[LLVMdev] Optimization passes organization and tradeoffs

David Greene dag at cray.com
Wed May 21 14:47:51 PDT 2008


On Wednesday 21 May 2008 16:09, Nicolas Capens wrote:
> Hi David,
>
> Thanks for the info, but I'm not using any of the command line options. I'm
> dynamically generating code at run-time. So unless I'm missing some way to
> set these command line options at run-time I think I need another API for
> controlling register allocation and scheduling.

You'll just have to pick one and instantiate that Pass to run.  I was just
pointing out the command-line options so you could try them out.

> I was also under the impression that linear scan register allocation is
> quite cheap, faster than graph coloring. Or does it use an advanced
> variation (with aggressive coalescing), while other register allocation
> algorithms are more straightforward?

Yes, Linear Scan is cheaper than graph coloring, but it is still more
expensive than the local spiller or simple spiller.

                                           -Dave



More information about the llvm-dev mailing list