Displaying 3 results from an estimated 3 matches for "rinard".
Did you mean:
minard
2015 Feb 05
5
[LLVMdev] RFC: Recursive inlining
...nrolling, and
> > Chandler suggested we should model it as such.
> >
> >
> > We originally thought he meant by this that the mechanics of inlining
> > would stay the same, but the heuristics would have to be changed.
> > This is the approach taken by Rugina & Rinard [1], which is almost
> > the sum of the literature on this topic I could find, which
> > surprised me. However, the actual suggestion was to remove the
> > recursion entirely, and replace it with a loop.
> >
> >
> > Before I get into the mechanics of this, I'...
2015 Feb 05
5
[LLVMdev] RFC: Recursive inlining
...f times) to inline. In this respect it is
similar to loop unrolling, and Chandler suggested we should model it as
such.
We originally thought he meant by this that the mechanics of inlining would
stay the same, but the heuristics would have to be changed. This is the
approach taken by Rugina & Rinard [1], which is almost the sum of the
literature on this topic I could find, which surprised me. However, the
actual suggestion was to remove the recursion entirely, and replace it with
a loop.
Before I get into the mechanics of this, I'd like to explain what the
possible optimization opportunit...
2015 Feb 18
5
[LLVMdev] RFC: Recursive inlining
...and
> > Chandler suggested we should model it as such.
> >
> >
> > We originally thought he meant by this that the mechanics of
> > inlining would stay the same, but the heuristics would have to be
> > changed.
> > This is the approach taken by Rugina & Rinard [1], which is almost
> > the sum of the literature on this topic I could find, which
> > surprised me. However, the actual suggestion was to remove the
> > recursion entirely, and replace it with a loop.
> >
> >
> > Before I get into the mechanics of this, I'...