As we head for March conferences such as DATE, HDLCon and the Synopsys Users Group (SNUG), the C language design debate is likely to rear its ugly head. But emotions and vendor posturing may obscure the real role of C-language design, which is not what most designers think.

Judging from postings in the E-Mail Synopsys Users Group (ESNUG), many RTL designers seem to think there's an evil conspiracy of vendors and CAD managers to take away Verilog and force a move to C/C++. If there is, it won't work. If today's RTL ASIC designers need a higher level of abstraction, Superlog makes far more sense than C/C++, at least for Verilog users.

The point that may often be missed is that C/C++ language design is already in widespread use today. It's being used by many major electronics manufacturers, particularly in Europe, for system-level modeling and silicon intellectual property (IP) design. In this sense, C language design is an old methodology, not a new one.

Most of these companies, however, use their own proprietary C/C++ class libraries, making IP exchange very difficult. That's where SystemC comes in. One standardized library will allow companies to mix and match C language IP from different sources. SystemC is really about IP modeling, not forcing RTL designers into a new methodology.

Another legitimate place for C language tools is to ease the gap between C and HDL models. Today, there usually needs to be a complete rewrite. Any way to automate the translation process will help.

C language models can also help speed RTL simulation. There's always been an ability to bring in C language models through the Verilog programming language interface (PLI), but some vendors are looking at more efficient ways to create this link. C/C++ language approaches to testbench generation, such as Cadence Design Systems' TestBuilder library, may also appeal to some designers.

C is not going to replace Verilog for ASIC design, but who says we're going to design ASICs forever? I suspect that an increasing number of high-volume, fast time-to-market applications will be based on programmable or reconfigurable architectures.

This is where C language design can really shine. In the ultimate design system, you wouldn't even care what goes into hardware or software-you'd just write C/C++ code, and everything else would happen "under the hood." A C compiler that's intelligent enough to do all this would be a great prize indeed.