Meeting date: 06 jun 2006 Members (asterisk for those attending): *Arpad Muranyi, Intel Corp. *Bob Ross, Teraspeed Consulting Group *Todd Westerhoff, Cisco Systems *Mike LaBonte, Cisco Systems Paul Fernando, NCSU Barry Katz, SiSoft Walter Katz, SiSoft Ken Willis, Cadence Design Systems *Ian Dodd, Mentor Graphics *Lance Wang, Cadence Design Systems Richard Ward, Texas Instruments Doug White, Cisco Systems Sanjeev Gupta, Agilent Joe Abler, IBM John Shields, Mentor Graphics ------------- Review of ARs: - Mike forward diagram to John Shields for comments - TBD - Ian ask Gary Pratt about progress getting models - Gary is still working on it ------------- Next week discuss inputs to Matlab model: - Mentor uses impulse response/channel simulation approach. - Impulse response created with t-line or AMS simulator - Custom simulator uses this to simulate millions of cycle - Has to be done for whole channel. Joe Abler: - Developed EQ architecture: - DFE algorithm - CDR algorithm - Simulation can pull in xtalk channels. - It fully models digital control algorithm. - Supports fast/typ/slow corners. - Uses S parameters for packages. - Written in C, not Matlab: - It does Matlab-like calculations for you (home-grown). - Matlab is used for results outputs only. - Simulation is mostly in the time domain. - Not transistor models, relies on linearity. - Real receiver models are an advantage: - In old simulations the receiver was just a load. - Checking receiver performance was done by post-analysis. - How to simulate IBM receiver with another driver? - It's not interoperable. - Why was it written in C? - First define partition of: - What is in GUI - What is in serdes - What are inputs - What are outputs - Writing it "from scratch" makes it not interoperable. - Someone has to own the simulator. - Would like to get away from this. - The B element reads IBIS info and interfaces with simulator - This is an interpreter. - AMS simulators are powerful, but can they do hardware-software co-design? - SystemC is a platform for this. - A DLL-like solution would be good. - Many simulators have these interfaces. - Encrypted AMS and machine binary are the same to users. - Is this in conflict with the IBIS mission? - People had trouble learning IBIS, this may be harder. - We need to look into interface specification details. - Why are AMS-like languages proliferating? - SystemVerilog - SystemC - No perfect solution ------------- Next meeting: Tuesday 13 Jun 2006 12:00pm PT