AMD Phenom II X4 965 BE

Back in April, Benchmark Reviews covered the release of AMD's flagship quad-core processor, the Phenom II X4 955 BE. Nearly four months later, we have that opportunity once again. Today, we look at the new king of the mountain over at the AMD, the Phenom II X4 965 Black Edition processor model HDZ965FBGIBOX. Coming in at 3.4GHz, it not only bests the previous flagship model by 200Mhz, but also happens to be the highest clocked CPU on the market. In today's review, we find out just how well that exta clockspeed translates to performance.

Aside from operating frequency, the 3.4GHz Phenom II X4 965 isn't much different than its 3.2GHz counterpart. In fact, our sources at AMD confirm that both chips are binned from the same revision of silicon. That's not necessarily a bad thing, though, as this revision has already proven to be an excellent performer. You can also expect the same architectural benefits, including socket AM2-AM3 compatibility, 8MB total L2 and L3 cache, and a 128-bit DDR2-DDR3 memory controller.

At $245 MSRP, the Phenom II X4 965 BE is also shaping up to be a great value. If you recall, this is actually the same price that the X4 955 BE was introduced at back in April. Thanks to a competitive market, it looks like AMD has had to keep their prices down. We've aslo been told to expect a few bundle deals during launch that should knock the price down even more.

About Advanced Micro Devices, Inc (AMD)

Advanced Micro Devices (NYSE: AMD) is an innovative technology company dedicated to collaborating with customers and partners to ignite the next generation of computing and graphics solutions at work, home, and play.

Over the course of AMD's three decades in business, silicon and software have become the steel and plastic of the worldwide digital economy. Technology companies have become global pacesetters, making technical advances at a prodigious rate - always driving the industry to deliver more and more, faster and faster.

However, "technology for technology's sake" is not the way we do business at AMD. Our history is marked by a commitment to innovation that's truly useful for customers - putting the real needs of people ahead of technical one-upmanship. AMD founder Jerry Sanders has always maintained that "customers should come first, at every stage of a company's activities."

We believe our company history bears that out.

Phenom II X4 965 BE Features

The industry's first true Quad core x86 processor

• True quad-core and triple-core designed from the ground up for better communication between cores.
  • Benefit: Cores can communicate on die rather than on package for better performance

AMD64 with Direct Connect Architecture

• Helps improve system performance and efficiency by directly connecting the processors, the memory controller, and the I/O to the CPU.
• Designed to enable simultaneous 32- and 64-bit computing
• Integrated Memory Controller
  • Benefits: Increases application performance by dramatically reducing memory latency
  • Scales memory bandwidth and performance to match compute needs
  • HyperTransport^TM Technology provides up to 16.0GB/s peak bandwidth per processor-reducing I/O bottlenecks
  • Up to 37GB/s total delivered processor-to-system bandwidth (HyperTransport bus + memory bus)

AMD Balanced Smart Cache

• Shared L3 cache (either 6MB or 4MB)
• 512K L2 cache per core
  • Benefit: Shortened access times to the highly accessed data for better performance.

AMD Wide Floating Point Accelerator

• 128-bit floating point unit (FPU)
• High performance (128bit internal data path) floating point unit per core.
  • Benefit: Larger data paths and quicker floating point calculations for better performance.

HyperTransport^TM Technology

• One 16-bit link at up to 4000MT/s
• Up to 8.0GB/s HyperTransport^TM 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 PC2-8500 (DDR2-1066); PC2-6400 (DDR2-800), PC2-5300 (DDR2-667), PC2-4200 (DDR2-533) or PC2-3200 (DDR2-400) SDRAM unbuffered DIMMs - AM2+
• Support for unregistered DIMMs up to PC2 8500(DDR2-1066MHz) and PC3 10600 (DDR3-1333MHz) - AM3
• Up to 17.1GB/s memory bandwidth for DDR2 and up to 21GB/s memory bandwidth for DDR3
  • Benefit: Quick access to system memory for better performance.

AMD Virtualization^TM (AMD-V^TM) Technology With Rapid Virtualization Indexing

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

AMD PowerNow!^TM Technology (Cool'n'Quiet^TM Technology)

• Enhanced power management features which automatically and instantaneously adjusts performance states and features based on processor performance requirements
• For quieter operation and reduced power requirements
  • Benefit: Enables cooler and quieter platform designs by providing extremely efficient performance and energy usage.

AMD CoolCore^TM Technology

• Reduces processor energy consumption by turning off unused parts of the processor. For example, the memory controller can turn off the write logic when reading from memory, helping reduce system power.
• 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.

Dual Dynamic Power Management^TM

• Enables more granular power management capabilities to reduce processor energy consumption.
• Separate power planes for cores and memory controller, for optimum power consumption and performance, creating more opportunities for power savings within the cores and memory controller.
  • Benefit: Helps improve platform efficiency by providing on demand memory performance while still allowing for decreased system power consumption

HDZ965FBGIBOX Specifications

• Core Frequency: 3.4GHz
• TRAY OPN# HDZ965FBK4DGI
• PIB OPN# HDZ965FBGIBOX
• L1 Cache Sizes: 64K of L1 instruction and 64K of L1 data cache per core (512KB total L1 per processor)
• L2 Cache Sizes: 512KB of L2 data cache per core (2MB total L2 per processor)
• L3 Cache Size: 6MB (shared)
• Total Cache (L2+L3): 8MB
• Memory Controller Type: Integrated 128-bit wide memory controller *
• Memory Controller Speed: Up to 2.0GHz with Dual Dynamic Power Management
• Types of Memory Supported: Unregistered DIMMs up to PC2-8500 (DDR2-1066MHz) -AND- PC3-10600 (DDR3-1333MHz)
• HyperTransport 3.0 Specification: One 16-bit/16-bit link @ up to 4.0GHz full duplex (2.0GHz x2)
• Total Processor-to-System Bandwidth: Up to 37.3GB/s total bandwidth [Up to 21.3 GB/s memory bandwidth (DDR3-1333) + 16.0GB/s (HT3)]
• Up to 33.1GB/s total bandwidth [Up to 17.1 GB/s memory bandwidth (DDR2-1066) + 16.0GB/s (HT3)]
• Packaging: Socket AM3 938-pin organic micro pin grid array (micro-PGA)
• Fab location: GLOBALFOUNDARIES Fab 1 module 1 in Dresden, Germany (formerly AMD Fab 36)
• Process Technology: 45-nanometer DSL SOI (silicon-on-insulator) technology
• Approximate Die Size: 258 mm2
• Approximate Transistor count: ~ 758 million
• Max Temp: 65 Celsius
• Nominal Voltage: 0.825-1.425V
• Max TDP: 140 Watts
• MSRP: $245 (US)

Closer Look: Phenom II X4 965 BE

As I mentioned in the intro, the Phenom II X4 965 BE that we're testing today is based on the same C2 revision of Deneb as the X4 955. So without any changes to the silicon, the X4 965 is simply a higher binned chip capable of running at 3.4 GHz. Fortunately, we have both processors on hand to measure what kind of performance difference to expect. Before we get to the benchmarks, though, let's take a closer look at the chip and recap some of the features of this architecture.

Despite the screenshot above, AMD's Phenom II X4 965 is a socket AM3 processor. That means full support for DDR3 memory, as well as backward compatibility with DDR2 and AM2+ motherboards. As an X4 part, all four cores of the Deneb die are enabled, as well as 512KB L2 cache per core and the full 6MB L3 shared cache. To reach 3.4GHz, though, it looks like AMD had to increase the voltage to 1.4 volts. In fact, a quick glance back at the specifications reveals a TDP of 140 watts. Given some of the undervolting figures I've seen with the X4 955, though, I have a feeling this chip could run on less. Of course, as a Black Edition model, we're more interested in how well it will overclock. We'll cover this in detail toward the end of this article, but here's a little preview:

After some tweaking, we were able to get our Phenom II X4 965 up to 3.9GHz, a full 100 MHz higher than we achieved during our X4 955 review. We also increased the CPU-NB to 2.8GHz, which, as you'll see in the overclocking section, helps quite a bit with memory intensive applications.

Testing and Results

Before I begin any benchmarking or overclocking, I thoroughly stress the CPU and memory by running Prime95 on all available cores for 12 hours. If no errors are found, I move on to a gaming stress test. To do this, I use Prime95 again to stress the processor, while running an instance of FurMark's stability test on top of this. If the computer survives this test for 2 hours without lockup or corruption, I consider it to be stable and ready for overclocking. After achieving what I feel is stable overclock, I run to these tests again for certainty. The goal of this stress testing is to ensure the clock speeds and settings are stable before performing any benchmarks. After all, what good are performance measures if the system cannot reliably produce them.

Once the hardware is prepared, we begin our testing. Each benchmark test program begins after a system restart, and the very first result for every test will be ignored since it often only caches the test. This process proved extremely important in the World in Conflict and Supreme Commander benchmarks, as the first run served to cache maps allowing subsequent tests to perform much better than the first. Each test is completed five times, with the average results displayed in our article.

Intel LGA775 Test System

• Processors: Intel Core 2 Duo E7400, Core 2 Quad Q9450
• Motherboard:Asus P5Q3
• System Memory: 2x2GB OCZ Platinum DDR3 (1333MHz@6-6-6-20 & 1066MHz@6-6-6-15)
• Video:ASUS EAH4870 DK TOP
• Disk Drive: Western Digital 6400AAKS 640GB
  
• Optical Drive: LITE-ON iHAS122-04 DVD Burner
• PSU: Corsair TX850W
• Enclosure:Cooler Master CM Storm
• Operating System: Windows XP Professional SP3

AMD Socket AM3 Test System

• Processors:AMD Phenom II X4 965 BE, Phenom II X2 550, Athlon II X2 250
• Motherboard: Asus M4A79T Deluxe
• System Memory: 2x2GB OCZ Platinum DDR3 (1600MHz@7-7-7-24)
• Video:ASUS EAH4870 DK TOP
• Disk Drive: Western Digital 6400AAKS 640GB
• Optical Drive: LITE-ON iHAS122-04 DVD Burner
• PSU: Corsair TX850W
• Enclosure:Cooler Master CM Storm
• Operating System: Windows XP Professional SP3

Benchmark Applications

• EVEREST Ultimate Edition v5.01 by Lavalys
• Passmark PerformanceTest
• PCMark05 v1.2.0 by Futurmark Corporation
• Crysis Benchmark Tool
• Devil May Cry 4 Benchmark
• SPECviewperf v10.0 by Standard Performance Evaluation Corporation

EVEREST Benchmark Tests

EVEREST Ultimate Edition is an industry leading system diagnostics and benchmarking solution for enthusiasts PC users, based on the award-winning EVEREST Technology. During system optimizations and tweaking it provides essential system and overclock information, advanced hardware monitoring and diagnostics capabilities to check the effects of the applied settings. CPU, FPU and memory benchmarks are available to measure the actual system performance and compare it to previous states or other systems. Furthermore, complete software, operating system and security information makes EVEREST Ultimate Edition a comprehensive system diagnostics tool that offers a total of 100 pages of information about your PC.

As one might expect, there isn't much difference between the Phenom II X4 965 and the 955 when it comes to memory bandwidth. Both have the same memory controller and are using the same DDR3 modules. The higher clock speed of the X4 965 only slightly tips these benchmarks in its favor.

Moving on to the CPU integer tests, things get a little more interesting. Compared to the X4 955, the X4 965's extra 200 MHz equates to roughly a 6% increase in the Queen and AES benchmarks, and a 3.5% increase in the PhotoWorxx benchmark.

The same increase in performance can be seen in the floating point tests. Here, the X4 965 has a clear 6% lead over the 955 across the board. As we've seen in past reviews, however, the Intel chips perform slightly better in the single-precision Julia test and the X4 965 has to settle second place.

Passmark PerformanceTest

PassMark PerformanceTest is a PC hardware benchmark utility that allows a user to quickly assess the performance of their computer and compare it to a number of standard 'baseline' computer systems. The Passmark PerformanceTest CPU tests all benchmark the mathematical operations, compression, encryption, SSE, and 3DNow! instructions of modern processors.

In our tests there were several areas of concentration for each benchmark, which are combined into one compound score. This score is referred to as the CPU Mark, and is a composite of the following tests: Integer Math, Floating Point Math, Find Prime Numbers, SSE/3DNow!, Compression, Encryption, Image Rotation, and String Sorting. For this review, we've also decided to run the memory benchmark, which results in a composite score based on the following tests: small block allocation, cached read, uncached read, write performance, and large block allocation.

From past reviews, we've seen that Intel processors consistently score higher in this benchmark. Once again, that's the case. The extra clock speed of the Phenom II X4 965 allows it to inch closer to the Intel Q9450, but it still can't quite pass it up. The X4 965 does, however, perform about 6.5% faster than the X4 955. What's interesting is that this performance increase also applies to the memory score. It appears that the at least some memory tests used by Passmark were affected by the difference in clock speed.

PCMark05 Benchmark Tests

Using synthetic benchmarks to compare one product to another has some distinct advantages when testing similar hardware, yet I have never found myself completely satisfied by the process. I have come to understand that they're important for comparing "apples to apples", and that the results are usually very consistent. But as with any synthetic benchmark, the numbers can often mean very little more than just numbers. We don't take a high score on a synthetic benchmark to mean that a product will/should perform well, and neither should you. The difference between projected performance and actual performance is the difference between fire and the fire-fly.

PCMark is a series of computer benchmark tools developed by Futuremark. The tools are designed to test the performance of the user's CPU, read/write speeds of RAM and hard drives. We have used these tests to simulate a battery of applications and tasks, which will produce results we can compare to other systems using similar hardware.

In contrast to Passmark's rankings, PCMark05 places the Phenom II X4 965 well ahead of the Intel Q9450. Not surprising, it also pulls ahead of the X4 955 in the overall and CPU benchmarks. What is surprising, however, is that once again there is a significant increase in memory scores. This appears to be more of a reflection of the memory tests used in PCMark05 than pure memory performance though. As we saw in Everest, both the X4 965 and the 955 are about matched in terms of memory bandwidth.

Crysis Gaming

Crysis uses a new graphics engine: the CryENGINE2, which is the successor to Far Cry's CryENGINE. CryENGINE2 is among the first engines to use the Direct3D 10 (DirectX10) framework of Windows Vista, but can also run using DirectX9, both on Vista and Windows XP.

Roy Taylor, Vice President of Content Relations at NVIDIA, has spoken on the subject of the engine's complexity, stating that Crysis has over a million lines of code, 1GB of texture data, and 85,000 shaders. To get the most out of modern multicore processor architectures, CPU intensive subsystems of CryENGINE 2 such as physics, networking and sound, have been re-written to support multi-threading.

Crysis offers an in-game benchmark tool, which is similar to World in Conflict. This short test does place some high amounts of stress on a graphics card, since there are so many landscape features rendered. For benchmarking purposes, Crysis can mean trouble as it places a high demand on both GPU and CPU resources. Benchmark Reviews uses the Crysis Benchmark Tool by Mad Boris to test frame rates in batches, which allows the results of many tests to be averaged.

When testing a CPU, the settings we choose are a bit different than a typical video card review. As you'll see in the charts below, modern games can easily max out the processing abilities of the video card and mask any differences between the CPUs. For this reason, we generally start at low resolutions and quality settings and slowly ramp them up until we hit the limit of the video card. This allows us to quickly distinguish differences between CPUs and identify any other limiting hardware.

At 1024x768 and low quality settings, we see the greatest variation in frame rates. While the Phenom II X4 955 previously tied the Intel Q9450, with the release of the X4 965, we finally have a clear winner. As we increase resolution and quality settings, though, all three of our quad core processors are enough to keep up with the HD 4870.

Devil May Cry 4 Gaming

Devil May Cry 4 is one of the newest additions to Benchmark Reviews' testing suite. Fortunately for us, Capcom recognized the community's interest in hardware testing and included a built in benchmarking tool with this game. In fact, it is even included it in the demo. The benchmarking tool runs through four different scene, all of which I highly recommend watching. However, for the purposes of our review, we only report the scores of the most challenging tests: scenes 2 and 4.

Similar to the Crysis gaming benchmarks, we will start testing DMC 4 at lower settings to reduce the impact of GPU limitations. From there, we'll slowly increase resolution and quality settings until we reach the limit of our HD 4870.

Scene two of our Devil May Cry 4 benchmark is a bit unpredictable. At the lower resolution, we see a nice trend with the Phenom II X4 965 and 955 taking the lead. However, at higher resolutions, the results are mixed and X4 965 falls behind. Let's see if scene four can clear things up.

Scene four is a little more consistent. Here, the X4 965 manages to pull ahead of the pack in the low resolution test. Still, the scores even out at higher resolutions and settings as we reach the limit of our video card.

SPECperfview CATIA Tests

SPECviewperf is a portable OpenGL performance benchmark program written in C. It was developed by IBM. Later updates and significant contributions were made by SGI, Digital (Compaq, HP), 3Dlabs (Creative Labs) and other SPECopc project group members. SPECviewperf provides a vast amount of flexibility in benchmarking OpenGL performance. Currently, the program runs on most implementations of UNIX, Windows XP, Windows 2000, and Linux.

SPECviewperf parses command lines and data files, sets the rendering state, and converts data sets to a format that can be traversed using OpenGL rendering calls. It renders the data set for a pre-specified amount of time or number of frames with animation between frames. Finally, it outputs the results. SPECviewperf reports performance in frames per second. Other information about the system under test - all the rendering states, the time to build display lists (if applicable), and the data set used - are also output in a standardized report.

A "benchmark" using SPECviewperf is really a single invocation of SPECviewperf with command-line options telling the SPECviewperf program which data set to read in, which texture file to use, what OpenGL primitive to use to render the data set, which attributes to apply and how frequently, whether or not to use display lists, and so on. One quickly realizes that there are an infinite number of SPECviewperf "benchmarks" (an infinite number of data sets multiplied by an almost infinite number of command-line states).

As this is a single-threaded benchmark, our quad core and dual core processors are now on a level playing field. The dual cores do put up a decent fight, particularly the Intel E7400. However, the Phenom II X4 965 takes the lead, with a 4.5% increase over the 955.

Phenom II X4 965 BE Overclock Results

Overclocking the Phenom II X4 965's core frequency was a breeze thanks to its unlocked Black Edition multiplier. As we've seen with the rest of the Phenom II family, this processor also benefits from overclocking the cpu-northbridge. Changing this frequency directly impacts the memory controller and L3 cache. To illustrate the point, I've run our benchmarks at four different settings. Up first are the stock results, which I simply copied over from the previous pages for reference. I then overclocked just the northbridge to 2.8 GHz to see what kind of effect it had. Following those benchmarks, I set the northbridge back to stock and overclocked just the core frequency to 3.9 GHz. Finally, I combined the two for a total of 3.9GHz core frequency and 2.8 GHz northbridge frequency. The results are below.

The memory benchmarks clearly demonstrate the importance of the northbridge frequency. As expected, the northbridge-only overclock and the northbridge+CPU overclock perform the best. The CPU-only overclock does show some improvement over stock, but not nearly as much as when combined with an overclocked northbridge.

Moving on to the CPU benchmarks, the results are reversed. The Everest AES and Queen tests, as well as PCMark05's CPU test, do an excellent job of isolating the CPU. In these benchmarks, overclocking the northbridge has a negligible effect. However, the Everest PhotoWorxx benchmark breaks the trend and confirms Everest's claim that this benchmark tests the CPU and memory subsystem. What's interesting is that overclocking the northbridge has a greater effect than overclocking the core frequency. As expected though, the combined overclock takes the lead and even appears to have a synergistic effect.

The gaming benchmarks tell a story similar to what we saw in the earlier sections of the review. At higher resolutions and quality settings, the video card is likely to hit a wall before the processor does. At lower resolutions though, we do see a significant effect from overclocking. Once again, the combined overclock is the performance leader, while the northbridge-only and cpu-only overclocks seem to vary depending on the game.

Catalyst 9.8 Driver Comparison

Shortly before launch, AMD gave us access to an unreleased version of their Catalyst 9.8 driver. According to AMD, the new driver should provide significant improvements in peformance when Crossfire systems are paired an AMD processor. While we didn't have a crossfire build to test with, we did check the new driver against our current Catalyst 9.4 driver used for testing.

I wasn't expecting much of a performance increase given that this driver release is geared toward Crossfire systems. However, at lower resolutions, it does appear to have had an effect. Given more time, I plan to benchmark other games. For the time being, though, we'll have to stick with this little preview. Be sure to check it out for yourself when the Catalyst 9.8 driver is released next week.

AMD Phenom II Final Thoughts

Flagship processors tend not to retain their title for very long and that's exactly the case here. Less than four months after its introduction, the Phenom II X4 955 is being replaced by the X4 965. It's a good move for AMD, though, particularly with the Fall shopping season gearing up and Intel's Lynnfield launch looming around the corner.

It's also good news for consumers and an interesting milestone for the industry: the X4 965 is the first quad-core processor ever to ship at 3.4GHz, making it the highest clocked processor on the market. Combined with the solid performance of AMD's C2 revision of Deneb and the unlocked Black Edition multipliers, the X4 965 is a welcome addition to the Phenom II family.

HDZ965FBGIBOX Conclusion

Throughout our benchmarks today, the 3.4GHz Phenom II X4 965 consistently topped the charts. Compared to the previous 3.2 GHz X4 955 flagship model, the extra 200 MHz led to an average performance increase of roughly 6%, making it the fastest Phenom II to date. It also put up a decent fight against Intel's Q9450, pulling ahead in nearly every benchmark.

The X4 965 earns top marks for construction quality as well. We torture all of our system with a combination of Prime95 and FurMark before running any benchmarks and not once did we notice instability. We've come to expect this kind of quality from modern processors, and the Deneb-based X4 965 certainly does not disappoint.

Based on the same C2 revision of Deneb as the X4 955, the Phenom II X4 965 also carries the same architectural benefits. This includes full AM3-DDR3 support, 8MB combined L2 and L3 cache, and backwards compatibility with AM2+ and DDR2. The only downside seems to the slightly increased TDP: 140 watts versus the 955's 125 watts.

Fortunately, that TDP didn't hinder our overclocking results. Thanks to the unlocked multipliers, we were able to achieve 3.9 GHz core clockspeed with our X4 965 BE, up 100MHz from our X4 955 review. Some thorough benchmarking also illustrated the importance of the cpu-northbridge frequency. After bringing this frequency up to 2.8GHz, we noticed a significant improvement in memory-intensive benchmarks as well as some of our gaming tests.

Given the increase in performance compared to the X4 955, it's refreshing to see that AMD is actually introducing X4 965 at the same price point. You can find the Phenom II X4 965 on newegg.com for $249. Of course, that also means a price drop for X4 955, which may entice some. Fortunately, AMD is planning several bundle deals that promise to bring the price down during launch.

Pros:

+ Highest clocked processor on the market
+ AM3 & DDR3 support
+ Unlocked multiplier
+ 8MB of total cache
+ AMD OverDrive 3.0
+ Excellent price/performance

Cons:

- 140W TDP

Ratings:

• Performance: 9.50
• Construction: 9.25
• Functionality: 9.00
• Overclock: 9.50
• Value: 8.50

Final Score: 9.15 out of 10.

Excellence Achievement: Benchmark Reviews Golden Tachometer Award.

Questions? Comments? Benchmark Reviews really wants your feedback. We invite you to leave your remarks in our Discussion Forum.

Related Articles:

• Biostar TA890GXB-HD mATX AMD Motherboard
• AMD A10-5800K Trinity Desktop Processor
• Thermaltake Jing CPU Cooler CLP0574
• OCZ Vendetta Exposed HDT CPU Cooler
• ASUS M4A88TD-V EVO/USB3 Motherboard
• SandForce SF-2000 Series SSD Overview
• Intel Core i7-3770K Ivy Bridge Processor
• Noctua NH-U12P Heatpipe CPU Cooler
• AMD Phenom-II X4-980 BE Processor
• Thermalright SI-128 CPU Cooler