https://bugs.llvm.org/show_bug.cgi?id=42357
Bug ID: 42357
Summary: Regression in SVN r363875, "[LFTR] Rename variable to
minimize confusion [NFC]"
Product: libraries
Version: trunk
Hardware: PC
OS: All
Status: NEW
Severity: enhancement
Priority: P
Component: Scalar Optimizations
Assignee: unassignedb...@nondot.org
Reporter: mar...@martin.st
CC: listm...@philipreames.com, llvm-bugs@lists.llvm.org
Created attachment 22130
--> https://bugs.llvm.org/attachment.cgi?id=22130&action=edit
Reproduction C code
Since SVN r363875, "[LFTR] Rename variable to minimize confusion [NFC]",
compiling GLEW for i686 mingw almost doubled in compile time, and produces an
output almost double in size:
$ time clang-prev -target i686-w64-mingw32 -c -O3 glew-preproc-i686.c -w && ls
-lh glew-preproc-i686.o
real 1m3.655s
user 1m3.325s
sys 0m0.331s
-rw-rw-r-- 1 martin martin 615K Jun 21 21:30 glew-preproc-i686.o
$ time clang-new -target i686-w64-mingw32 -c -O3 glew-preproc-i686.c -w && ls
-lh glew-preproc-i686.o
real 1m58.397s
user 1m57.998s
sys 0m0.400s
-rw-rw-r-- 1 martin martin 1.1M Jun 21 21:33 glew-preproc-i686.o
When optimizing the output from "clang -target i686-w64-mingw32 -S -O3
glew-preproc-i686.c -Xclang -disable-llvm-passes -emit-llvm", things behave
like this:
$ time opt-old -O3 glew-preproc-i686.ll -o glew-preproc-i686.bc && ls -lh
glew-preproc-i686.bc
real 0m56.537s
user 0m56.286s
sys 0m0.253s
-rw-rw-r-- 1 martin martin 2.5M Jun 21 21:38 glew-preproc-i686.bc
$ time llc glew-preproc-i686.bc && ls -lh glew-preproc-i686.s
real 0m8.260s
user 0m8.159s
sys 0m0.100s
-rw-rw-r-- 1 martin martin 5.0M Jun 21 21:42 glew-preproc-i686.s
$ time opt-new -O3 glew-preproc-i686.ll -o glew-preproc-i686.bc && ls -lh
glew-preproc-i686.bc
real 1m19.033s
user 1m17.550s
sys 0m0.252s
-rw-rw-r-- 1 martin martin 3.8M Jun 21 21:40 glew-preproc-i686.bc
$ time llc glew-preproc-i686.bc && ls -lh glew-preproc-i686.s
real 0m37.613s
user 0m37.341s
sys 0m0.273s
-rw-rw-r-- 1 martin martin 11M Jun 21 21:45 glew-preproc-i686.s
--
You are receiving this mail because:
You are on the CC list for the bug._______________________________________________
llvm-bugs mailing list
llvm-bugs@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
