All, I have seen no serious argument why LTI is not a valid assumption for Interconnect today or in the foreseeable future. Unlike silicon, interconnect modeling has been incredibly stable for the last two decades, and yes interconnect modeling has been done at speeds over 7Gig for a long time in the communications industry. I do not believe that anybody seriously claims the need for non-LTI models for simulation. There is clearly a need for interconnect solution space exploration along the lines of what was implemented for AMI parameters. Now consider the advantages of netlist of blocks, each block represented by one of a short list of models we all know and understand. A netlist that each EDA vendor understands can be simply translated into his own simulators. Now consider timely (desperate?) need for package and connector manufactures to be able to deliver accurate models. And consider the need that system developers have in reliably designing next generation high-speed systems. I would hope we can put this LTI discussion to rest, and start focusing on decisions like the required set of modeling blocks, the format of the subckt netlisting format, ... Walter -----Original Message----- From: ibis-macro-bounce@xxxxxxxxxxxxx [mailto:ibis-macro-bounce@xxxxxxxxxxxxx]On Behalf Of Muranyi, Arpad Sent: Thursday, June 26, 2008 7:59 PM To: IBIS-ATM Subject: [ibis-macro] Re: Question on seeting the EMD direction Let's put aside time constants, mechanical or cross talk effects. My goal is to get a consensus on whether it is safe to assume LTI for a new interconnect language (or specification) we are just starting to develop for the present day and future designs. I simply do not want it to be obsolete by the time it is finished, or I do not want to have to develop one specification after the other because by the time one is done, another one is needed because the one we just finished is not flexible enough to take on some badly needed features. I do not want to repeat history. To me it is quite embarrassing that we spent 7 or so years on something like ICM, or 3-4 years on something like the IBIS-AMS extensions, etc... to find out that people are not interested in using them for various reasons. Lastly, regarding the old saying about user interfaces (on the bottom of this message). I am a little surprised to hear this quote from someone who favors the C language for AMI modeling over other, somewhat simpler, but limiting languages, such as VHDL-AMS, Verilog-A(MS), Matlab, and the like... After all, C is a "...language which is capable of expressing any engineering problem..." isn't it? Either way, my goal is not necessarily to achieve a language that is capable of expressing ANY ENGINEERING problem, I just want to propose a modeling language that is flexible and expandable enough so that it doesn't paint itself into the corner by the time it is released... Comments? Arpad ======================================================== --------------------------------------------------------------------- IBIS Macro website : http://www.eda.org/pub/ibis/macromodel_wip/ IBIS Macro reflector: //www.freelists.org/list/ibis-macro To unsubscribe send an email: To: ibis-macro-request@xxxxxxxxxxxxx Subject: unsubscribe