NS32FV16 Datasheet, PDF (53/104 Pages) Texas Instruments – Advanced Imaging/Communication Signal Processors

3 0 Functional Description (Continued)
Operation
aligned real X Y
extended Z
for (n40 n ^k LENG n00)
Z 2n 4 (extended) (Z 2n 0 X n low *
Y n low)
Z 2n01 4 (extended) (Z 2n01 0 X n high
* Y n high)
VARMIN Vector Aligned Real Minimum
The VARMIN instruction compares corresponding elements
of the X and Y real vectors and writes the smaller of the two
in the corresponding place in the Z integer vector
Syntax
EXEC VARMIN
15
10000
Operation
11 10
0
100 0101 1111
Note When PARAM CLR is set to ‘‘1’’ only multiplication is performed
without addition When PARAM SUB is set to ‘‘1’’ the ‘‘a’’ sign is
replaced by a ‘‘b’’ sign
aligned real X Y Z
for (n40 n ^k LENG n00)
VCMAD Vector Complex Multiply and Add
The VCMAD instruction multiplies the corresponding ele-
ments of the X and Y complex vectors and adds or sub-
tracts the result to the complex vector Z This result is
placed in the Z output vector
Syntax
EXEC VCMAD
te 15
10000
Operation
11 10
0
100 1110 0000
complex X Y Z
for (n40 n ^k LENG n00)
le ^{Z n 4 (complex) (Z n 0 X n * Y n )}
Note When PARAM COJ is set to ‘‘1’’ X n is multiplied by the conjugate
of Y n When PARAM CLR is set to ‘‘1’’ only multiplication is per-
formed without addition When PARAM SUB is set to ‘‘1’’ the ‘‘a’’
sign is replaced by a ‘‘b’’ sign
o _{3 4 5 13 Clipping and Min Max Instructions}
VARABS Vector Aligned Real Absolute Value
The VARABS instruction computes the absolute value of
s ^{each element in the real vector X and places the result in}
the corresponding place in the Y output vector
Syntax
EXEC VARABS
b 15
10000
Operation
11 10
0
100 0001 1111
O ^{aligned real X Z}
Z n low 4 min (X n low Y n low)
Z n high 4 min (X n high Y n high)
VARMAX Vector Aligned Real Maximum
The VARMAX instruction compares corresponding ele-
ments of the X and Y real vectors and writes the larger of
the two in the corresponding place in the Z integer vector
Syntax
EXEC VARMAX
15
10000
Operation
11 10
0
100 0110 0110
aligned real X Y Z
for (n40 n ^k LENG n00)
Z n low 4 max (X n low Y n low)
Z n high 4 max (X n high Y n high)
VRFMIN Vector Real Find Minimum
The VRFMIN instruction scans the X real vector and returns
the address of the element with the smallest value The
resulting address is placed in Z 0
Syntax
EXEC VRFMIN
15
11 10
0
10000
100 0110 1101
for (n40 n ^k LENG n00)
Z n low 4 abs (X n low)
Z n high 4 abs (X n high)
Note There is no representation for the absolute value of 0x8000 Whenev-
er an absolute value of 0x8000 is needed OVF SAT is set to ‘‘1’’ and
the maximum positive number 0x7FFF is returned
52