than IEEE 754
- C++ compilers/platforms that don't use IEEE754 floating point