0

9.7kviews

Compare Fixed Point and Floating Point Implementation.

written 7.2 years ago by teamques10 ★ 70k

•   modified 6.4 years ago

discrete time signal processing

ADD COMMENT FOLLOW SHARE EDIT

1 Answer

0

1.2kviews

written 7.2 years ago by teamques10 ★ 70k

•   modified 6.4 years ago

Sr no	Fixed Point Implementation	Floating Point Implementation
1	It has limited dynamic range. For example 16-Bit integer represents a maximum range 65,536.	It has large dynamic range. For example single bit positive number has range $2^{-149}$ to $(3.409×10)^{38}$ and double precision number has a range $(-10)^{-308}$ to $(10)^{308}$.
2	Difference between two successive number whether it is small valued or large valued remains same.	The difference between two successive small valued number large valued number is different.
3	Error due to rounding and truncation are large; so it has less precision	Floating point representation gives larger precision.
4	Output of multiply and add stage produces error in the algorithm.	Due to large dynamic range such errors are not produced.
5	Software implementation is complicated.	Software implementation is easy.
6	Less computational power.	More computational power.
7	Addition of two numbers does not affect the precision.	Addition of two number, usually affect the precision.
8	Rounding and truncation must be a part of program.	It is not necessary to specify rounding and truncation.
9	Overflow error occurs because the size of intermediate register is comparatively small.	The size of intermediate register is around 80-Bits, so overflow error does not occur.
10	Requires less registers and less number of input-output pins.	Requires large registers and more number of input-output pins.

ADD COMMENT SHARE EDIT