Category:Floating-Point
This category includes topics related to floating-point computation.
Not listed here are the RTL floating-point data types, and related procedures and functions. You can find most of these in the System and System.Math units.
The FPU (Floating Point Unit) handles floating-point calculations on 32-bit systems and the SSE (Streaming SIMD Extensions) handles floating-point calculations on 64-bit systems. You can prevent floating-point exceptions from occuring by using SetExceptionMask to mask some or all of them on both FPU and SSE. SetExceptionMask sets the floating-point exception bits in FPU Control Word and SSE's MXCSR Control/Status registers.
The FPU Control Word register is illustrated below:
|
Bit |
Description | |
|---|---|---|
| 15 | Reserved | |
| 14 | Reserved | |
| 13 | Reserved | |
| 12 | Infinity control | |
| 11 | Rounding control | |
| 10 | ||
| 9 | Precision control | |
| 8 | ||
| 7 | Reserved | |
| 6 | Reserved | |
| 5 | Exception masking | Precision |
| 4 | Underflow | |
| 3 | Overflow | |
| 2 | Zero divide | |
| 1 | Denormal operand | |
| 0 | Invalid operation | |
See Also
- W1066 Lost Extended floating point precision. Reduced to Double (Delphi)
- Intel and Floating-Point
- x87 and SSE Floating Point Assists in IA-32: Flush-To-Zero (FTZ) and Denormals-Are-Zero (DAZ)
- Intel® 64 and IA-32 Architectures Software Developer's Manuals
- IA-64 Floating-Point Operations and the IEEE Standard for Binary Floating-Point Arithmetic
Pages in category "Floating-Point"
The following 22 pages are in this category, out of 22 total.
F
- Float
- Float and Double Limits
- Float.h Index
- Floating Point Constants
- Floating point precision control (Delphi for x64)
- Floating-Point Comparison Routines
- Floating-Point Exception: Divide by 0, Domain or Overflow
- Floating-Point Exception: Partial Loss of Precision or Underflow
- Floating-Point Exception: Stack Fault
- Floating-Point Exceptions
- Floating-Point Number Control Routines
- Floating-Point Rounding Routines
- FPU