AMD Piledriver FX-4350, FX-6350 & FX-8350 Review

Brand: AMD
Model: Piledriver FX-4350, FX-6350 & FX-8350
Website: http://www.amd.com/us/products/desktop/processors/amdfx/Pages/amdfx.aspx
RRP FX-4350: £95 (At time of the review)
RRP FX-6350: £105 (At time of the review)
RRP FX-8350: £155 (At time of the review)

I am going to be taking a look at AMDs latest “Piledriver” line-up to see how they fare in todays review. I will be taking a look at the top of the line components from the four, six and eight core lines. Whilst this review will definitely demonstrate how much extra performance the extra cores offer over their smaller siblings when a heavily multithreaded application is being used, it will also demonstrate how well the additional cores scale. For example, going from a four core to an eight core CPU will not deliver a 100% improvement in performance. AMD had the best scaling in the Phenom II days. Since then, they haven’t been able to get the cores to scale quite as well although the reason for that is unclear to me.

AMD have been playing catch-up with Intel for a long time now. Are they finally beginning to close the gap in terms of IPC (Instructions Per Cycle) and performance in general? In order to find out, I will be using a mid-ranged Intel Core i5-3570K to see how well the AMD CPUs stack up against Intel’s “mid-range” processor. The reason for this is simple and that is the i5-3570Ks price point. Whilst it is more expensive when compared to the AMD offerings, it is a quad core with four threads which should in theory make the FX-8350 shine in heavily multithreaded applications. However, will that be the case? Or will Intel still lead the way even though it has half the number of cores?

AMD have a pretty longwinded features and specifications list so without further ado, let’s examine the specs and then get started with the review.

AMD FX Processors

We call it the new AMD FX 8-Core Processor Black Edition and it’s unlocked for your overclocking pleasure.¹ Experience unmatched multitasking and pure core performance with the industry’s first 32nm 8-core desktop processor. Get the speed you crave with AMD Turbo CORE Technology to push your core frequencies to the limit when you need it most. Go beyond the limits of maximum speed with easy-to-use AMD OverDrive™ and AMD Catalyst Control Center™ software suites. But the best part of all? You’ll get all this impressive performance at an unbelievable price. You’ll be asking yourself “what competition?” in no time.

AMD FX 8-Core Processors

• The industry’s first and only native 8-core desktop processor for unmatched multitasking and pure core performance with all-new “Bulldozer” architecture.
• New 32 nanometer die shrink designed to reduce leakage for improved efficiency, increased clock rate headroom and better thermals.

AMD Turbo CORE Technology

• The AMD FX Processors come equipped with AMD Turbo CORE Technology. AMD Turbo CORE Technology is a performance boosting technology that helps increase performance on the applications that need it the most.

New Instruction Capabilities

• AVX
  • Advanced Vector Extensions increase parallelism tailored for scientific and 3D applications that use heavy floating point calculations
• FMA4 and XOP
  • Floating Point Vector Multiply -Accumulate  improves throughput and performance on many vector functions (integer and floating point)
• AES
  • Advanced Encryption Standard noticeably increase performance on the latest encryption applications like TrueCrypt and benchmarks like PCMark

AMD Balanced Smart Cache

• Shared L3 cache ( up to 8MB)
  • Improved scheduling and pre-fetch capabilities
  • 64-ways (16-ways/sub-cache)
  • Increased data queue sizes
  • Coherency for 8-cores

AMD Wide Floating Point Accelerator

• Shared FP Scheduler
• Dual 128-bit Floating point engines – capable of teaming together for 256-bit AVX instructions or operating separately with each core.

HyperTransport™ Technology

• One 16-bit link at up to 5600MT/s
• Up to 8.0GB/s HyperTransport™ I/O bandwidth; Up to 16GB/s in HyperTransport Generation 3.0 mode
• Up to 37GB/s total delivered processor-to-system bandwidth (HyperTransport bus + memory bus)

Benefit: Quick access times to system I/O for better performance.

Integrated DRAM Controller with AMD Memory Optimizer Technology

• A high-bandwidth, low-latency integrated memory controller
• Supports up to DDR3-1866²
• Supports new low voltage memories of 1.35V and 1.2V
• Up to 29.9GB/s memory bandwidth for DDR3
• New Pre-Fetcher improvements
• Direct communications to each core in Dual-Core module (APIC registers in each core)

Benefit: Optimized memory controller to feed more cores

AMD Virtualization™ (AMD-V™) Technology with IOMMU

• Silicon feature-set enhancements designed to improve the performance, reliability, and security of existing and future virtualization environments by allowing virtualized applications with direct and rapid access to their allocated memory.
• IOMMU is an extension to AMD64 architecture to support address translation and access protection on DMA transfers
  • Security for User Level application and Virtual Machine guest operating system
    • Address translation and access control
    • Device isolation
    • Device assignment in virtualized systems
    • Security & trusted boot support
    • Unified interrupt management

Benefit: Helps virtualization software to run more securely and efficiently enabling a better experience when dealing with virtual systems

AMD PowerNow!™ Technology (Cool’n’Quiet™ Technology)

• Enhanced power management features which automatically and instantaneously adjusts performance states and features based on processor performance requirements
• C6 power state for cache flush, and voltage down individual core
• CC6 power state allows all cores in C6 to power even lower
  • For quieter operation and reduced power requirements
• Separate memory controller power control
• IO-based c-state interface
• Works automatically without the need for drivers or BIOS enablement.
• Power can be switched on or off within a single clock cycle, saving energy with no impact to performance.

Benefit: Helps users get more efficient performance by dynamically activating or turning off parts of the processor.

Enthusiasts around the globe are ever hopeful that AMD will switch to LGA for their socket system soon but it is unlikely to happen in the near future. Remember the days when you’d pull out your CPU with the cooler unless you swivelled it back and forth to break the bond between the two? That’s still an issue today, something which is no longer an issue with Intel due to their switch to LGA way back in the socket 775 days. I for one would be very happy if AMD switched to LGA but I believe this hasn’t happened as of yet due to backwards compatibility for older motherboards with newer CPUs. Anyway, rant over. Lets take a look at the differences between AMD and Intel CPUs.

As you can see from the image above, the AMD chip is quite a bit larger than the Intel chip. Bigger doesn’t always mean better, though. The heat spreader is not actually much taller (if at all) than, even though it looks like it.

The difference between AMD and Intel is immediately obvious. AMD has pins, Intel does not. Long gone are those days where you could bend a pin on an Intel CPU, unfortunately AMD have yet to follow with this one as mentioned previously.

AMD Test Setup:

CPUs: AMD Piledriver FX-4350, FX-6350 & FX-8350
Motherboard: ASUS Sabertooth 990FX
RAM: G.SKILL RipJawsX 2400MHz CAS 11
Graphics card: BFG GTX295

Intel Test Setup:

CPU: Intel Core i5-3570K & Intel Core i7-3770K
Motherboard: ASUS Sabertooth Z77
RAM: G.SKILL RipJawsX 2400MHz CAS 11
Graphics card: BFG GTX295

Methodology:

All benchmarks will be run on a fresh install of Windows 7 64-bit to ensure that there are minimal background processes taking place to give you a better idea of the true performance behind the processor(s).

Benchmarks:

AIDA64 – CPU & Memory Tests
CINEBENCH 11.5
SiSoftware SANDRA 2013 – CPU & Memory Tests
X264 HD Version 5

I won’t go into detail about the Intel system due to the fact that they are only used in this review to show comparisons, and this review isn’t about the Intel chips. This review is about the AMD chips so I will go into greater depth on those CPUs instead. I’m sure that is what you lot are after anyway… right?

Before I start, I should mention that I used a maximum of 1.5 volts on the CPUs. Note that the LLC (Load Line Calibration) is set to Extreme in the BIOS, hence why the CPU voltages show as 1.524 volts and not 1.5 volts.The reason for this is that it is still considered relatively “safe” and my cooling set up could handle it. I would be cautious advising such a voltage on unsuitable cooling. Only the FX-8350 pushed the Silver Arrow Extreme to its real limits but the voltage had to be kept the same in order to allow for a fair test. The IMC voltage was set to 1.3 volts (the Sabertooth seems to think 1.4 volts is fine!) to keep it happy with the 2400MHz kit of RAM that’s installed.

I’ll start off with the quad core, the FX-4350 and work my way up through the chain. This CPU comes clocked at 4.2GHz at stock which is a high clock speed to begin with. However, the good news is that the Sabertooth puts it at only 1.33 volts for that speed which means it should have a fairly high amount of wiggle room left if we push it up to 1.5 volts. Well, in fact, I managed to push it to 5.1GHz without any hassles at all. It happily completed every single test that I threw at it, even being bombarded with 100% load for over an hour.

Moving on to the hex core, the FX-6350 was next on the torture list. The FX-6350 comes clocked at 3.9GHz at stock which is lower than the quad and octa core variants. Although this may seem like a downside, it isn’t. It still managed to overclock to 5GHz and remain stable with 1.5 volts. It is 100MHz slower than the quad core but this could be down to the silicone lottery rather than it just being that way. Either way, 5GHz is a great result.

Finally, we have the octa core on show and is just as good of an overclocker as the rest. The only thing I appear to be limited by is the voltage and the temperatures it hits. It easily clocked up to 5GHz without any effort at all and by setting the CPU voltage to 1.5 volts. It remained completely stable throughout testing, even if the temperatures were border line too high. However, it remained within thermal limitations (just) so this is happily classed as a pass by me.

So, it seems that these chips can hit 5GHz without little effort and providing that your cooling is able to cope with the heat which they kick out. However, what does this mean in terms of performance when they are stacked against an Intel variant? A 5GHz, eight core CPU is definitely something which is something worth bragging over. Although, the performance is probably more important than those seeking some 5GHz E-peen.

This simple integer benchmark focuses on the branch prediction capabilities and the misprediction penalties of the CPU. It finds the solutions for the classic “Queens problem” on a 10 by 10 sized chessboard. At the same clock speed theoretically the processor with the shorter pipeline and smaller misprediction penalties will attain higher benchmark scores. For example — with HyperThreading disabled — the Intel Northwood core processors get higher scores than the Intel Prescott core based ones due to the 20-step vs 31-step long pipeline. CPU Queen test uses integer MMX, SSE2 and SSSE3 optimizations.

SANDRA 2013 is a pretty stringent benchmark, capable of testing your systems limits.  It is a pretty extensive suite of benchmarks but i have narrowed down the more relevant ones  to compare performance.

x264 HD Benchmark is a benchmark that allows you to measure how fast your PC can encode a 1080p video clip into a high quality x264 video file. It allows for an easy comparison because everyone running it will use the same video clip and software. The x264 video encoder has a fairly accurate internal benchmark (in frames per second) for each pass of the video encode and it also uses multi-core processors very efficiently. All these factors make the x264 HD Benchmark an ideal tool in comparing the video encoding performance of different processors and systems

In order to find out the idle wattage of the processor(s), I booted the system and let it sit in Windows for five minutes without touching it to let it truly settle. The lowest value reported over the space of 10 seconds was then reported. For the load wattages, I loaded the CPU with Cinebench which gives the CPU a quick 100% blasting and then the maximum numbers were recorded and noted down. I opted to not use Prime95 as that seems to make the AMD CPUs chew more power than you’d ever imagine possible. It’s a well-known problem and therefore I avoided it.

Well, well, well… where to begin with this one? Truthfully, I’m not completely sure. So, lets just dive in and battle through it. Starting with the overclocking of these chips…

As seen in the review, these chips all managed to hit 5GHz or even 5.1GHz with air cooling. Granted it is a top-end air cooler, it shows you that it is possible to hit these clock speeds on air. For those of you thinking that a 5GHz chip must be a monster and perform like an unstoppable beast, I would have to caution you in your thinking there when you compare it to a more expensive counterpart from Intel. In some instances which have been witnessed during this review, the eight core from AMD falls short of a CPU with half the cores, without Hyper Threading. It’s not ideal by any means. However, having said that, I do feel that these CPUs do have a place in the market, and that one would be in the area where people cannot afford to splash out on high-end Intel parts but they still want a capable CPU – AMD fit that bill perfectly.

Power consumption is still somewhat of an issue with the AMD CPUs, and it is surprising how much power they actually do draw. It would be fantastic if AMD could get these numbers down but from a personal prospective, I cannot see it happening any time soon as they’ve been stuck on a 125w TDP for years and years now. Nothing seems to suggest (to me) that this would change soon which is a shame. Perhaps Steamroller (next generation) will see these numbers come down. Once you overclock these CPUs, especially the FX-8350, the power consumption shoots through the roof and there’s no stopping it.

Right, lets talk video encoding performance. Now, of course this is only one example as I used X264 but this is a very common program to use to benchmark the video encoding power which is why I opted to use it. I ran the latest version so that it was a full 1080P render (rather than version 4 which is 720P) as that is what a lot of people are rendering today. Video rendering by nature (due to the software) always loved to use more cores and threads which should have made the FX-8350 shine and it should have allowed it to have performed much better than it did. It got beaten by the overclocked i5-3570K in the first pass but it pulled ahead in the second passing, although I suspect the 400MHz difference between the two played a big part in that. Both the quad and hex core chips fell horrifically short of the i5 and even further behind the i7. Even when are running at 5GHz, they just cannot keep up with a stock clocked i5-3570K. However, they did still get the job done, at almost half the price of an i5, it just means that rendering times will be that much longer.

When these CPUs were put through their paces in Cinebench, it showed how much of a performance difference there is between Intel and AMD. Cinebench renders an image using on the CPU cores and it renders a a block at a time per CPU core. IE: Four cores means it will render four parts of the image at a time, six will render six and so on, until the rendering of the image is complete. The faster it completes the render, the higher your score will be. The FX-4350 was bottom of the bunch, even when overclocked to 5.1GHz, it couldn’t compete with a six core FX-6350 at stock. The FX-6350 manages to edge out a win over the i5 at stock speeds when it is clocked to 5GHz but it loses out when the i5 is overclocked to 4.6GHz which coincidentally also beats the FX-8350 at stock clock speeds. The fastest of the AMD bunch is obviously going to be the one with the most cores so the overclocked FX-8350 comes out in second place and it is rivalled by the i7 although it loses out once the overclocked i7 comes into play. Unfortunately for AMD, this shows us exactly how much work they need to put into their next CPUs in order to begin the long catch up with Intel. Something with half the amount of cores manages to maintain a steady and fairly healthy lead.

In order to make sure that this isn’t the most dragged out conclusion in the world, I will carry on to actually summarise the results as best as I can. Judging by the graphs and my testing which was done in this review, it is clear to see how AMD stack up against Intel. However, do not take this as a negative point and keep in mind that the Intel CPUs are vastly more expensive than the AMD offerings.

AMD may not have the best performing chip on the market, but they certainly do provide excellent bang for buck ratio which makes them an ideal choice to go for if you are on a budget but still want a very capable system. I feel that the FX-4350 and FX-6350 definitely deserve our value award. I feel that the FX-8350 is priced a little too steeply, though.

[ratings]