forked from OSchip/llvm-project
7fb39dfa7c
Summary: In D49565/r337503, the type id record writing was fixed so that only referenced type ids were emitted into each per-module index for ThinLTO distributed builds. However, this still left an efficiency issue: each per-module index checked all type ids for membership in the referenced set, yielding O(M*N) performance (M indexes and N type ids). Change the TypeIdMap in the summary to be indexed by GUID, to facilitate correlating with type identifier GUIDs referenced in the function summary TypeIdInfo structures. This allowed simplifying other places where a map from type id GUID to type id map entry was previously being used to aid this correlation. Also fix AsmWriter code to handle the rare case of type id GUID collision. For a large internal application, this reduced the thin link time by almost 15%. Reviewers: pcc, vitalybuka Subscribers: mehdi_amini, inglorion, steven_wu, dexonsmith, llvm-commits Differential Revision: https://reviews.llvm.org/D51330 llvm-svn: 343021 |
||
|---|---|---|
| .. | ||
| Inputs | ||
| bad-read-from-vtable.ll | ||
| branch-funnel-threshold.ll | ||
| branch-funnel.ll | ||
| constant-arg.ll | ||
| devirt-single-impl-check.ll | ||
| devirt-single-impl.ll | ||
| expand-check.ll | ||
| export-nothing.ll | ||
| export-single-impl.ll | ||
| export-uniform-ret-val.ll | ||
| export-unique-ret-val.ll | ||
| export-unsuccessful-checked.ll | ||
| export-vcp.ll | ||
| import-indir.ll | ||
| import-no-dominating-assume.ll | ||
| import.ll | ||
| non-constant-vtable.ll | ||
| pointer-vtable.ll | ||
| soa-vtable.ll | ||
| struct-vtable.ll | ||
| uniform-retval-invoke.ll | ||
| uniform-retval.ll | ||
| unique-retval.ll | ||
| vcp-accesses-memory.ll | ||
| vcp-decl.ll | ||
| vcp-no-this.ll | ||
| vcp-non-constant-arg.ll | ||
| vcp-too-wide-ints.ll | ||
| vcp-type-mismatch.ll | ||
| vcp-uses-this.ll | ||
| virtual-const-prop-begin.ll | ||
| virtual-const-prop-check.ll | ||
| virtual-const-prop-end.ll | ||
| vtable-decl.ll | ||