llvm dev - Jun 2013 - [LLVMdev] RFC - Profile Guided Optimization in LLVM

On Wed, Jun 12, 2013 at 2:55 PM, Jakob Stoklund Olesen <stoklund at
2pi.dk>wrote:
> That sounds plausible to me. It seems like we might need a way of
> representing call graph profiling in addition to the existing branch
> probabilities?
>
Agreed. An important consideration here is WPO vs. LTO vs. TU-at-a-time
call graphs.

> FWIW, the greedy register allocator’s live range splitting algorithm is
> designed to consume profile information so it can push spill code into cold
> blocks. The primary interface is SpillPlacement::getBlockFrequency() which
> currently returns an estimate based on loop depth only.
>
It doesn't use MachineBlockFrequency? If it does, it will get a lot more
than loop depth: __builtin_expect, cold function attribute, and static
branch heuristics. If it doesn't it should, and then it will immediately
benefit from this.
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20130612/a851059d/attachment.html>

On Wed, Jun 12, 2013 at 3:05 PM, Chandler Carruth <chandlerc at
google.com>wrote:
> FWIW, the greedy register allocator’s live range splitting algorithm is
>> designed to consume profile information so it can push spill code into
cold
>> blocks. The primary interface is SpillPlacement::getBlockFrequency()
which
>> currently returns an estimate based on loop depth only.
>>
>
> It doesn't use MachineBlockFrequency?

Err, MachineBlockFrequencyInfo -- the analysis pass. Sorry.
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20130612/9bb1a359/attachment.html>

On Jun 12, 2013, at 3:05 PM, Chandler Carruth <chandlerc at google.com>
wrote:
> 
> On Wed, Jun 12, 2013 at 2:55 PM, Jakob Stoklund Olesen <stoklund at
2pi.dk> wrote:
> That sounds plausible to me. It seems like we might need a way of
representing call graph profiling in addition to the existing branch
probabilities?
> 
> Agreed. An important consideration here is WPO vs. LTO vs. TU-at-a-time
call graphs.
>  
> FWIW, the greedy register allocator’s live range splitting algorithm is
designed to consume profile information so it can push spill code into cold
blocks. The primary interface is SpillPlacement::getBlockFrequency() which
currently returns an estimate based on loop depth only.
> 
> It doesn't use MachineBlockFrequency? If it does, it will get a lot
more than loop depth: __builtin_expect, cold function attribute, and static
branch heuristics. If it doesn't it should, and then it will immediately
benefit from this.
It predates the block frequency interface. It just needs to be hooked up,
patches welcome. It would also be nice to remove the floating point computations
from the spill placement code.

Thanks,
/jakob

On Wed, Jun 12, 2013 at 3:10 PM, Jakob Stoklund Olesen <stoklund at
2pi.dk>wrote:
> It predates the block frequency interface. It just needs to be hooked up,
> patches welcome. It would also be nice to remove the floating point
> computations from the spill placement code.

Cool, if Diego doesn't beat me to it, I may send you a patch as that seems
easy and obviously beneficial.
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20130612/09e7857c/attachment.html>