Re: GSoC proposal: partial redundancy elimination

Thanks for your feedback.

I've posted a new version at the same location
(https://gist.github.com/bcdarwin/94d9dbb34e4d1c47b40f). If you have
more suggestions, let me know.

Regards,

Ben

On 1 May 2013 22:08, Matthew Fluet <[email protected]> wrote:

> Ben,
>
> Looks like a good start for the proposal.  It would be good to develop the
> timeline in a little more detail.  You'll probably want some time at the
> beginning to learn MLton's SSA intermediate representation (which is not
> quite the "traditional" SSA presented in most compiler textbooks; see
> http://mlton.org/pipermail/mlton/2007-February/029597.html for some details)
> and consider how VanDrunen's algorithms will need to be modified for this
> SSA IL.  It might be reasonable to set as an early- or midterm-deliverable
> an GVN-PRE "analysis" pass -- that is, a pass that performs the necessary
> analysis of an SSA function to identify redundancies, but does not yet
> "transform" the function to eliminate the redundancies.  Similarly, I would
> suggest using as simple data structures as possible for the first
> implementation, and move to more sophisticated data structures at a later
> stage.  Extending to mutable objects would be nice; I'm not as familiar with
> that extension of GVN-PRE.  Benchmarking spills and memory accesses would be
> nice, but I agree that it isn't very easy to get at in MLton.

------------------------------------------------------------------------------
Introducing AppDynamics Lite, a free troubleshooting tool for Java/.NET
Get 100% visibility into your production application - at no cost.
Code-level diagnostics for performance bottlenecks with <2% overhead
Download for free and get started troubleshooting in minutes.
http://p.sf.net/sfu/appdyn_d2d_ap1