archived-llvm

mirror of https://github.com/RPCS3/llvm.git synced 2026-01-31 01:25:19 +01:00

Files

Dehao Chen ded3597f2c Use the first instruction's count to estimate the funciton's entry frequency.

Summary: In the current implementation, we only have accurate profile count for standalone symbols. For inlined functions, we do not have entry count data because it's not available in LBR. In this patch, we use the first instruction's frequency to estimiate the function's entry count, especially for inlined functions. This may be inaccurate due to debug info in optimized code. However, this is a better estimate than the static 80/20 estimation we have in the current implementation.

Reviewers: tejohnson, davidxl

Reviewed By: tejohnson

Subscribers: sanjoy, llvm-commits, aprantl

Differential Revision: https://reviews.llvm.org/D38478

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@315369 91177308-0d34-0410-b5e6-96231b3b80d8

2017-10-10 21:13:50 +00:00

Inputs

Directly return promoted direct call instead of rely on stripPointerCast.

2017-10-06 17:04:55 +00:00

branch.ll

Do not set branch weight if the branch weight annotation is present.

2017-03-23 14:43:10 +00:00

calls.ll

Remove the sample pgo annotation heuristic that uses call count to annotate basic block count.

2017-03-06 17:49:59 +00:00

cov-zero-samples.ll

Canonicalize the representation of empty an expression in DIGlobalVariableExpression