Hardware Previews / AMD FX-8150 Preview 1: Linux Performance (Results: x264 HD, C-Ray, TTSIOD 3D Renderer, SciMark 2, PyBench)

Everybody knows x264, the freeware used by a lot of people to encode their videos to H.264/AVC Pro, the most common video codec for high definition content and archiving.
Like most of the video encoders out there, x264 is multi-threaded and can take advantage of all the cores of our processors, no matter how many, it will "eat them up".
If you're doing a 2-pass encoding you'll notice that the first pass utilizes 2 to 4 cores, while the second pass utilizes all the cores. That's because of how the encoding algorithm of H.264/AVC Pro work.
One would expect the FX-8150 to be a tad slower than the Core i5 & Core i7 in the first pass ( due to the lower IPC and the fact that only 2 to 4 cores are being utilized, thus it has no advantage against the quad-core Core i5/i7's ) and leave the 2500K & the 2600K behind in the highly multi-threaded second pass.
However that is not entirely true. The FX-8150 is only clearly faster than the 2500K in total, while it's a lot slower than both Intel processors in the first pass ( it is even slower than the Phenom II x6 ), surpasses the 2500K in the 2nd pass and is marginally slower than the 2600K.
If encoding videos is your thing, it looks like the 2600K is a better choice than the FX-8150 ( and there are even faster solutions, namely the Core i7 980X & 990X, and the soon to be launched 6-core Core i7 3930K & 3960X for the X79 [ LGA2011 ] platform ).
As you can see from our overclocked 2500K test, it takes 4.5GHz for the 2500K to marginally surpass the FX-8150 in the 2nd pass, and enjoy a marginal advantage in total time taken to encode a video.

C-Ray is a volume raycasting rendering software ( raycasting is a raytracing method ), raytracing rendering programs are multi-threaded, and since raytracing is the future of CG and gaming, is a hot topic nowadays.
While we're not ready yet for real-time raytraced games, the industry uses some "scripted" techniques here and there, but also use raytracing for still images and short video animations.
This time the FX-8150 manages to surpass the 2600K too. It doesn't leave the Core i5/i7's miles behind, but still it's a rather healthy advantage.

TTSIOD 3D Renderer is a... 3D rendering software that supports OpenMP & Intel Threading Building Blocks with many different rendering modes.
The version included in the Phoronix Test Suite is entirely CPU based, there's no OpenGL ( GPU ) acceleration.
In TTSIOD 3D the FX-8150 is faster than the Core i5-2500K but it's much slower than the Core i7-2600K.
Overclocking the 2500K to 4.5GHz is enough for it to match the 2600K. A FX-8150 overclocked to 4.5GHz or so will probably be able to match the 2600K running at 3.4GHz.

SciMark 2 is a Java benchmark for scientific and numerical computing, it measures our CPU's FPU ( Floating Point Unit ) performance and gives us a final score in Mflops ( Millions of FLoating point Operations Per Second ). Who would use Java for scientific calculations is beyond me, but let's get past that and take a look at the results, shall we ?
Maybe we shouldn't. The FX-8150's performance in this test is very disappointing. It loses badly to 2500K. An overclocked 2600K would be untouchable in this test.

PyBench is a benchmark that measures the performance of our CPU in various functions like Nested For Loops, While, Bubble Search, Built-in Function Calls, etc in Python ( programming language ).
In PTS PyBench runs a collection of tests looped 3 or 4 times and reports the average total time.
Sadly the FX-8150 is the slowest CPU of the pack again, the 2500K and the 2600K are tied up, and the overclocked 2500K enjoys a healthy lead against the stock clocked 2600K and is nearly 100% faster than the stock FX-8150.

Let's go to page #4 now...