GCC 5.2 Compiler Benchmarks With ARM Cortex-A57 A Mixed Bag

Following Monday's NVIDIA Jetson TX1 performance overview one of the first follow-up tests I wanted to carry out was to see how the performance would evolve if using a newer compiler than what's shipped in Ubuntu 14.04. This current long-term support release ships GCC 4.8 while out since then was GCC 4.9 and now GCC 5.2.1 with GCC 6 coming in just a few months.

In this article are some benchmarks using the Jetson TX1 when running open-source tests using the stock GCC 4.8.4 compiler and then trying out GCC 4.9.3 and GCC 5.2.1. The same compiler flags were used each time when building the benchmarks under each of the different compilers using the automated Phoronix Test Suite. GCC 4.9 and GCC 5.2 were obtained from the Ubuntu Toolchain PPA. All tests are built on the Jetson TX1 without any cross-compilation or other steps.

The Jetson TX1 was left in the same configuration during testing. Part of the reason for wanting to do this compiler testing is due to the ever-maturing work still within GCC (and LLVM/Clang) for greater ARM performance. Under the stock GCC 4.8 compiler, with a few benchmarks I was also occasionally running into bugs:

This article is just for looking at GCC 4.8.4 vs. GCC 4.9.3 vs. GCC 5.2.1 while in a follow-up article will be some GCC vs. Clang compiler benchmarks for this Tegra X1 development board with four Cortex-A57 SoCs.