Bulldozer is a codename that has been tossed around for years now and finally we get to find out what’s behind the name. The Bulldozer-based desktop processors are based on the "Zambezi" 32nm architecture and will feature up to 8-cores. This means it features the world’s first 8-core desktop solution which has been designed from the ground up.

AMD is kick starting the new FX line-up with seven new processors, which include 8-core, 6-core and 4-core models. The flagship processor is the FX-8150 which features a base clock of 3.6GHz with a Turbo Core clock of 3.9GHz and a Max Turbo clock of 4.2GHz.

The FX-8150 has a TDP headroom of 125 watts and of course it features 8-cores. However it also features an 8MB L2 cache along with an 8MB L3 cache, which is standard across all 8-core processors. Accompanying the FX-8150 is the FX-8120 and FX-8100 processors, the FX-8120 features a base clock of 3.1GHz while the FX-8100 works at 2.8GHz.

AMD has set the MSRP of the FX-8150 at just $245, which is only slightly more than the asking price of the Core i5-2500K, while it is considerably less than the Core i7-2600K. The FX-8120 is cheaper again at just $205, which is amazing given this 8-core processor is cheaper than the Core i5-2500K.

Along with the three 8-core processors is a single 6-core known as the FX-6100, which comes clocked at 3.3GHz with a Turbo Core frequency of 3.6GHz and a Max Turbo clock of 3.9GHz. While the L3 cache remains the same at 8MB, the L2 cache has been reduced to 6MB as 1MB is allocated per core. The FX-6100 features an MSRP of just $175, which seems like an incredible deal given the Phenom II X6 1100T currently costs $190 and the Core i5-2600K even more at $220.

Finally there are three 4-core models, which include the FX-4170, FX-B4150 and FX-4100. These quad-core parts feature a 4MB L2 cache along with the full 8MB L3 cache. The FX-4170 is the highest clocked Bulldozer processor with a standard operating frequency of 4.2GHz, that said the Turbo Core feature is disabled and the Max Turbo frequency is just 100MHz above the base clock.

Bulldozer is designed to provide the perfect balance between performance, cost and power consumption for multi-threaded applications. It focuses on high-frequency and resource sharing to achieve optimal throughput for next generation applications. As mentioned previously, the AMD FX processors offer up to eight power-efficient cores. These represent the first generation of a new execution-core family from AMD (Family 15h).

The Bulldozer concept is based on a 2-core design which shares latency-tolerant functionality, smoothes bursty/inefficient usage, and provides dynamic resource allocation between threads. Each core has its own 16KB L1 cache along with a 1MB L2 cache, while the L3 cache is shared. The other units are now effectively shared between two cores and include: Fetch, Decode, Floating point pipelines, and the L2 cache.

So how does the new Bulldozer architecture perform and can it stand up to Intel’s latest Sandy Bridge processors? When testing we will be focusing on how the FX-8150 flagship processor compares to the Core i5-2500K and Core i7-2600K processors. Finally, we will be evaluating how the new FX processors compare not only in terms of performance, but more importantly value.

Hardware - AMD Phenom II X6 1100T (3.30GHz) - AMD Phenom II X4 980 (3.70GHz) - AMD FX-8150 (3.60GHz) - AMD FX-8120 (3.10GHz) - AMD FX-6100 (3.30GHz) - AMD FX-4170 (4.20GHz) - x2 4GB G.Skill DDR3-1866 (CAS 8-9-8-24) - Crucial m4 256GB (SATA 6Gb/s) - Gigabyte GeForce GTX 580 SOC (1.5GB) - Asrock Fatal1ty 990FX Professional (AMD 990FX) Software - Microsoft Windows 7 Ultimate SP1 64-bit - Nvidia Forceware 285.38

Hardware - Intel Core i7-975 Extreme Edition (3.33GHz) - Intel Core i7-920 (2.66GHz) - x3 2GB G.Skill DDR3-1600 (CAS 8-8-8-20) - Kingston HyperX 240GB (SATA 6Gb/s) - Gigabyte GeForce GTX 580 SOC (1.5GB) - Gigabyte G1.Sniper (Intel X58) Software - Microsoft Windows 7 Ultimate SP1 64-bit - Nvidia Forceware 285.38

Hardware - Intel Core i7-2600K - Intel Core i5-2500K - x2 4GB G.Skill DDR3-1866 (CAS 8-9-8-24) - Crucial m4 256GB (SATA 6Gb/s) - Gigabyte GeForce GTX 580 SOC (1.5GB) - Gigabyte G1.Sniper2 (Intel Z68) Software - Microsoft Windows 7 Ultimate SP1 64-bit - Nvidia Forceware 285.38

Hardware - Intel Core i5-750 - x2 4GB G.Skill DDR3-1600 (CAS 8-8-8-20) - Crucial m4 256GB (SATA 6Gb/s) - Gigabyte GeForce GTX 580 SOC (1.5GB) - Gigabyte P55A-UD7 (Intel P55) Software - Microsoft Windows 7 Ultimate SP1 64-bit - Nvidia Forceware 285.38

When compared to the Phenom II the new FX processors provide significantly more memory bandwidth. Read throughput has been increased by roughly 60%, while write performance is almost 40% greater. That said, while the bandwidth performance has been improved greatly, it is still considerably slower than the Intel Sandy Bridge processors.

The L2 cache performance has also been improved, at least in part when compared to the Phenom II range. Here the FX-8150 provided 30% more read throughput than the Phenom II X6 1100T, though having said that the write performance was slightly slower. Finally, once again when compared to the Sandy Bridge processors, Bulldozer is considerably slower when measuring L2 read/write performance.

The SPECviewperf v11 benchmark is clearly not optimized for processors with more than four cores, as the highly clocked 4.2GHz AMD FX-4170 came away with the win. Rendering an average of 10.56fps in the SolidWorks test, it was slightly faster than the FX-8120 and FX-8150 processors. When compared to the Phenom II X6 1100T the FX-8150 was 39% faster, while it also beat the Core i7-2600K by a 17% margin.

The Maya benchmark delivered similar performance trends. Again the AMD FX-4170 delivered the best performance, while the FX-8150 was not far behind. When compared to the Phenom II X6 1100T the FX-8150 was 55% faster, while it was just 7% faster than the Core i7-2600K.

The CINEBENCH R11.5 benchmark saw the AMD FX-8150 render an average of 48.96fps, making it 9% slower than the Core i5-2500K and Core i7-2600K processors. It was on the other hand 4% faster than the Phenom II X4 980 and 6% faster than the Phenom II X6 1100T.

The AMD FX-8150 performed exceptionally well when running the in-built WinRAR benchmark, as it delivered a multithreading result of 4287KB/s making it 18% faster than the Core i7-2600O and 32% faster than the Core i5-2500K. Moreover it was 52% faster than the Phenom II X6 1100T and 66% faster than the Phenom II X4 980.

The Bulldozer based FX-8150 delivered roughly the same performance as the Core i5-2500K and Core i7 920 when testing with Excel 2010, which is a giant leap when compared to the Phenom II processors. For example when compared to the Phenom II X6 1100T the FX-8150 was 30% faster.

When running the in-built WinRAR benchmark we found that the AMD FX-8150 was by far the fastest processor. However in our own custom WinRAR compression test the results are quite different. Here the FX-8150 took 118 seconds to complete the 700MB file compression test, making it 22% slower than the Core i7-2600K and 17% slower than the Core i5-2500K. That said, it was able to match the performance of the Core i7 920.

When compared to the Phenom II X6 1100T the FX-8150 provided a significant performance boost when testing with Adobe Photoshop CS5, as it completed the test 26% faster. This was enough to match the performance of the Core i5-2500K. That said, the FX-8150 was still 32% slower than the Core i7-2600K.

The FX-8120 delivered roughly the same performance as the FX-8150, while the FX-6100 was still 8% faster than the Phenom II X6 1100T. Oddly the FX-4170 was 4% slower than the Phenom II X4 980, which is surprising given they are both quad-core processors and the FX-4170 has a rather significant clock speed advantage.

When testing with Fritz Chess 12 using the built-in benchmark we found that the FX-8150 was able to deliver 11682 kilo nodes per second, making it just 1% faster than the Phenom II X6 1100T. That said it was also 12% slower than the Core i7-2600K and 10% faster than the Core i5-2500K.

Interestingly clock for clock the FX processors are likely no faster than the Phenom II in this test, as the FX-8120 was 6% slower than the Phenom II X6 1100T. More over the FX-4170 was 24% slower than the Phenom II X4 980 despite being clocked 14% higher.

Using HandBrake we found that the AMD FX-8150 was capable of 113.8fps, making it just 2% slower than the Core i5 2500K, while it was also 5% slower than the Core i7-2600K. When compared to the Phenom II X6 1100T the FX-8150 did deliver 15% more performance.

The lower clocked FX-8120 delivered the same performance as the Core i7 975 Extreme Edition, while the FX-6100 was 3% slower than the Phenom II X6 1100T. This means clock for clock the six-core battle was won by the older Phenom II X6. Furthermore, we find when comparing quad-core performance, that the Phenom II X4 980 was much faster than the FX-4170.

The x264 HD Benchmark 4.0 provided us with more interesting results. The Pass 1 test for example saw the FX-8150 deliver just 121.3fps, making it 2% slower than the Phenom II X6 1100T, 15% slower than the Core i5-2500K, and 17% slower than the Core i7-2600K.

The FX-8210 on the other hand was only 1% faster than the Phenom II X4 980, while the FX-4170 was 11% slower than the Phenom II X4 980.

However the Pass 2 test results were quite different, here the FX-8150 was the fastest processor tested delivering 37.1fps, beating the Core i7-2600K by half a percent. The FX-6100 was still 6% slower than the Phenom II X6 1100T, while the FX-4170 was also 15% slower than the Phenom II X4 980.

The last encoding benchmark uses the TMPGEnc 4.0 Xpress program where we measure the avi to mpg conversion time. Here the AMD FX-8150 took 8 minutes and 5 seconds to complete the task, making it 27% slower than the Core i5-2500K and 31% slower than the Core i7-2600K. The FX-8150, FX-8120 and FX-6100 all delivered roughly the same performance. The FX-4170 on the other hand was 2% slower than the Phenom II X6 1100T and 5% slower than the Phenom II X4 980.

The Dirt 3 results are interesting and if you look at the above graph you will notice the five top spots are filled with Intel processors, while AMD fills the bottom six slots. However if you look closer there is a difference between the 1680x1050 and 1920x1200 performance trends. While Intel clearly dominated at 1680x1050 (where you could say there was less of a GPU bottleneck) the AMD processors performed better at 1920x1200.

At 1920x1200 the FX-8150 matched the performance of the Core i7 2600K, while it beat the Core i5-2500K. The FX-8120 matched the Core i7 975 EE, while the FX-6100 matched the Core i7 920. The FX-4170 on the other hand was the slowest processor tested.

With the exception of the Phenom II X4 980, we find that again AMD fills the bottom of our graph when measuring 1680x1050 performance. However at 1920x1200 the FX-4170 was actually the fastest processor tested, delivering 88fps. The rest of the FX line-up rendered between 82–84fps, allowing them to deliver similar performance to the Core i5 and Core i7 processors.

The Witcher 2 also shows mixed results when comparing 1680x1050 and 1920x1200 performance data. At 1680x1050 the FX processors performed poorly, with the exception of the FX-8150 which managed 77.6fps, allowing it to deliver similar performance to that of the Core i7 920.

The 1920x1200 performance saw the FX-8150 deliver roughly 1fps more than the FX-8120 and FX-6100 processors. That put it in line with the Core i7 975 EE and 920 processors, while making them slower than the Phenom II X4 980 and Phenom II X6 1100T.

The Crysis 2 performance was also disappointing for the FX processors. This game does not benefit from having more than four threads available, and for that reason the FX-4170 delivered the best performance at 1680x1050. This also meant that the FX-8150, FX-8120 and FX-6100 processors were all slower than the Phenom II X4 980.

Power consumption is still an issue for AMD, as we found the FX-8150 sucking down 252 watts under full load. Although this means that it is 5% less hungry than the Core i7 920, it also used 52% more power than the Core i7-2600K, or 86 watts more. Idle consumption is also an issue, as 100 watts means that the FX-8150 still uses 32% more power (24 watts) than the Core i7-2600K.

The FX-6100 was slightly better than the Phenom II X6 1100T when measuring both idle and load consumption, though we should stress only slightly. It is a similar situation when comparing the FX-4170 and Phenom II X4 980.

Overclocking the FX-8150 from the standard 3.6GHz to 4.4GHz boosted the Dirt 3 performance at 1680x1050 by 4%. The 1920x1200 performance on the other hand remained the same at 92fps.

The 4.4GHz overclocked delivered just under 3% more performance in the Pass 1 test and 12% more for the Pass 2 test. This allowed the FX-8150 to overtake the Phenom II X6 1100T, while it was still much slower than the Core i5-2500K and Core i7-2600K processors.

The CINEBENCH R11.5 OpenGL test saw no performance gain when overclocking the FX-8150 processor, while the CPU score was increased by 17%.

There are a few reasons for our disappointment with Bulldozers performance. Firstly the FX-8150 really struggled to handle the Core i5-2500K in a manner that we thought it should. If you look at our application benchmarks, such as Excel 2010 and Photoshop CS5, the FX-8150 was only just able to match the Core i5-2500K, while it was still much slower when testing with WinRAR.

Although it could be considered a great success, matching the performance of a high performance processor such as the Core i5-2500K, keep in mind the FX-8150 is more expensive and does consume considerably more power. Another factor to consider is that the Core i5-2500K features just four cores with four threads, whereas the FX-8150 has eight cores at its disposal.

Even so, when we ran our encoding benchmarks, which focus on tests capable of using all eight cores, the FX-8150 didn’t exactly shine. The Core i5-2500K was slightly faster when using HandBrake, while it was worlds faster in TMPGEnc XPress, not to mention it blasted the FX-8150 in the x264 HD Benchmark. To think that the FX-8150 had twice as many cores to play with and it still came up short, really speaks volumes about the core efficiency of the Sandy Bridge processors.

When it comes to real-world gaming using realistic quality settings at resolutions gamers are going to play at, either processor will suffice. Still, the Core i5-2500K was again the faster option of the two, leaving the FX-8150 unable to claim a victory.

The FX-8150 is certainly cheap at $245 for an eight-core processor, but if you break down the performance the Core i5-2500K still seems like the better deal. The FX-8120 on the other hand, which is essentially the same processor as the FX-8150 as both are fully unlocked, costs just $205 and at this price is cheaper than the Core i5-2500K.

Picking between the Core i5-2500K and the FX-8120 is a much harder decision, and depending on your needs you could really go either way. For gaming we would probably stick with the Core i5-2500K for now, but those looking at using heavily threaded programs the FX-8120 could be the way to go.

Then there is the cheaper $165 six-core FX-6100 to consider, and frankly we were most disappointed with this processor. While it did show strong gains over the Phenom II X6 1100T in programs such as Excel, WinRAR and Photoshop, it was considerably weaker when testing with Fritz Chess 12. Moreover it was slower in the HandBrake and x264 HD Benchmark, while it was also slower in virtually every game we threw at it.

The FX-6100 is so much slower than the Core i5-2500K that there is really no point in making that comparison. At $165 it is priced to compete with the Core i5-2300 ($180) or the Core i3-2130 ($150), so we will have to look into making that comparison shortly.

Disabling half the cores of the FX-8150 and overclocking it to 4.2GHz with a turbo clock of 4.3GHz to mimic the FX-4170 provided us with unbelievably poor performance. Although this configuration should represent the performance of the FX-4170 very accurately, it is hard to believe that the quad-core version of Bulldozer will be so much slower than existing Phenom II X4 processors in most tests.

Overclocking performance is not all that fantastic either, as we were able to push the FX-8150 to just 4.4GHz on air. When compared to the 4.1GHz of our Phenom II X6 1100T that’s not bad, but when compared to the 5.2GHz possible when using the Core i5-2500K or Core i7-2600K it’s not great. Granted this extreme overclock has only been possible on the Asus Maximus IV Extreme-Z, although all other P67 and Z68 motherboards reach at least 4.7–4.8GHz.

Overall we have been disappointed with the Bulldozer launch as we really were hoping for a lot more from the eight-core processors. After all this time it is disappointing to find that these new processors do little to improve AMD’s situation, as they struggled to compete with the now 9 month old Sandy Bridge processors. Moreover, in many cases the Phenom II, which is now well over 18 months old, was able to deliver better performance.

Still, this is just the start for Bulldozer, and there is much more to be seen from the FX range. Things can certainly improve and we are interested to see how the FX processors handle the upcoming Battlefield 3 video game.