TMS320C67x Extensions to the Control Register File
2-15
CPU Data Paths and Control
Table 2–8. Floating-Point Adder Configuration Register Field Descriptions
Bit Position
Width
Field Name
Function
31–27
5
Reserved
26–25
2
Rmode .L2
Value 00: Round toward nearest representable floating-point number
Value 01: Round toward 0 (truncate)
Value 10: Round toward infinity (round up)
Value 11: Round toward negative infinity (round down)
24
1
UNDER .L2
Set to 1 when result underflows
23
1
INEX .L2
Set to 1 when result differs from what would have been computed had
the exponent range and precision been unbounded; never set with
INVAL
22
1
OVER .L2
Set to 1 when result overflows
21
1
INFO .L2
Set to 1 when result is signed infinity
20
1
INVAL .L2
Set to 1 when a signed NaN (SNaN) is a source, NaN is a source in
a floating-point to integer conversion, or when infinity is subtracted
from infinity
19
1
DEN2 .L2
src2 is a denormalized number
18
1
DEN1 .L2
src1 is a denormalized number
17
1
NAN2 .L2
src2 is NaN
16
1
NAN1 .L2
src1 is NaN
15–11
5
Reserved
10–9
2
Rmode .L1
Value 00: Round toward nearest even representable floating-point
number
Value 01: Round toward 0 (truncate)
Value 10: Round toward infinity (round up)
Value 11: Round toward negative infinity (round down)
8
1
UNDER .L1
Set to 1 when result underflows
7
1
INEX .L1
Set to 1 when result differs from what would have been computed had
the exponent range and precision been unbounded; never set with
INVAL
6
1
OVER .L1
Set to 1 when result overflows
5
1
INFO .L1
Set to 1 when result is signed infinity
4
1
INVAL .L1
Set to 1 when a signed NaN is a source, NaN is a source in a floating-
point to integer conversion, or when infinity is subtracted from infinity
3
1
DEN2 .L1
src2 is a denormalized number
2
1
DEN1 .L1
src1 is a denormalized number
1
1
NAN2 .L1
src2 is NaN
0
1
NAN1 .L1
src1 is NaN
