Wish the code was available to take a peek at. Talking about optimizing transformations, the Stalin scheme compiler is quite freakily good at that.It would have been interesting to run this code through Stalin and see what happens.
From the author Je rey Mark Siskind's research statement:
It uses the results of flow analysis to perform life-time analysis,
escape analysis, points-to analysis, and must-alias analysis. ...
It also uses the above analysis to support flow-directed region-based
storage management, where run-time garbage collection is replaced
with static allocation and deallocation on a per-abstract-value
and per-program-point basis. It also performs flow-directed
lightweight CPS conversion,.. to support extremely efficient first-class continuations.
It is quite remarkable that even without any type annotations Stalin can hold its own against a hand written C and often beat it as well.
Stalin download doesn't seem to have changed since 2006.
Is "Stalin can hold its own against a hand written C and often beat it" just your paraphrase of wikipedia's quote from a bare assertion in Jeffrey Siskind's Purdue research statement - or due you have some numbers to share?
The Sourceforge version might be more updated. But I doubt it will differ by much from what you have, the changes in the version are in the minor numbers.
Numbers circa 2008: take a look at section 6 in ftp://ftp.ecn.purdue.edu/qobi/TR-ECE-08-02.pdf and ftp://ftp.ecn.purdue.edu/qobi/TR-ECE-08-03.pdf These are numbers from StalinGrad. A automatic differentiation engine written in Stalin (hence the Grad).
With all that said, any reason to distrust research statements ? Those are usually taken seriously in academia. One only jeopardizes his position by bluffing.
From the author Je rey Mark Siskind's research statement:
It is quite remarkable that even without any type annotations Stalin can hold its own against a hand written C and often beat it as well.http://en.wikipedia.org/wiki/Stalin_(Scheme_implementation)
(Has escaped chrismonsantoization)