[LLVMbugs] [Bug 21013] New: File takes relatively long in llvm but is parsed relatively quickly

bugzilla-daemon at llvm.org bugzilla-daemon at llvm.org
Sat Sep 20 16:17:01 PDT 2014


http://llvm.org/bugs/show_bug.cgi?id=21013

            Bug ID: 21013
           Summary: File takes relatively long in llvm but is parsed
                    relatively quickly
           Product: clang
           Version: trunk
          Hardware: PC
                OS: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: LLVM Codegen
          Assignee: unassignedclangbugs at nondot.org
          Reporter: nicolasweber at gmx.de
                CC: llvmbugs at cs.uiuc.edu
    Classification: Unclassified

Created attachment 13058
  --> http://llvm.org/bugs/attachment.cgi?id=13058&action=edit
repro

Instruments says that 2.3s is spent in clang::EmitBackendOutput and of that
2.2s in llvm::legacy::PassManagerImpl::run().

Parser::ParseTopLevelDecl takes only 0.8s.

Building with -O0 instead of -O2 -g makes the compile over twice as fast.

Filing with example, so folks can check if anything surprising is going on. If
not, to the WontFix it goes.

(Times on a 2 year old MBP, with a release build of clang as of today.)

The attached file is a preprocessed version of ninja's src/build_test.cc, it
can also be built by checking out ninja and running

  CXX=path/to/clang ./configure.py
  ninja ninja_test
  touch src/build_test.cc
  ninja ninja_test -v

in the checkout.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-bugs/attachments/20140920/9100320a/attachment.html>


More information about the llvm-bugs mailing list