forked from OSchip/llvm-project
d275a06487
Summary: AutoFDO compilation has two places that do inlining - the sample profile loader that does inlining with context sensitive profile, and the regular inliner as CGSCC pass. Ideally we want most inlining to come from sample profile loader as that is driven by context sensitive profile and also retains context sensitivity after inlining. However the reality is most of the inlining actually happens during regular inliner. To track the number of inline instances from sample profile loader and help move more inlining to sample profile loader, I'm adding statistics and optimization remarks for sample profile loader's inlining. Reviewers: wmi, davidxl Subscribers: hiraditya, llvm-commits Tags: #llvm Differential Revision: https://reviews.llvm.org/D70584 |
||
|---|---|---|
| .. | ||
| Inputs | ||
| branch.ll | ||
| calls.ll | ||
| cold-indirect-call.ll | ||
| compressed-profile-symbol-list.ll | ||
| cov-zero-samples.ll | ||
| coverage-warning.ll | ||
| discriminator.ll | ||
| early-inline.ll | ||
| entry_counts.ll | ||
| entry_counts_cold.ll | ||
| entry_counts_missing_dbginfo.ll | ||
| flattened.ll | ||
| fnptr.ll | ||
| function_metadata.ll | ||
| gcc-simple.ll | ||
| indirect-call-gcc.ll | ||
| indirect-call.ll | ||
| inline-act.ll | ||
| inline-callee-update.ll | ||
| inline-cold-callsite-samplepgo.ll | ||
| inline-cold.ll | ||
| inline-combine.ll | ||
| inline-coverage.ll | ||
| inline-mergeprof.ll | ||
| inline-stats.ll | ||
| inline-topdown.ll | ||
| inline.ll | ||
| nodebug.ll | ||
| nolocinfo.ll | ||
| offset.ll | ||
| profile-format-compress.ll | ||
| profile-format.ll | ||
| profile-sample-accurate.ll | ||
| propagate.ll | ||
| remap.ll | ||
| remarks.ll | ||
| section-accurate-samplepgo.ll | ||
| summary.ll | ||
| syntax.ll | ||
| uncompressed-profile-symbol-list.ll | ||
| warm-inline-instance.ll | ||