News for hackers. Stuff that matters.
Re:STOKE – a stochastic optimizer for x86_64 assembly by erichocean on Tuesday January 19, 2016 @11:20PM (#10936068)
It seems crazy that stochastic optimization written by a couple of grad students is better than 30 years of the very best programmers writing optimizing compilers by hand.

Take something like Inria's certified C compiler, then run STOKE (with loop optimizations available) and you've got a provably correct optimizing compiler that's as fast or faster than GCC, Clang, VS, and Intel but without the compilation bugs.

Really cool stuff.

Re:STOKE – a stochastic optimizer for x86_64 assembly by UniQP on Wednesday January 20, 2016 @05:23AM (#10937118)
There are a lot of compiler optimizations that uses global analyses/optimizations that are not covered by optimizing a sequence of assembly instructions. Thus, I doubt that CompCert + STOKE would reach the code quality of production compilers.

Nevertheless, I agree that superoptimizers ofter find optimizations that are not considered by the compiler engineers. Another good example is the Optgen tool [1] that also identified missing (machine-independent) optimizations in all production compilers.

[1] http://pp.ipd.kit.edu/optgen/