There is a difference between EDA technologies that are cool and technologies that are actually cool for designers. There is this nice article by EDA femme fatale Peggy Aycinena that details the DVCon keynote by Synopsys Vice-President John Chilton. Synopsys evaluated the design infrastructure costs and their relation to productivity.

Some very interesting findings came of it :

• Changing from Verilog/VHDL to SystemVerilog builds verification into the design process - I think what they mean here is implementation processes; because of what he says later on
• SystemC is important for transaction level modeling and simulation - well SystemC is also verifiable ( I happen to make my living off that!). However, currently SystemVerilog is closer to the language that implementation tools operate on.
• Parallel processing - Now this is controversial . George Harper puts it best on his (ongoing??) series on parallelism. It is not very easy to parallelise a lot of programs. Simulation may be one of them - they key is to break the design into blocks that dont interfere with each other . A sort of conflict graph
• Better place ‘n route - like TeraRoute and K-Route?
• Finally, full design suites - blechhhh. Well, I cant say I blame the Big Three to sneak this one in. What with the recent spate of buffet deals.

Well, this is interesting - I wonder if these are all that is important. I have heard designers ask for better usability, better look-ahead (prototyping) and … 64-bit support. Wonder why those did’nt come up.

del.icio.us Tags: eda