[Pipelines] Additional unrolling in LTO #536
Conversation
Some workloads require specific sequences of events to happen to fully simplify. This adds an extra full unrolling pass to help these cases on the cores with branch predictors. It helps produce simplified loops, which can then be SROA'd allowing further simplification, which can be important for performance. This is added under own flag - spending extra compile time to get extra performance on specific user request. Originally patch by David Green (david.green@arm.com)
|
Very minor nit: the first patch should start |
There was a problem hiding this comment.
LGTM, but when merging it'd be good to have the commit message mirror the commit message changes you just made to the patch file when squash-merging:
Some workloads require specific sequences of events to happen
to fully simplify. This adds an extra full unrolling pass to help these
cases on the cores with branch predictors. It helps produce simplified
loops, which can then be SROA'd allowing further simplification, which
can be important for performance.
The feature adds extra compile time to get extra performance and
is enabled by the opt flag 'extra-LTO-loop-unroll' (off by default).
Some workloads require specific sequences of events to happen
to fully simplify. This adds an extra full unrolling pass to help these
cases on the cores with branch predictors. It helps produce simplified
loops, which can then be SROA'd allowing further simplification, which
can be important for performance.
Feature adds extra compile time to get extra performance and
is enabled by the opt flag 'extra-LTO-loop-unroll' (off by default).
Original patch by David Green (david.green@arm.com)