Short Word Floating-Point Format; Packing For Floating-Point Data - Analog Devices ADSP-21261 SHARC Hardware Reference Manual

Short Word Floating-Point Format

The DSP supports a 16-bit floating-point data type and provides conver-
sion instructions for it. The short float data format has an 11-bit mantissa
with a 4-bit exponent plus sign bit, as shown in
floating-point numbers reside in the lower 16 bits of the 32-bit float-
ing-point field.
15 14
e 3
s • • •
HIDDEN BIT
Figure 2-4. 16-Bit Floating-Point Format

Packing for Floating-Point Data

Two shifter instructions,
unpacking conversions between 32-bit floating-point words and 16-bit
floating-point words. The
ing-point number to a 16-bit floating-point number. The
instruction converts the 16-bit floating-point numbers back to 32-bit
IEEE floating-point numbers. Each instruction executes in a single cycle.
The results of the
Table 2-3.
ADSP-2126x SHARC Processor Hardware Reference
11 10
e 0 1 . f 10
and
FPACK FUNPACK
instruction converts a 32-bit IEEE float-
FPACK
and
FPACK FUNPACK
Processing Elements
Figure
• • •
BINARY POINT
, perform the packing and
operations appear in
2-4. The 16-bit
0
f 0
FUNPACK
Table 2-2 and
2-7