Running GCC 4.9 On AMD's AM1 Kabini With Jaguar Cores
Using the AMD Athlon 5350 AM1 APU with its four "Jaguar" cores operating at 2.05GHz, I ran some benchmarks from Ubuntu 14.04 Linux comparing the performance of binaries compiled under GCC 4.8.2 and this week's GCC 4.9.0 RC1. Is GCC 4.9 better able to exploit the potential out of AMD's Jaguar microarchitecture? Let's see.
initial support for AMD's Jaguar and there's no explicitly marked Jaguar improvements for the upcoming GCC 4.9 release, but there's a lot of improvements overall for this annual update to the GNU Compiler Collection, so I decided to see how it affects these socketed Kabini APUs. Benchmarks from the other Athlon and Semprons are still forthcoming and on Thursday I delivered some preliminary results when upgrading the Linux kernel and Mesa, for those that haven't checked back since the AM1 Platform launch earlier in the week.
With this testing of the AMD Athlon 5350 quad-core 2.05GHz APU I compared the performance using GCC 4.8.2 and 4.9.0 RC1 built from source in the same configuration. I then re-built all of the Phoronix Test Suite tests with the same CFLAGS/CXXFLAGS between compiler releases. The testing is quite straightforward and similar to other Phoronix compiler comparisons in the past.
Most of the tests show only very minor changes in performance for the Athlon 5350 between GCC 4.8.2 and GCC 4.9.0 RC1.
GCC 4.9.0 RC1 looks like it's compiling binaries faster though then with GCC 4.8.2.
See all the rest of the benchmark results within OpenBenchmarking.org's 1404126-KH-GCC48TUNE46.