[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]

Re: gEDA-user: VHDL Compiler



John Sheahan wrote:
in Europe, VHDL is the FPGA language of choice. In the US, it's about half-and-half. In Japan, they prefer Verilog. Apparently, in Japan they prefer to keep using what has worked well in the past.


I've worked with counterexamples to all of these - not that I'd dispute the general thrust much. For asic I would have guessed the USA leaned to verilog a tad. And Australia is VHDL. (all .001% of the world market)..

Perhaps the key - as you observe - is to emphasize translator tools to reduce the wastage wherever possible?
Unfortunately some things are hard to translate, although less so with the arrival of Verilog 2000, and less so for synthesisable code. Off the top of my head (I'm sure there are better examples) bidirectional transmission gates, the Verilog conditional operator where the result of the expression is 'x', 4-value logic vs. n-value logic, the general issue of VHDL's more sophisticated type system (enumerations, multi-dimensional arrays, records, etc).

Anyway, what does it really matter if we have two languages? Sometime soon we'll have a choice of a lot more - SystemVerilog, SystemC, 'e' when it becomes synthesisable, Handel-C if it ever takes off, and so on. There's no point reducing everything to the lowest common denominator. Imagine telling a programmer that they weren't allowed to write in C, C++, Java, ML, and a million other languages just because someone had already written Fortran. IMO, the more the merrier - it shows that the HW world is finally starting to take language design seriously, albeit 20 years too late.

Evan