Re: Back end generator alternatives?

Mr.E wrote:
> From my research BEG appears to be a great product but I emailed the
> company to find out about their product BEG and Firm. I got no
> response to a few emails so I called them in Germany. I was told the
> product could be from $10-25k depending upon the level of optimization
> I wanted. Based on google searches I found the cost over the years
> was anywhere from $250 to $2000. I'd like to use a backend generator
> to save myself time and frustration. I want the best but at the same
> time I'm not trying to get taken for a ride.


I fear you confused prices for compiler built with BEG with BEG itself
(Could you give an URL?). An early research version of BEG, generating
Modula-2, was available in binary form free of charge. It's more than
ten years old now. Since then we have developed BEG into a commercial
product, generating C, with many more features. Pricing depends on
options, desired optimization modules, amount of training and other
services, and pricing model (runtime royalties vs. flat fee); which
can add up to more than 25k.


Your emails apparently got eaten by our overeager spamfilter.
My sincere apologies.


>
> iBurg was used in the LCC compiler seemed like an excellent option. I
> love what Fraser & Hanson have done over the years with their
> generator but it isnt comercially available.


iBurg is a free implementation of code selectors very similar to BEGs
see [1]. BEG, however, generates complete back ends including local
and/or global register allocation, instruction scheduling, DAG
matching for SSA, and other optimizations etc.


iBurg is good at what it's doing. You, however, have to do
register allocation and other back end / optimization stuff yourself.


Helmut Emmelmann


[1] Christopher W. Fraser and David R. Hanson and Todd A. Proebsting:
Engineering Efficient Code Generators Using Tree Matching and Dynamic
Programming, TR-386-92, 1992

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