J
Jan Decaluwe
Guest
On 03/18/2011 04:18 PM, rickman wrote:
as clichés can be.
Many HDLs have been proposed. Most were in the "close to hardware"
camp. They are mostly forgotten. The market clearly prefers
the "outsiders" that look more like "programming languages".
Of course, the market may be (temporarily) wrong, but to back that
up you better come up with very strong arguments and fresh ideas.
Programming is of course about languages, compilers and methodolgies.
However, a point frequently missed by hardware designers is that
there is a wide variety of sub-discplines and application domains,
with very different characteristics and constraints. The best
programmers know their application domain very well, they master
the coding patterns that work well for their purpose, and they
understand what they can expect from their compilers.
A good HDL designer should of course do exactly the same. He
should understand hardware design very well. But he should also
understand what kind of coding patterns are efficiently
handled by his compiler (synthesis tool). Focussing on
"describing hardware" is counter-productive, because it tends
to make designers blind for useful coding patterns.
Interpreting HDL design as a specific software discpline has
a tremendous advantage: you benefit from all advances in tools
and methodologies from the overall software world. The best
talent that gives us real breakthroughs is there - that is
just a matter of numbers.
We hardware designers are much less special than we often
like to think. By realizing that, we can reduce our development
costs and increase our productivity.
Jan
--
Jan Decaluwe - Resources bvba - http://www.jandecaluwe.com
Python as a HDL: http://www.myhdl.org
VHDL development, the modern way: http://www.sigasi.com
Analog design automation: http://www.mephisto-da.com
World-class digital design: http://www.easics.com
This is just a bunch of clichés, as misleading and boring
as clichés can be.
Many HDLs have been proposed. Most were in the "close to hardware"
camp. They are mostly forgotten. The market clearly prefers
the "outsiders" that look more like "programming languages".
Of course, the market may be (temporarily) wrong, but to back that
up you better come up with very strong arguments and fresh ideas.
Programming is of course about languages, compilers and methodolgies.
However, a point frequently missed by hardware designers is that
there is a wide variety of sub-discplines and application domains,
with very different characteristics and constraints. The best
programmers know their application domain very well, they master
the coding patterns that work well for their purpose, and they
understand what they can expect from their compilers.
A good HDL designer should of course do exactly the same. He
should understand hardware design very well. But he should also
understand what kind of coding patterns are efficiently
handled by his compiler (synthesis tool). Focussing on
"describing hardware" is counter-productive, because it tends
to make designers blind for useful coding patterns.
Interpreting HDL design as a specific software discpline has
a tremendous advantage: you benefit from all advances in tools
and methodologies from the overall software world. The best
talent that gives us real breakthroughs is there - that is
just a matter of numbers.
We hardware designers are much less special than we often
like to think. By realizing that, we can reduce our development
costs and increase our productivity.
Jan
--
Jan Decaluwe - Resources bvba - http://www.jandecaluwe.com
Python as a HDL: http://www.myhdl.org
VHDL development, the modern way: http://www.sigasi.com
Analog design automation: http://www.mephisto-da.com
World-class digital design: http://www.easics.com