AMD/Intel Benchmarks: Building The Mainline Linux x86_64 Kernel With LLVM Clang
With the Hackbench Linux kernel scheduler benchmark, using the Clang compiler to build Linux 5.3 only hurt the performance.
The PostgreSQL performance was about the same on the two kernel builds.
The semaphores performance was coming in slightly higher on the Clang built kernel.
The socket activity micro-benchmark meanwhile saw the best performance on both systems out of the GCC built kernel.
The context switching time also ended up being slightly higher with the Clang built kernel.
The Apache performance was edging slightly higher on the GCC built kernel.
When running 97 different tests across the two systems and the kernel builds from GCC and Clang, the race was very close. In looking at the geometric mean, the end result was a wash for the Core i9 9900K while the Ryzen 9 3900X performance was ever so slightly higher using GCC over Clang. to build the Linux 5.3 mainline kernel. There's still some shortcomings as mentioned to building the Linux kernel with Clang, but it's great that it's now largely working with the mainline Clang compiler and mainline Linux x86_64 kernel.
If you enjoyed this article consider joining Phoronix Premium to view this site ad-free, multi-page articles on a single page, and other benefits. PayPal or Stripe tips are also graciously accepted. Thanks for your support.