Real type in vhdl - floating vs fixed point

EDAboard.com Forum Index -> PLD, SPLD, GAL, CPLD, FPGA Design -> Real type in vhdl - floating vs fixed point

Author

Message

vhdl15

Joined: 03 Nov 2009
Posts: 1

03 Nov 2009 11:05 Real type in vhdl

Hello

I am working on codes of correcting errors.
I must handle real values in entries and exits to perform decoding.
The problem is that the real type is not in synthesizable VHDL, and if I'm going to use the binary representation of the real (floating point) I'll lose in accuracy or accuracy is very important for decoding.
Can you help me it is urgent

Thank you very much

Google
AdSense
Google Adsense

03 Nov 2009 11:05 Ads

avimit

Joined: 16 Nov 2005
Posts: 417
Helped: 69
Location: Fleet, UK

03 Nov 2009 11:45 Re: Real type in vhdl

First you need to decide if you will be using floating point or fixed point data representation.
Second thing is: You will have to compromise on accuracy while you use these representation.
You will have to do analysis of what accuracy you need and what accuracy you can get using fixed point/floting point representation, then make a decision.
HardWare design is always a compromise, and you will have to do some Smile