Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Register Log in

A simple question for experts

Status
Not open for further replies.
S

sheikh

Advanced Member level 4
Joined
Sep 10, 2007
Messages
104
Helped
1
Reputation
2
Reaction score
1
Trophy points
1,298
Activity points
2,008
Hello Dear experts
I have an ASM an I extracted a Data Path from it. This Data Path contains 7 ADD/SUB and 8 Multipliers units and when I see it on paper, it seems a little big. I don't have any experience in real implementation of a design in to a FPGA, So please tell me according to your experience, if I use this amount of ADD/SUB and multipliers, does it consume a lot of sources of a FPGA (spartan3, for instance) ? Is it necessary to use hardware replication technique, for instance, to reducing the amount of hardware, or not?
Regard
Mostafa
 

this is a pretty small design, even if the data was quite wide. The adders arnt much logic and you can use embedded multipliers.

- - - Updated - - -

This is assuming integer arithmatic. If you're talking floating point its a whole other ball game.
 
  • Like
Reactions: sheikh

    S

    sheikh

    Points: 2
    Helpful Answer Positive Rating
Thanks a lot dear TrickyDicky for fast and clear answer. So there isn't any worry about it. Also, Could you please tell me how can I find that a design consumes a lot of sources or not? I mean, Do I need to read the FPGA datasheet to find out the amount of sources and then check the number of units applied in a design and decide according to it, or I need more experience? ( because lots of fpga family exist).
Thanks in advance

- - - Updated - - -

I talkin floating point.
 

Of course he's talking floating point. Fixed point would just be too easy. Could you explain why your problem needs floating point, and could not be solved with fixed point by just a little extra thinking?

As for resource usage, either some experience so you get a feel for it. Or just synthesize the design and check out resource usage in the synth report.
 
  • Like
Reactions: sheikh

    S

    sheikh

    Points: 2
    Helpful Answer Positive Rating
Thanks dear mrflibble, well I want to use the values of pixels in an image and my software program in matlab used the values in double format so i think for more accuracy it can be better to use the floating point (32bits) in my design. but as I said I was confused by the size of design ( consider it that I have another units, such as color convertor, but they need less units than this one)
Regards
Mostafa
 

dude you can use xilinx ISE to see if your design fits on FPGA and if so how much area does it consume
ISE clearly states how much area your design take after synchronizing it
and for the software, don't worry. you can instal xilinx web pack without having to pay anything
 

Thanks sina.parsnejad
 

Ah damnation, just lost a medium long post. Well, short version is: think harder, use fixed point. For your average image processing task you can do the job just fine with fixed point. The fact that matlab uses doubles doesn't mean your digital design suddenly has to use floating point.
 
  • Like
Reactions: sheikh

    S

    sheikh

    Points: 2
    Helpful Answer Positive Rating
Thanks a lot Dear mrflibble:D OK, I will think harder:))
 

Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top