Intel Xeon Cascade Lake Compiler Performance - GCC 9/10 vs. LLVM Clang 8/9

At least for the newest Intel Xeon "Cascade Lake" processors, the LLVM Clang compiler is running incredibly well compared to the long-standing GNU Compiler Collection (GCC). Overall, LLVM clang is now nearly at performance parity to GCC 9 and the in-development GCC 10 compilers. Here are some Linux compiler benchmarks using the dual Intel Xeon Platinum 8280 server built around the Gigabyte S3461-3R0.

Of our many Cascade Lake benchmarks since its launch back in April, one area we hadn't looked at until now was the LLVM Clang compiler performance on these latest Xeon server processors. From an Ubuntu 19.04 host with the Linux 5.1 kernel, GCC 9.1.0, GCC 10.0.0 20190616, LLVM Clang 8.0.1, and LLVM Clang 9.0 SVN were all freshly built in their release/optimized mode. A variety of C/C++ benchmarks were then carried out while using "-O3 -march=cascadelake" as the CFLAGS/CXXFLAGS.

Across the dozens of benchmarks carried out using the Phoronix Test Suite it ended up being an incredibly competitive race between GCC and Clang, though there's still certain workloads very visibly favoring one compiler over the other. Here are those latest Intel Xeon compiler benchmark numbers.