Displaying 3 results from an estimated 3 matches for "functionpassbatcher".
2006 Jan 10
0
[LLVMdev] passmanager, significant rework idea...
...for the function passes. Since we don't need to run them in any
specific order, we might as well run the function passes in the order
that the callgraphsccpasses are being run in. If there is no
CallGraphSCCPassBatcher active, the passmanager would check to see if
there is a FunctionPassBatcher active, and if not it would create one.
Finally, note that all of this behavior is specific to "Module
PassManagers", so ideally none of the logic would be used/touched by the
FunctionPassManager stuff etc. Given this, it shouldn't be in something
generic like the base PassUnit c...
2006 Jan 10
3
[LLVMdev] passmanager, significant rework idea...
The patch below basically hammers out some ideas as to where I'd like
to take the passmanager in LLVM. I've tried thinking things through,
but I'm still a n00b, so some criticism would be more than welcome. =)
Starting from line 191 down. If you're wondering why I created a
patch, well that's because I found thinking in passmanagert.h the most
productive.
--
Regards.
2006 Jan 10
1
[LLVMdev] Re: passmanager, significant rework idea...
...asses. Since we don't need to run them in any
> specific order, we might as well run the function passes in the order
> that the callgraphsccpasses are being run in. If there is no
> CallGraphSCCPassBatcher active, the passmanager would check to see if
> there is a FunctionPassBatcher active, and if not it would create one.
As stated in the inliner improvements notes on your site, there seems
to be a significant effect based on the order of application and
traversal method used, wouldn't this hold true for other passes? I
could be wrong, but this seems weird to me. This i...