Forums

Fixed Point DSP Vs Floating Point DSP

Started by Sandeep Chikkerur June 26, 2003
Hi,

  I searched the internet to find out the difference between a fixed
point & a  floating point DSP.
To my surprise, I did not find a well defined difference.
   
Plz, can anyone differentiate between the two so that a layman can
also understand ?

Thanx in advance...
On Thu, 26 Jun 2003 03:19:53 -0700, Sandeep Chikkerur wrote:

> I searched the internet to find out the difference between a fixed > point & a floating point DSP. > To my surprise, I did not find a well defined difference.
That would surprise me, too.
> Plz, can anyone differentiate between the two so that a layman can > also understand ?
What purpose would that serve, if that layman did not understand what the difference between fixed point and floating point arithmetic was? And if a person was sufficiently knowledgeable that they *did* know the difference between fixed and floating point arithmetic, then it would be a trivial matter indeed, to point out to them that fixed point processors have no hardware to perform floating point arithmetic, but that floating point ones do. Cheers, -- Andrew
Sandeep Chikkerur wrote:
> > Hi, > > I searched the internet to find out the difference between a fixed > point & a floating point DSP. > To my surprise, I did not find a well defined difference. > > Plz, can anyone differentiate between the two so that a layman can > also understand ? > > Thanx in advance...
A fixed-point DSP does only fixed-poing arithmetic, and usually has hardware to facilitate multiplication. A floating-point DSP has extra hardware for floating-point addition and multiplication (addition is a bit more involved than multiplication) and often omits the hardware to facilitate fixed-poing multiplication. Expect some difference in the size of a few registers. A matter of curiosity (mine): What difference had you imagined? Jerry -- Engineering is the art of making what you want from things you can get. �����������������������������������������������������������������������
Sandeep Chikkerur wrote:
> > Hi, > > I searched the internet to find out the difference between a fixed > point & a floating point DSP. > To my surprise, I did not find a well defined difference. > > Plz, can anyone differentiate between the two so that a layman can > also understand ? > > Thanx in advance...
A fixed-point DSP does only fixed-point arithmetic, and usually has hardware to facilitate multiplication. A floating-point DSP has extra hardware for floating-point addition and multiplication (addition is a bit more involved than multiplication) and often omits the hardware to facilitate fixed-poing multiplication. Expect some difference in the size of a few registers. A matter of curiosity (mine): What difference had you imagined? Jerry -- Engineering is the art of making what you want from things you can get. �����������������������������������������������������������������������
In article <d5d88eb5.0306260219.36f87462@posting.google.com>, 
sandeep_mc81@yahoo.com says...
> Hi, > > I searched the internet to find out the difference between a fixed > point & a floating point DSP. > To my surprise, I did not find a well defined difference. > > Plz, can anyone differentiate between the two so that a layman can > also understand ? > > Thanx in advance... >
heres a few points about how I understand it. In my experience, the processor type influenced our filter choices: - floating point DSPs can do floating point operations quickly (eg 1 clock cycle). Fixed point DSPs do floating point operations slowly (eg 50 clock cycles) because they use compiler tricks to implement "floating point emulation". - many dsp algorithms depend critically on the absence/presence of quantization due to integer arithmetic. For example, all the IIR stability considerations are further complicated by fixed point arithmetic quantization issues. - When dealing with a fixed point processor, you use a filter design package that includes a 'quantization analyser'. In Matlab, you design the filter, look at the response the see if its suitable and stable, then you turn on the quantization plugin. This recalculates the filters transfer function to take into account the effects of numeric rounding/truncation. Often, the original filter is stable but after adding in the effect of quantization, the filter is unstable. It depends also on what quantization parameters you choose. You set these params to match your processor. - before choosing a fixed/floating point processor, you need to understand your application. Will it require floating point arithmetic? (eg do I need FIR's, IIR's, alot of IIR's) Can you get away with floating point emulation in a fixed point proc? (Maybe I only need a single low pole IIR?) -- http://www.mobilecomms.com.au http://www.nexiondata.com