Freescale Semiconductor MC68881 - Computational Accuracy; System Control Operations

Table 4-8. Conditional Test Mnemonics (Continued)
Exception on Unordered No Exception on Unordered
LE ,Less Than or Equal OLE Ordered Less Than or Equal
LT 'Less Than OLT Ordered Less Than
NGE Not (Greater Than or Equal) UGE Unordered or Greater Than Equal
NGL Not (Greater Than or Less Than) UEQ Unordered or Equal
NGLE Not (Greater Than or Less Then or Equal) UN Unordered
NGT Not Greater Than UGT Unordered or Greater Than
NLE Not (Less Than or Equal) ULE Unordered or Less Than or Equal
NLT Not Less Than ULT Unordered or Less Than
SEO Signaling Equal EQ Equal,
SNE Signaling Not Equal NE Not Equal
SF Signaling Always False F Always False
ST Signaling Always True T Always True
4.2.5 System Control Operations
The system control instructions communicate with the operating system via a conditional
trap instruction (FTRAPcc) and save or restore (FSAVE or FRESTORE) the nonuser visible
portion of the FPCP during context switches in a virtual memory or other type of multi-
tasking system. The conditional trap instruction uses the same conditional tests as the
program control instructions and allows an optional 16- or 32-bit immediate operand to
be included as part of the instruction for passing parameters to the operating system. Table
4-9 summarizes the system control instructions.
Table 4-9. System Control Operations
Instruction Operand Syntax Operand Size Operation
FRESTORE (ea) None State Frame t Internal Registers
FSAVE (ea) None internal Registers ~ State Frame
FTRAPcc None None If Condition True,
#xxx W,L Then Take Exception
4.3 COMPUTATIONAL ACCURACY
Whenever an attempt is made to represent a real number in a binary format of finite
precision, there is a possibility that the number cannot be represented exactly; this is
commonly referred to as round-off error. Furthermore, when two inexact numbers are used
in a calculation, the error present in each number is reflected and possibly aggravated in
the result.
One of the major reasons that the
IEEE Standard for Binary Floating-PointArithmetic (ANSI/
IEEE Std. 754,1985) was developed is to define the error bounds for calculation of binary
MC68881/MC68882 USER'S MANUAL FREESCALE
4-5