GCC vs. Clang Compilers On The Intel Core i9 With Clear Linux
For those curious about the GCC versus LLVM Clang compilers with Intel's new Core i9 7900X, earlier this month I had ran some compiler benchmarks on this high-end processor.
I simply forgot to post these GCC vs. Clang i9-7900X benchmark results earlier, but here they are for those interested. The tests were done with the performance-oriented Clear Linux distribution.
GCC 7.1.1 and LLVM Clang 4.0.1 were the releases used. Tests were done with the CFLAGS/CXXFLAGS of "-O3 -march=haswell" and "-O3 -march=native" on both compilers for also looking at the impact of the Skylake-X tuning compared to Haswell. Skylake-X notably adds AVX-512 support.
Interestingly on fast storage such as this NVMe SSD system, using Clang now leads to faster performance with SQLite over GCC.
GCC leads in FFTW. The Haswell vs. Skylake-X tuning doesn't lead to a large performance difference.
In most of our tests done internally, the Skylake-X tuning doesn't lead to any major benefits at this point and using AVX-512 for some benchmarks can lead to slower performance.
See many more Clang vs. GCC compiler benchmarks on the Intel Core i9 via this OpenBenchmarking.org result file.
I simply forgot to post these GCC vs. Clang i9-7900X benchmark results earlier, but here they are for those interested. The tests were done with the performance-oriented Clear Linux distribution.
GCC 7.1.1 and LLVM Clang 4.0.1 were the releases used. Tests were done with the CFLAGS/CXXFLAGS of "-O3 -march=haswell" and "-O3 -march=native" on both compilers for also looking at the impact of the Skylake-X tuning compared to Haswell. Skylake-X notably adds AVX-512 support.
Interestingly on fast storage such as this NVMe SSD system, using Clang now leads to faster performance with SQLite over GCC.
GCC leads in FFTW. The Haswell vs. Skylake-X tuning doesn't lead to a large performance difference.
In most of our tests done internally, the Skylake-X tuning doesn't lead to any major benefits at this point and using AVX-512 for some benchmarks can lead to slower performance.
See many more Clang vs. GCC compiler benchmarks on the Intel Core i9 via this OpenBenchmarking.org result file.
11 Comments