When all 8 cores of the FX-8350 are fully saturated, it can be up to ~20% faster than the i5-4590.
When workloads only actually saturate up to 4 cores, the i5-4590 is up to ~50% faster than the FX-8350.
real-time workloads have very real inter-core scaling limitations. You can lean more about this with a Google search for "amdahls law"
What this means, is that, even though the FX-8350 can achieve higher total execution throughput than the i5-4590, that performance advantage will have a very difficult time ever translating to better performance in a game engine. Real-time workloads scale better with core performance than with core count, and in this case, "clock speed" is not a useful measure of core performance as these architectures are radically different. (A piledriver "core" is tiny compared to a haswell core).