Not really surprising.
Getting the most out of any processor requires processor-specific optimization. Unfortunately for AMD, Intel has the lion's share of the market, so developers pay more attention to getting software to run well on Intel processors. Some of the top tier games that get used for benchmarks have been hand-optimized for Intel, as have productivity applications such as video encoders and Photoshop. (The last two have also benefited historically from Intel having better SIMD implementations. That is probably still true. But an A-series AMD processor with properly optimized OpenCL code might be better still.)
Intel is in the developer tools business as well. They sell a compiler that generates code that is very good for Intel processors and very bad for AMD. Any application that is built with Intel tools is going to make AMD look bad.
Finally, there is the OS issue. Because of the way AMD used paired cores with some shared elements (cache and FPU), getting the most out of the FX series processors requires changes to the process scheduler. (The simplified version: threads of the same process and multiple instances of the same application should be assigned to paired cores; unconnected applications should be spread to different core pairs whenever possible. That maximizes the effectiveness of the shared cache. The shared FPU is of little concern unless you have applications that do math with long doubles; it can do two 64 bit operations simultaneously but only one 128 bit operation.) The most popular OS on the market, Windows 7, has not made the necessary adjustments, nor has any earlier version. Windows 8 and later have, as have recent Linux kernels. Mac OS probably has not, but Apple has never made a computer with an AMD processor so it isn't relevant unless you own a Hackintosh.