Efficient fixed-point multiplication by 2^{K}

`b = pow2(a,K)`

`b = pow2(a,K)`

returns the value of `a`

shifted
by `K`

bits where `K`

is an integer
and `a`

and `b`

are `fi`

objects.
The output `b`

always has the same word length and
fraction length as the input `a`

.

In fixed-point arithmetic, shifting by `K`

bits
is equivalent to, and more efficient than, computing *b* = *a**2^{k}.

If `K`

is a non-integer, the `pow2`

function
will round it to `floor`

before performing the calculation.

The scaling of `a`

must be equivalent to binary
point-only scaling; in other words, it must have a power of 2 slope
and a bias of 0.

`a`

can be real or complex. If `a`

is
complex, `pow2`

operates on both the real and complex
portions of `a`

.

The `pow2`

function obeys the `OverflowAction`

and `RoundingMethod`

properties
associated with `a`

. If obeying the `RoundingMethod`

property
associated with `a`

is not important, try using the `bitshift`

function.

The `pow2`

function does not support `fi`

objects
of data type `Boolean`

.

The function also does not support the syntax `b = pow2(a)`

when `a`

is
a `fi`

object.

Example 4. Example 1

In the following example, `a`

is a real-valued `fi`

object,
and `K`

is a positive integer.

The `pow2`

function shifts the bits of `a`

3
places to the left, effectively multiplying `a`

by 2^{3}.

a = fi(pi,1,16,8) b = pow2(a,3) binary_a = bin(a) binary_b = bin(b)

MATLAB^{®} returns:

a = 3.1406 DataTypeMode: Fixed-point: binary point scaling Signedness: Signed WordLength: 16 FractionLength: 8 b = 25.1250 DataTypeMode: Fixed-point: binary point scaling Signedness: Signed WordLength: 16 FractionLength: 8 binary_a = 0000001100100100 binary_b = 0001100100100000

Example 5. Example 2

In the following example, `a`

is a real-valued `fi`

object,
and `K`

is a negative integer.

The `pow2`

function shifts the bits of `a`

4
places to the right, effectively multiplying `a`

by 2^{–4}.

a = fi(pi,1,16,8) b = pow2(a,-4) binary_a = bin(a) binary_b = bin(b)

MATLAB returns:

a = 3.1406 DataTypeMode: Fixed-point: binary point scaling Signedness: Signed WordLength: 16 FractionLength: 8 b = 0.1953 DataTypeMode: Fixed-point: binary point scaling Signedness: Signed WordLength: 16 FractionLength: 8 binary_a = 0000001100100100 binary_b = 0000000000110010

Example 6. Example 3

The following example shows the use of `pow2`

with
a complex `fi`

object:

format long g P = fipref('NumericTypeDisplay', 'short'); a = fi(57 - 2i, 1, 16, 8) a = 57 - 2i s16,8 pow2(a, 2) ans = 127.99609375 - 8i s16,8

Was this topic helpful?