Re: End of optimization...

mayan@sandbridgetech.com wrote:
>
> If you have a radical new architecture _AND_ you've decided to write a
> compiler for it, clearly you're going to need people who understand how
> to develop new optimization techniques AND who know how to implement
> them - i.e. senior people.
>
> A Ph.D. would be a step in the right direction; but make sure its a
> Ph.D. with LOTS of implementation experience.


There is some movement towards processors with instruction sets
designed for machine generated code. There remains a lot of research
to be done on the design of this type of instruction set. The two
instruction sets of this type that I have looked at in significant
detail (One we wrote a C compiler for) both would been almost
impossible to hand code any significant application.


> Open source is a possibility, but you'd have to do a major
> module. Part of the problem there is that gcc's intermediate form is
> (was?) not particularily amenable to a large class of optimizations.


Open source is seductive, for retargeting processors of similar
architecture and applications it works quite well. For retargeting to
other application area's and architectures it saves little and costs a
lot, always competing with the earlier software design constraints. As
the previous comment points out the software architecture may prevent
newer optimizations and compiler implementations.


> According to a colleague of mine, this is borne out somewhat by the
> trends at recent conferences - there is not much optimization work
> being published. Instead, papers seem to be about areas such as
> program verification.


I agree with the general comment that not much is being published on
optimization. There is a lot of work being done. Most of the
optimization work I have recently seen has been development and
optimizing (pun intended) processor intensive optimization algorithms
  e.g. ram allocation in paged architectures .




> Also the faster CPUs and larger memories available have allowed
> techniques that were previously considered egregiously resource hungry
> to become acceptable and move from research labs into mainstream
> compilers.


I think we just made the same point.




Walter Banks

Return to the comp.compilers page.
Search the comp.compilers archives again.