AMD Compiler Optimization Benchmarks With GCC 4.10 (GCC 5.0)

[sdack]

Just in case you are not being aware of it: btver1 and bdver1 are not the same type of CPU. btver1 stands for the Bobcat APU, which is probably the CPU inside your E-350. bdver1 refers to the Bulldozer CPU like the FX8150. The later is also the one with SSE4.1, SSE4.2 and AVX instructions, which are missing from the Bobcat. So while the article shows results for bdver1 will this be of no meaning for your E-350. You should expect results similar to the Phenom (aka K10, amdfam10 or Barcelona) with your E-350 though.

Since you are trying to squeeze some more performance out of your E-350s... take a look at ZSWAP for the Linux kernel if you have not done so already. It is said to work miracles. It inserts compression of memory pages just before these get swapped. You get less swapping and swapping will also be faster, and it feels like having more memory. You will either need to enable it first in the kernel configuration if it has not been done yet (it is marked as experimental) and then add/change in /etc/default/grub (or wherever your distro keeps the GRUB configuration):

GRUB_CMDLINE_LINUX_DEFAULT="zswap.max_pool_percent =25 zswap.compressor=lz4 zswap.enabled=1 quiet"

The ZSWAP feature needs to be enabled on boot and it is easiest to pass the arguments with grub to the kernel. Even when you have a desktop or server and think you do not need it should you at least take a look at it. With LZ4 as compressor/decompressor does it achieve compression ratios in the range from 1:2 to 1:4 and with speeds of 200MB/s-500MB/s for compression and 1GB/s-2GB/s for decompression. Even when you use a very fast SSD for swapping will it reduce the I/O and extend its life.

[Adarion]

And Michael still refuses to use Gentoo for benchmarks.

But now let's take these results in consideration for EVERY kind of benchmark out there. Int_el (R) (TM) optimized benchmarks being run on AMD CPUs/APUs - and then people blare about AMD being slower than intel ones. Well, maybe in some parts intel's raw performance is better. But it is not that much better than what all these hardware testers make you think. Especially when they are running their blob benchmark software.