Testing LLVM Clang 3.5's Code Generation Optimizations
For those curious about the performance of LLVM Clang in its current development form when testing the common code generation options for optimizing the performance (and in some cases size) of the resulting binaries, here's some fresh compiler benchmarks.
Just as some extra benchmarks for the weekend while finishing out the month, I ran some new benchmarks comparing common optimization levels for LLVM/Clang with the latest 3.5 development code as of earlier this month. The configurations tested for this article included:
-O0
-O1
-O2
-Os
-Oz
-O3
-Ofast
-O3 -march=native
There aren't any breakthrough changes with the code generation optimizations for Clang 3.5 on x86_64, but simply here's some fresh benchmark results with an array of open-source C/C++ tests done from an Intel Core i7 4790K Devil's Canyon system running Ubuntu 14.04 with the Linux 3.16 development kernel. The code generation options were controlled through the CFLAGS/CXXFLAGS.
Take these reference compiler results as you wish or feel free to run your own compiler benchmarks using the Phoronix Test Suite.