diff options
author | Vedant Kumar <vsk@apple.com> | 2019-10-21 11:48:38 -0700 |
---|---|---|
committer | Vedant Kumar <vsk@apple.com> | 2019-12-04 10:10:55 -0800 |
commit | e18531595bba495946aa52c0a16b9f9238cff8bc (patch) | |
tree | 489f93c6cda69f1e7255d0aa0fc596e466f30628 /llvm/test/tools/llvm-cov | |
parent | a6febd53c94afea4fa88503328ad800d9c3ab1a2 (diff) | |
download | bcm5719-llvm-e18531595bba495946aa52c0a16b9f9238cff8bc.tar.gz bcm5719-llvm-e18531595bba495946aa52c0a16b9f9238cff8bc.zip |
[Coverage] Revise format to reduce binary size
Revise the coverage mapping format to reduce binary size by:
1. Naming function records and marking them `linkonce_odr`, and
2. Compressing filenames.
This shrinks the size of llc's coverage segment by 82% (334MB -> 62MB)
and speeds up end-to-end single-threaded report generation by 10%. For
reference the compressed name data in llc is 81MB (__llvm_prf_names).
Rationale for changes to the format:
- With the current format, most coverage function records are discarded.
E.g., more than 97% of the records in llc are *duplicate* placeholders
for functions visible-but-not-used in TUs. Placeholders *are* used to
show under-covered functions, but duplicate placeholders waste space.
- We reached general consensus about giving (1) a try at the 2017 code
coverage BoF [1]. The thinking was that using `linkonce_odr` to merge
duplicates is simpler than alternatives like teaching build systems
about a coverage-aware database/module/etc on the side.
- Revising the format is expensive due to the backwards compatibility
requirement, so we might as well compress filenames while we're at it.
This shrinks the encoded filenames in llc by 86% (12MB -> 1.6MB).
See CoverageMappingFormat.rst for the details on what exactly has
changed.
Fixes PR34533 [2], hopefully.
[1] http://lists.llvm.org/pipermail/llvm-dev/2017-October/118428.html
[2] https://bugs.llvm.org/show_bug.cgi?id=34533
Differential Revision: https://reviews.llvm.org/D69471
Diffstat (limited to 'llvm/test/tools/llvm-cov')
-rwxr-xr-x | llvm/test/tools/llvm-cov/Inputs/binary-formats.v3.macho64l | bin | 0 -> 64456 bytes | |||
-rw-r--r-- | llvm/test/tools/llvm-cov/binary-formats.c | 1 |
2 files changed, 1 insertions, 0 deletions
diff --git a/llvm/test/tools/llvm-cov/Inputs/binary-formats.v3.macho64l b/llvm/test/tools/llvm-cov/Inputs/binary-formats.v3.macho64l Binary files differnew file mode 100755 index 00000000000..7da1822483e --- /dev/null +++ b/llvm/test/tools/llvm-cov/Inputs/binary-formats.v3.macho64l diff --git a/llvm/test/tools/llvm-cov/binary-formats.c b/llvm/test/tools/llvm-cov/binary-formats.c index 8836971d423..009583c00bc 100644 --- a/llvm/test/tools/llvm-cov/binary-formats.c +++ b/llvm/test/tools/llvm-cov/binary-formats.c @@ -7,5 +7,6 @@ int main(int argc, const char *argv[]) {} // RUN: llvm-cov show %S/Inputs/binary-formats.macho32l -instr-profile %t.profdata -path-equivalence=/tmp,%S %s | FileCheck %s // RUN: llvm-cov show %S/Inputs/binary-formats.macho64l -instr-profile %t.profdata -path-equivalence=/tmp,%S %s | FileCheck %s // RUN: llvm-cov show %S/Inputs/binary-formats.macho32b -instr-profile %t.profdata -path-equivalence=/tmp,%S %s | FileCheck %s +// RUN: llvm-cov show %S/Inputs/binary-formats.v3.macho64l -instr-profile %t.profdata -path-equivalence=/tmp,%S %s | FileCheck %s // RUN: llvm-cov export %S/Inputs/binary-formats.macho64l -instr-profile %t.profdata | FileCheck %S/Inputs/binary-formats.canonical.json |