[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [oc] ocrp1



Aloha!

Lars Segerlund wrote:
>  Well there might be a compelling reason to do so in order to get to 
> grips with integrating a fpga in a product :-) , this is the reason for 
> my curiosity about building the proto board myself.
> 
>  I really want to be sure that I understand the timing problems and so 
> on that are encountered in using fpga's since I am working on an 
> upcoming product which will incorporate fpga's.

Which conforms to the personal/education reason I wrote about. Good reasons 
for doing some sort of test board and lab work. The question is if building 
your own general FPGA proto board would be the best research vehicle to 
investigate those things?

>  Also I am currently investigating different synthesis toolsand options 
> so it would be nice to do towo things at once :-)

I would not evaluate a toolchain and debug/develop a proto board (including 
BSP, support SW) at the same time. One of the benefits of having a COTS board 
with toolchain is to have large portions of the infrastructure ready made and 
(hopefully) debugged. Change one parameter at a time and analyze the result.

>  The reason for NOT choosing some vendor board is thus to be a bit 
> vendor independent, learn from experience and get some practice before 
> it we have to get it right.

But, even if you build your board yourself you need to decide on a given 
technology and the vendor independency will be moot. Unless you are not 
planning to optimize the device utilization at all, the vendor specifics must 
be tgought about pretty early on anyways.

Since FPGA different technologies have so varied and fixed resources (between 
vendors) even your architecture might be affected - even if you do 
encapsulation on block level. A good example is partitioning between logic and 
memory which affect the architecture quite a lot depending on which vendor you 
choose.

In short: Vendor independency and FPGAs is a hard thing to do. What one should 
do instead (IMHO) is design architectures that are flexible, adaptable and 
able to achieve good utilization of the technologies. It's a bit different 
from being independent.

Just my 3.14152 Euro.

Good luck!
-- 
Med vänlig hälsning, Yours

Joachim Strömbergson - Alltid i harmonisk svängning.
VP, Research & Development
----------------------------------------------------------------------
InformAsic AB / Hugo Grauers gata 5B / SE-411 33 GÖTEBORG / Sweden
Tel: +46 31 68 54 90  Fax: +46 31 68 54 91  Mobile: +46 733 75 97 02
E-mail: joachim.strombergson@informasic.com  Home: www.informasic.com


--
To unsubscribe from cores mailing list please visit http://www.opencores.org/mailinglists.shtml