[ibis-macro] Re: Distribution of compiled 'C'
- From: Joe M Abler <abler@xxxxxxxxxx>
- To: ian_dodd@xxxxxxxxxx
- Date: Wed, 2 Aug 2006 08:57:02 -0400
IP protection is a major issue. DLL's provide a reasonable level of
protection, but not complete. The other control over the IP that I'll
have as a technology supplier is to control distribution to my customer
set under NDA protection. I would not be willing to have models released
into a general EDA library that is distributed by an EDA vendor. My
preference would be to control the compile myself, but if there's some
reason I'd need to work with the EDA vendor to do that I'm willing to
consider it. That however would be an arrangement between the EDA vendor
and myself, and the model would still come back to me for my controlled
distribution. The end user would still see the tool delivered by the EDA
vendor and the models distributed in DLL format by independent technology
suppliers.
Thanks, Joe
Joe Abler abler@xxxxxxxxxx
IBM Systems & Technology Group 919-254-0573
High Speed Serial Link Solutions 919-254-9616 (fax)
3039 Cornwallis Road
Research Triangle Park, NC 27709
"Dodd, Ian" <ian_dodd@xxxxxxxxxx>
Sent by: ibis-macro-bounce@xxxxxxxxxxxxx
08/01/2006 05:05 PM
Please respond to
ian_dodd@xxxxxxxxxx
To
<ibis-macro@xxxxxxxxxxxxx>
cc
Subject
[ibis-macro] Distribution of compiled 'C'
All,
I personally think that Cadence's proposal of using compiled 'C' is not
very practical - there are too many platforms and compiler switches that
produce binaries that are incompatible with each other. How many IC
vendors wishes to produce binaries for Windows, Linux, Solaris, HP Unix
........
Will they be provided as single threaded and multi-threaded .dll's? What
parameter passing convention will be expected? How about the version of
'C' libraries (ANSI or non ANSI)?
I am beginning to think that models written and distributed in 'C' would
be far better - that way the EDA tool vendor provides the compiler and
controls the compile options.
This is along the lines of the methodology used for VHDL-AMS, Verilog-AMS
and System-C today.
We could use encryption to protect IP.
(I still think that compilation of 'C' provides little protection for IP -
I used to be a really dab hand with a disassembler!)
Comments anyone?
Ian
Other related posts:
