[PATCH] [Unroll] Rework the naming and structure of the new unroll heuristics.

[Chandler Carruth]

Hi mzolotukhin, hfinkel,

The new naming is (to me) much easier to understand. Here is a summary
of the new state of the world:

- '*Threshold' is the base line threshold for full unrolling. It is
  measured against the estimated unrolled cost as computed by
  getUserCost in TTI (or CodeMetrics, etc). We will exceed this
  threshold when unrolling loops where unrolling exposes a significant
  degree of simplification of the logic within the loop.
- '*MaxThreshold' is an absolute cap computed the same way as
  '*Threshold' but is never exceeded even due to profitable
  circumstanctes.
- '*PercentCostSavedThreshold' is the percentage of the loop's estimated
  dynamic execution cost which needs to be saved by unrolling to allow
  full unrolling in excess of the '*Threshold'.

When actually analyzing the loop, we now produce both an estimated
unrolled cost, and an estimated rolled cost. The rolled cost is notably
a dynamic estimate based on our analysis of the expected execution of
each iteration.

These estimates are pretty bad still, but we can make them much better,
and (to me) it is much more clear *how* to make them better when they
have reasonably descriptive names. For example, we may want to apply
estimated (from heuristics or profiles) dynamic execution weights to the
*dynamic* cost estimates. If we start doing that, we would also need to
track the static unrolled cost and the dynamic unrolled cost, as only
the latter could reasonably be weighted by profile information.

This patch is sadly not without functionality change for the new unroll
analysis logic. Buried in the heuristic management were several things
that surprised me. For example, we never subtracted the optimized
instruction count off when comparing against the unroll heursistics!
I don't know if this just got lost somewhere along the way or what, but
with the new accounting of things, this is much easier to keep track of
and we use the post-simplification cost estimate to compare to the
thresholds, and use the dynamic cost reduction ratio to select whether
we can exceed the baseline threshold.

My next series of patches will significantly improve the cost estimation
by handling dead control flows, weighting dynamic cost by the profile
(which both will amplify simplifications of nested loops and discount
simplifications of heavily predicated paths), and trying to track
instructions that are unnecessary after unrolling.

http://reviews.llvm.org/D9966

Files:
  include/llvm/Analysis/TargetTransformInfo.h
  lib/Transforms/Scalar/LoopUnrollPass.cpp
  test/Transforms/LoopUnroll/full-unroll-bad-geps.ll
  test/Transforms/LoopUnroll/full-unroll-heuristics.ll

EMAIL PREFERENCES
  http://reviews.llvm.org/settings/panel/emailpreferences/
-------------- next part --------------
A non-text attachment was scrubbed...
Name: D9966.26392.patch
Type: text/x-patch
Size: 21767 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20150525/16ecb3a6/attachment.bin>