AMD Phenom-II X6-1090T CPU Review
Whether you buy into the brand or not, AMD keeps the desktop and server processor market competitive and affordable. Intel recently released their 'Gulftown' Core i7-980X Extreme Edition desktop processor, which offers 3.33GHz to six unlocked CPU cores to the tune of $1150. For that kind of money, an entire high-end PC can be built using the AMD platform... which now also offers six processor cores and native SATA 6Gb/s connectivity. Enter the AMD Phenom-II X6-1090T Black Edition processor, model HDT90ZFBK6DGR.
AMD's Turbo CORE technology is now available in 'Thuban' Phenom-II AM3 desktop processors, beginning with the 2.8GHz X6-1055T and 3.2GHz X6-1090T CPUs. Turbo CORE senses when three of the six processor cores are not in use, and automatically boosts the clock speed up to 500MHz. Paired with the AMD 890FX chipset found on ASUS' Crosshair-IV Formula ROG motherboard, the AMD Phenom-II X6-1090T Black Edition CPU can reach 4.0GHz on all six cores with an additional 4.3GHz Turbo CORE. In this article, Benchmark Reviews tests the AMD X6-1090T against the Intel Core i7-920 and i7-890X processors in gaming, computing, and overclocking performance.
The Phenom-II X6-1090T is AMD's fastest and most powerful desktop processor, offering exceptional versatility. AMD has provided six real cores of processing power at 3.2GHz for less than $300. Turbo CORE activates when the CPU finds itself in a condition where three (or more) cores rest idle, which the processor then boosts the active cores up to 3.6GHz (on the 1090T) improving performance on less heavily-threaded workloads. With the launch of their flagship 3.2GHz Phenom-II X6-1090T Black Edition processor with Turbo CORE technology, AMD also announces the 2.8GHz Phenom-II X6-1055T. When either of the Phenom-II X6 CPUs is combined with either the AMD-890GX or new AMD-890FX chipset motherboards and a Radeon HD 5000-series graphics card, this creates the "Leo" platform.
Generally speaking, most users barely tap the potential of their computer system. Applications and video games are usually more affected by clock speed than they are processor cores, and many programs still are single-threaded. But again, this applies to most users. The differences between hardware enthusiasts or power-users and their casual PC user counterparts are acute and to the point that they have little in common with each other. While mainstream users concern themselves with browsing the web and checking email, enthusiasts are constantly looking to push their hardware with powerful overclocking experiments and power users create multiple virtual machine profiles for commercial application.
Although the consumer software industry has been slow to develop applications that can utilize multiple CPU cores and/or threads, there are many commercial programs which handle four or more processor threads very well... and even a few video games. Our benchmarks will determine how important multiple cores are to gaming later in this article, but it goes without question that Virtual-Machine applications receive immediate benefits from the added Hyper-Threading. Some already-popular productivity software titles also utilize multi-threaded processor cores, such as: Adobe Photoshop and Premiere-Pro, AutoDesk Maya and 3DS-Max, Microsoft Excel and Windows Live Movie Maker, Sony Vegas and Acid, and also VirtualDub.
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 X6-1090T Features
Turbo CORE Technology
AMD Turbo CORE technology enables maximum performance for any given task; depending on the application workload. AMD Phenom II X6 processors can power extreme multitasking for productivity across 6 cores or shift power to three cores and boost frequency. AMD Turbo Core technology runs seamlessly in the background within the provided TDP limits and works with your Windows operating system automatically to maximize 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 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 (6MB)
- 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 Technology
- One 16-bit link at up to 4000MT/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 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 (AMD-V) 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! Technology (Cool'n'Quiet 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 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
- 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
HDT90ZFBK6DGR Specifications
- Model Number & Core Frequency: X6-1090T / 3.2GHz (Black Edition)
- Part Number: HDT90ZFBK6DGR
- True Six Core Processing
- AMD Turbo CORE Technology
- L1 Cache: 128KB (64KB Instruction + 64KB Data) x6 (six-core)
- L2 Cache: 512KB x6 (six-core)
- L3 Cache: 6MB Shared
- 45-nanometer SOI (silicon-on-insulator) technology
- HyperTransport 3.0 16-bit/16-bit link at up to 4000MT/s full duplex; or up to 16.0GB/s I/O bandwidth
- Up to 21GB/sec dual channel memory bandwidth
- Support for unregistered DIMMs up to PC2 8500 (DDR2-1066MHz) and PC3 10600
- Direct Connect Architecture
- AMD Balanced Smart Cache
- AMD Dedicated Multi-cache
- AMD Virtualization Technology (AMD-V)
- AMD PowerNow 3.0 Technology
- AMD Dynamic Power Management
- Multi-Point Thermal Control
- AMD CoolCore Technology
Closer Look: AMD X6-1090T
The year was 1998, and I purchased my first personal computer. It was a Packard Bell Multimedia-820 desktop system that featured a 233MHz Cyrix MII-300 processor, 32MB of EDO RAM, and a 3.2GB hard drive. For under $1000 the package deal included the system, CRT monitor, bubble jet printer, attached stereo speakers, mouse pad, and even a floppy disk organizer. Powered by the now-legendary Windows 98 Operating System, I was surfing the web with my 56K modem at speeds that would make people cry real tears by today's broadband standard. What made this system so memorable? Besides being my first computer, it was also the first time I would ever overclock a computer... taking that Cyrix MII CPU from 233MHz up to 266MHz with the move of a single jumper shunt. It was madness, I tell you.
That's what overclocking used to be like, way back in the day. You would buy one processor, and after reading instruction on some poorly designed GeoCities/Angelfire website you could turn up the speed and laugh at the suckers paying full retail. Somehow the industry lost its way, and profits prevailed over processor flexibility. Intel was the first to lock-down its processors, and soon AMD followed their lead. This would eventually spawn Intel's Extreme Edition processor series, and AMD's Black Edition counterpart. Like past Black Edition processors, AMD aims to rekindle the lost love for overclocking by offering an 'unlocked' clock multiplier at no additional charge. What's different this time around is that the overclocker can tweak six CPU cores, and add Turbo CORE to the formula.
In present-day 2010, you can still build an entire computer system for less than $1000... and it can even include a six-core 3.2GHz AMD Phenom-II X6-1090T processor. Times have certainly changed, and the technology has changed with it, but still some things remain the same. Let's take a closer look at the AMD X6-1090T model HDT90ZFBK6DGR retail kit:
The new 45nm Thuban-generation Phenom-II processors arrive in two flavors: 2.8GHz X6-1055T, and 3.2GHz X6-1090T. The lower-speed X6-1055T offers more headroom for Turbo CORE, which raises half the processor cores an additional 500MHz to reach 3.3GHz. The flagship AMD Phenom-II X6-1090T Black Edition CPU adds 400MHz for a peak 3.6GHz Turbo CORE speed. Of course, me being me, I overclocked the X6-1090T until it tapped out at 4.0GHz with a 4.3GHz Turbo CORE.
The new hexa-core architecture essentially extends the AMD Phenom-II X6 processor die to accommodate two additional CPU cores. Unlike Intel's 980X, the AMD Phenom-II X6 series does not scale the L3 cache memory amounts with the number of cores. Thuban CPU cache sizes remain identical to those of the AMD Phenom-II X4-965 processor. L1 cache totals 128KB per core (64KB Instruction + 64KB Data), while L2 cache is 512KB per core and L3 receives a shared 6MB. The AMD Phenom-II X6-series 45nm die size measure 346 mm2, packaged in an AM3 938-pin organic micro Pin Grid Array (micro-PGA). Like their latest X4-965 processor, all AMD X6 CPU's maintain a 125W TDP.
AMD/Global Foundries' 45-nanometer SOI (silicon-on-insulator) technology enables Thuban to benefit from a 4000MT/s full-duplex HyperTransport link, with one 16-bit/16-bit link at up to 6.0 GB/s I/O bandwidth. An integrated memory controller delivers up to 21 GB/sec dual-channel memory bandwidth, and accepts unregistered DIMMs up to PC2-8500 (DDR2-1066MHz) or PC3-10600 (DDR3-1333MHz). For each of the test systems used in this article, all memory modules were configured at 1333MHz CL7-7-7-20.
While the Thuban die size measure 346 mm2, the outer package is identical to other Phenom-II CPU profiles. This means that our heatsink guidance provided in the Best CPU Cooler Performance series still applies. Combine the AMD Phenom-II X6 CPU to any AMD-890 motherboard and add any ATI Radeon 5000-series video card, and voilà, you've just created what AMD describes as their 'Leo' platform.
Benchmark Reviews has already tested the ATI Radeon 5000-series ad nauseam, we'll offer several AMD-890FX motherboard reviews following this article. Some of the key features that make the AMD-890FX platform so valuable are the vast supply of PCI-Express 2.0 link lanes; which now total 44 (42 total lanes from 11-engines on the Northbridge for graphics, USB, SATA, etc, and two additional lanes supplied from one engine on SB850 Southbridge.
CPU Testing Methodology
It is the purpose of this review article to illustrate the differences in processor performance. To accomplish this, our tests utilize several benchmark tools to demonstrate synthesized real-world tasks. These tasks include measuring PC video game frame rate performance, multimedia encoding, professional wireframe engineering, and graphical concept creation.
Benchmark results are only as good as their test methodology, and there are several different methods to compare products to one-another. Our goal was to isolate each processor as much as possible by using identical hardware components for each computer system. Identical video card, memory, solid state drive, power supply, and monitor were used in every test. All comparisons use system memory configured to operate at 1333MHz with CL 7-7-7-20-1 timings, however the Intel X58-Express accommodates 6GB of system memory while the AMD 890FX platform accepts 4GB.
Since all of the benchmarks we use for testing represent different game engine technology and graphic rendering processes, our battery of tests will provide a diverse range of results for you to gauge performance on your own computer system. Although many gamers and enthusiasts are still using Windows XP, Benchmark Reviews has made the switch to Microsoft Windows 7 as the Operating System of choice for our test platforms. Using a master drive image, each test platform was built with a clone using the exact same O/S and application load-out.
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. Each test is completed five times, with the highest and lowest scored removed. The average results of the three remaining tests are displayed in our article.
Intel X58 Test Platform
AMD 890FX Test Platform
Benchmark Software
- Operating System: Windows-7 Ultimate Edition 64-Bit
- Lavalys EVEREST Ultimate Edition v5.50
- Futuremark PCMark Vantage v1.0.2 64-Bit
- TV and Movies
- Gaming
- Music
- Maxon CINEBENCH R11.5 64-Bit
- BattleField: Bad Company 2 (High Quality, HBAO, 8x Anti-Aliasing, 16x Anisotropic Filtering, Single-Player Intro Scene)
- SPECviewperf-10:
- UGS Teamcenter Visualization Mockup
- PTC Pro/ENGINEER 2001
- Dassault Systemes CATIA V5R12
- 3ds Max 3.1
- Dassault Systemes Solidworks 2004
- Alias Maya 6.5
- Unigine Heaven Benchmark 2.0 (DX11, Normal Tessellation, 16x AF, 4x AA)
Cinebench R11.5 Benchmarks
Maxon Cinebench is a real-world test suite that assesses the computer's performance capabilities. Cinebench is based on Maxon's award-winning animation software, Cinema 4D, which is used extensively by studios and production houses worldwide for 3D content creation. Maxon software has been used in blockbuster movies such as Spider-Man, Star Wars, The Chronicles of Narnia and many more. Cinebench Release 11.5 includes the ability to more accurately test the industry's latest hardware, including systems with up to 64 processor threads and the testing environment better reflects the expectations of today's production demands. A more streamlined interface makes testing systems and reading results incredibly straightforward.
The Cinebench R11.5 test scenario uses all of a system's processing power to render a photorealistic 3D scene, "No Keyframes" the viral animation by AixSponza. This scene makes use of various algorithms to stress all available processor cores. The OpenGL graphics card testing procedure uses a complex 3D scene depicting a car chase with which the performance of your graphics card in OpenGL mode is measured. During the benchmark tests the graphics card is evaluated by way of displaying an intricate scene that includes complex geometry, high-resolution textures, and a variety of effects to evaluate the performance across a variety of real-world scenarios.
The OpenGL tests in Cinebench force processor-first graphics computation before offloading work to the available video card. Cinebench R11.5 reveals performance results illustrated in the chart below:
Cinebench R11.5 Processor Benchmark Results
Single-core CPU performance offers a per-core baseline reference in Cinebench R11.5, and the speed differences between processors in our test group illustrate the separation. These differences are amplified as the CPU multi-core tests are conducted, producing 3.87 points for the 2.66GHz Intel Core i7-920 processor compared to 3.03 points for the 3.4GHz AMD Phenom-II X4-965. Intel's 3.33GHz Core i7-980X Extreme Edition CPU leads the group with a combined score of 8.28, while the stock and overclocked AMD Phenom-II X6-1090T produce 4.69 and 5.87 respectively.
In the Cinebench OpenGL tests, a Radeon 5870 is paired to the 2.66GHz Intel Core i7-920 to produce 56.34 FPS while the 3.4GHz AMD Phenom-II X4-965 generates 58.88 FPS. Intel's Core i7-980X Extreme Edition processor produces the highest score of 67.06 FPS, but an overclocked AMD Phenom-II X6-1090T nearly matches performance with 66.75 FPS.
Looking at performance from a cost-value standpoint, there's a lot more to be learned. Presuming a $300 MSRP, the AMD Phenom-II X6-1090T costs $63.97 per point or multi-core performance and $5.06 per frame of OpenGL performance. Compare this to the $72.35 per point and $4.97 per OpenGL frame for the Intel Core i7-920. The AMD X4-965 costs $61.39 per point, and $3.16 per OpenGL frame. Finally, the Intel Core i7-980X costs $138.89 per multi-core point and $17.14 per OpenGL frame. By this measure, the AMD Phenom-II X6-1090T and X4-965 both offer better 'bang for the buck'.
SPECviewperf 10 Tests
SPECviewperf 10 is a synthetic benchmark designed to be a predictor of application performance and a measure of graphics subsystem performance. SPECviewperf 10 provides the ability to compare performance of systems running in higher-quality graphics modes that use full-scene anti-aliasing, and measures how effectively graphics subsystems scale when running multithreaded graphics content. The SPECopc project group's SPECviewperf 10 is a performance evaluation software requiring OpenGL 1.5 and a minimum of 1GB of system memory. It currently supports 32/64-bit versions of the Microsoft Windows Operating System.
UGS Teamcenter Visualization Mockup
The tcvis-01 viewset is based on traces of the UGS Teamcenter Visualization Mockup application (also known as VisMockup) used for visual simulation. State changes such as those executed by the application - including matrix, material, light and line-stipple changes - are included throughout the rendering of the model. All state changes are derived from a trace of the running application. The state changes put considerably more stress on graphics subsystems than the simple geometry dumps found in older viewsets. This viewset uses the glDrawArrays primitive to pass data through the OpenGL API.
PTC Pro/ENGINEER 2001
The proe-04 viewset was created from traces of the graphics workload generated by the Pro/ENGINEER 2001 application from PTC. Mirroring the application, draw arrays are used for the shaded tests and immediate mode is used for the wireframe. The gradient background used by the Pro/E application is also included to better model the application workload.
Two models and three rendering modes are measured during the test. PTC contributed the models to SPEC for use in measurement of the Pro/ENGINEER application. The first of the models, the PTC World Car, represents a large-model workload composed of 3.9 to 5.9 million vertices. This model is measured in shaded, hidden-line removal, and wireframe modes. The wireframe workloads are measured both in normal and antialiased mode. The second model is a copier. It is a medium-sized model made up of 485,000 to 1.6 million vertices. Shaded and hidden-line-removal modes were measured for this model.
VisMockup produced some interesting results; primarily because the AMD Phenom-II X4-965 topped all of the charts. Since each processor was tested a minimum of three times each, we knew our results were solid. What I don't know is why Teamcenter Visualization favors the X4-965 more than it favors an overclocked X6-1090T. It's also interesting to see that the X6-1090T performs the same at 3.2GHz as it does with a 4.0GHz overclock and 4.3GHz Turbo CORE. Forgetting all of this, the important result is that AMD's X6-1090T outperforms the much more expensive i7-980X using UGS Teamcenter Visualization Mockup.
PTC Pro/ENGINEER 2001 offered very similar results, where the AMD X4-965 is substantially ahead of the others and the AMD Phenom-II X6-1090T has considerably outperformed the Intel i7-980X CPU. Since Pro/ENGINEER offered excellent separation between products, we'll illustrate another price-performance breakdown. Beginning with the Intel Core i7-920, Pro/ENGINEER costs $16.56 per point of performance. Because of the low price and high benchmark result, AMD's X4-965 costs only $7.33 per point. The opposite is true for the expensive Intel i7-980X, which costs $66.90 per point of Pro/ENGINEER performance. Finally, the AMD Phenom-II X6-1090T costs 13.27 per point of performance.
Dassault Systemes CATIA V5R12
The catia-02 viewset was created from traces of the graphics workload generated by the CATIAV5R12 application from Dassault Systemes. Three models are measured using various modes in CATIA. Phil Harris of LionHeart Solutions, developer of CATBench2003, supplied SPEC/GPC with the models used to measure the CATIA application. The models are courtesy of CATBench2003 and CATIA Community. The car model contains more than two million points. SPECviewperf replicates the geometry represented by the smaller engine block and submarine models to increase complexity and decrease frame rates. After replication, these models contain 1.2 million vertices (engine block) and 1.8 million vertices (submarine).
3ds Max 3.1
The 3dsmax-04 viewset was created from traces of the graphics workload generated by 3ds max 3.1. To ensure a common comparison point, the OpenGL plug-in driver from Discreet was used during tracing. The models for this viewset came from the SPECapc 3ds max 3.1 benchmark. Each model was measured with two different lighting models to reflect a range of potential 3ds max users. The high-complexity model uses five to seven positional lights as defined by the SPECapc benchmark and reflects how a high-end user would work with 3ds max. The medium-complexity lighting models uses two positional lights, a more common lighting environment.
CATIA V5R12 appears to favor the AMD processors over Intel, judging by our results. This seems pecular, since the Intel systems were given 6GB of triple-channel 1333MHz DDR3, and the AMD systems received only 4GB of the same system memory. Unfortunately, SPECviewperf doesn't elaborate on what variables the tests deems most critical.
3ds Max 3.1 differs from the other tests, primarily because Intel's i7-920 and i7-980X both appear to respond in similar fashion to the AMD X4-965. Unfortunately, the AMD Phenom-II X6-1090T doesn't impress us with the lowest 3ds Max benchmark score of the group.
Dassault Systemes Solidworks 2004
The sw-02 viewset was created from traces of the graphics workload generated by the Solidworks 2004 application from Dassault Systemes. The model and workloads used were contributed by Solidworks as part of the SPECapc for SolidWorks 2004 benchmark. State changes as made by the application are included throughout the rendering of the model, including matrix, material, light and line-stipple changes. All state changes are derived from a trace of the running application. The state changes put considerably more stress on graphics subsystems than the simple geometry dumps found in older viewsets.
Alias Maya 6.5
The maya-02 viewset was created from traces of the graphics workload generated by the Maya 6.5 application from Alias. The models used in the tests were contributed by artists at NVIDIA. Various modes in the Maya application are measured. Models used in this version of the Maya viewset contain many more vertices than those used in maya-01, better reflecting models used by animators in the real world.
The Solidworks 2004 tests appears to mirror the UGS Teamcenter Visualization Mockup results, and give the AMD Phenom-II X4-965 a heady lead over the others. Surprisingly, the Intel Core i7-920 and i7-980X both produce nearly identical results as the AMD Phenom-II X6-1090T.
Quite possibly the opposite of PTC Pro/ENGINEER 2001 and Dassault Systemes CATIA V5R12, Alias Maya 6.5 test results appear pro-Intel. The Intel Core i7-920 and i7-980X both dominate their AMD counterparts, which illustrates how important it is to match your software needs to the best hardware for the task.
Unlike Cinebench, which relied primarily on CPU clock speed for results, the six real-world application benchmarks in SPECviewperf offer a more specific determination of performance. Five of the six programs benefited most by the AMD Phenom-II X4-965, with the Phenom-II X6-1090T trailing right behind, and only one test (Maya) craved Intel processors more than any other.
Video Game Benchmarks
PC-based video games can depend heavily on the CPU if the attached GPU (Graphics Processing Unit) is less powerful, or the graphics settings are configured so low that they create no strain on the video card and rely purely system processing speed; a phenomenon known as CPU-dependence. The opposite is true when the video game has a powerful video card installed, and can handle all graphical demands without receiving assistance from the CPU. Considering that these processors cost $300 or more, it's reasonable to expect enthusiasts can also afford a high-end GPU. After concluding that gamers would never play at the CPU-bound 640x480 resolution, especially considering the most popular modern resolutions are 1280x1024 and 1680x1050, our gaming benchmarks use the ATI Radeon HD 5870 video card to best represent real-world usage.
It's been my personal pleasure to educate PC hardware enthusiasts, and show them the truth behind gaming performance. Nearly always, video game FPS performance does not benefit from premium system memory or expensive processors. Still, the marketing hype seems persuade gamers into thinking it matters. For example, Intel markets their Core i7-980X Extreme Edition as a component capable of boosting video game performance, and goes so far as to claim it is the "Ultimate Gaming Weapon" that will help you "Experience smoother and more realistic gaming made possible by distributing AI, physics, and rendering across six cores and 12 threads, bringing 3D gaming to life for the ultimate desktop experience". This is a bold claim, and one that I've seen before. Fortunately video cards have come a very long way from the days when they couldn't display maximum game settings without begging for help from the CPU. So how much of a difference do processors make in PC gaming?
Since many of our previous articles have demonstrated how processor clock speed is more relevant to FPS performance than the number of CPU cores, and that the GPU handles 99% of the workload, we'll jump right into proving how our test group differs. Starting with BattleField: Bad Company 2, which was tested because it's one of the few games that can handle multi-threaded processor cores, we see that the ATI Radeon HD 5870 video card can produce frame rates in the 70's at the 1920x1200 resolution (24-28" LCD monitors).
BattleField: Bad Company 2 proves that only 3 FPS stand between our group of high-end enthusiast processors. I've already pointed out that the graphics hardware does nearly all of the work in PC video games, so long as it's capable, and anyone able to afford one of these premium processors should have an equally-expensive video card for the best performance. The Intel Core i7-920 produces 73.9 FPS, but the more affordable AMD Phenom-II X4-965 delivers 75.6 FPS. Proving the clock speed precedes the number of CPU cores, the Intel Core i7-980X delivers 74.6 FPS and the AMD Phenom-II X6-1090T produces 72.3; both of which are lower frame rates than the 3.4GHz X4-965.
Switching to the extremely demanding Unigine Heaven (2.0) benchmark, frame rate performance may actually call upon the CPU. Using DirectX-11 settings with normal tessellation and 4x AA/16x AF, we discovered only 0.5 FPS difference between processors. This miniscule disparity further proves that gamers are better off spending their money on a better graphics card than a higher-end CPU.
As far as our gaming tests are concerned, the added CPU cores neglected to add any additional value to the 3D graphics experience, and should not be purchased solely for the purpose of improving frame rate performance. It could be argued that our tests haven't allowed the CPU to take control, but that same position would go against the central purpose of enthusiast hardware; you're not going to find a gamer with a premium processor using an undersized monitor and an entry-level video card.
Despite our best intentions, it seems that these physics-heavy multi-threaded graphics benchmarks weren't particular to which processor they we using and performed well with any of them. Based on these demonstrations, it's time for the community to realize that no processor can match the gaming performance of any modern graphics card.
PCMark Vantage Test Results
PCMark Vantage is an objective hardware performance benchmark tool for PCs running 32- and 64-bit versions of Microsoft Windows Vista or Windows 7. PCMark Vantage is well suited for benchmarking any type of Microsoft Windows Vista/7 PC: from multimedia home entertainment systems and laptops, to dedicated workstations and high-end gaming rigs. Benchmark Reviews has decided to use a few select tests from the suite to demonstrate simulate real-world processor usage in this article. Our tests were conducted on 64-bit Windows 7, with results displayed in the chart below.
TV and Movies Suite
- TV and Movies 1 (CPU=50%, RAM=2%, GPU=45%, SSD=3%)
- Two simultaneous threads
- Video transcoding: HD DVD to media server archive
- Video playback: HD DVD w/ additional lower bitrate HD content from HDD, as downloaded from net
- TV and Movies 2 (CPU=50%, RAM=2%, GPU=45%, SSD=3%)
- Two simultaneous threads
- Video transcoding: HD DVD to media server archive
- Video playback, HD MPEG-2: 19.39 Mbps terrestrial HDTV playback
- TV and Movies 3 (SSD=100%)
- TV and Movies 4 (CPU=50%, RAM=2%, GPU=45%, SSD=3%)
- Video transcoding: media server archive to portable device
- Video playback, HD MPEG-2: 48 Mbps Blu-ray playback
Gaming Suite*
- Gaming 1 (CPU=30%, GPU=70%)
- Gaming 2 (SSD=100%)
- Gaming 3 (CPU=75%, RAM=5%, SSD=20%)
- Two simultaneous threads
- CPU game test
- Data decompression: level loading
- Gaming 4 (CPU=42%, RAM=1%, GPU=24%, SSD=33%)
- Three simultaneous threads
- GPU game test
- CPU game test
- HDD: game HDD
Music Suite
- Music 1 (CPU=50%, RAM=3%, GPU=13%, SSD=34%)
- Three simultaneous threads
- Web page rendering - w/ music shop content
- Audio transcoding: WAV -> WMA lossless
- HDD: Adding music to Windows Media Player
- Music 2 (CPU=100%)
- Audio transcoding: WAV -> WMA lossless
- Music 3 (CPU=100%)
- Audio transcoding: MP3 -> WMA
- Music 4 (CPU=50%, SSD=50%)
- Two simultaneous threads
- Audio transcoding: WMA -> WMA
- HDD: Adding music to Windows Media Player
* EDITOR'S NOTE: Hopefully our readers will carefully consider how relative PCMark Vantage is as "real-world" benchmark, since many of the tests rely on unrelated hardware components. For example, per the FutureMark PCMark Vantage White Paper document, Gaming test #2 weighs the storage device for 100% of the test score. In fact, according to PCMark Vantage the video card only impacts 23% of the total gaming score, but the CPU represents 37% of the final score. As our tests in this article (and many others) has already proven, gaming performance has a lot more to do with the GPU than the CPU, and especially more than the hard drive or SSD (which is worth 38% of the final gaming performance score).
Beginning with the results from PCMark Vantage's TV and Movies benchmark, the Intel Core i7-980X leads the pack with a score of 6118, followed by the overclocked AMD Phenom-II X6-1090T with 6043. The stock speed AMD Phenom-II X6-1090T produces 5828, which is just slightly more than the 5740 points scored by the AMD X4-965. The Intel Core i7-920 produces 4985, and comes in last place for PCMark Vantage TV and Movies performance.
Taking my Editor's Note into consideration (see above), PCMark Vantage Gaming benchmark gives the test to Intel's Gulftown... and by a huge margin. It's too bad (for Intel) that real-world video games don't benefit nearly this much by the processor, or else the Video Game Benchmarks section of the Intel Core i7-980X review would have looked very different. It's a good thing that the PCMark Vantage Gaming benchmark can utilize up to sixteen processor cores... :)
Moving on to the Vantage Music tests, which weight the CPU for 75% of the score, we see just how well the added cores improve the transcoding process. While the Intel Core i7-920 delivers 8602 points and the Gulftown Core i7-980X improves upon it with 12665, the 3.4GHz AMD Phenom-II X4-965 tops them both with 12886. Even more impressive for AMD's Phenom-II series is the results for their X6-1090T Black Edition CPU, which produced 16292 points at stock, and 18227 overclocked to 4.0GHz with a 4.3GHz Turbo CORE. Clearly, this is the ideal route for enthusiasts who encode their own video and transcode various audio formats.
While I'm not in agreement with the weighted importance of certain hardware components in the gaming tests, nor do I think that 1024x768 resolution without anti-aliasing or anisotropic filtering effects is representative to the "consumer experience" as Futuremark puts it, I will agree that the TV & Movies and Music test results are fairly representative of the real-world experience. In both of these tests, the AMD Phenom-II X6-1090T Black Edition CPU was favored heavily over the others, especially the more expensive Intel Core i7-980X.
Everest CPU Benchmarks
Lavalys 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.
All of the benchmarks used in our test bed: Queen, Photoworxx, ZLib, and AES, all rely on basic x86 instructions, and consume very low system memory while also being aware of HyperThreading, multi-processors, and multi-core processors. While the EVEREST CPU tests really only compare the processor performance more than it measures platforms, it still offers a glimpse into what kind of power each platform possesses.
Queen and Photoworxx tests are synthetic benchmarks that operate the function many times and over-exaggerate by several magnitudes what the real-world performance would be like. The Queen benchmark focuses on the branch prediction capabilities and misprediction penalties of the CPU. It does this by finding possible solutions to the classic queen problem on a chessboard. At the same clock speed theoretically the processor with the shorter pipeline and smaller misprediction penalties will attain higher benchmark scores.
Since the benchmark tests in Everest aggressively exaggerate CPU performance, we also expect that the final scores are equally exaggerated as well. Showing a real edge over quad-core architecture in Queen tests, Intel's Gulftown pushes past the i7-920 and leads the pack using its raw processing power. Even the overclocked 4.0GHz AMD Phenom-II X6-1090T can't produce Queen CPU performance close enough to match the Intel 980X.
Like the Queen benchmark, the Photoworxx tests for penalties against pipeline architecture. The synthetic Photoworxx benchmark stresses the integer arithmetic and multiplication execution units of the CPU and also the memory subsystem. Due to the fact that this test performs high memory read/write traffic, it cannot effectively scale in situations where more than two processing threads are used. The EVEREST Photoworxx benchmark performs the following tasks on a very large RGB image:
- Fill
- Flip
- Rotate90R (rotate 90 degrees CW)
- Rotate90L (rotate 90 degrees CCW)
- Random (fill the image with random colored pixels)
- RGB2BW (color to black & white conversion)
- Difference
- Crop
If CPU PhotoWorxx in any indicator of pipeline architecture penalties, then the AMD Phenom-II X6-1090T is in trouble. Scoring significantly lower than the AMD X4-965, the X6-1090T can't even compete with the other processors in terms of amplified PhotoWorxx performance.
The Zip Library test measures combined CPU and memory subsystem performance through the public ZLib compression library. ZLib is designed as a free lossless data compression library for use on virtually any computer hardware and operating system. The ZLib data format is itself portable across platforms and has a footprint independent of input data that can be reduced at some cost in compression.
Since I'm also the webmaster for Benchmark Reviews, I find myself using file compression programs almost as much as I use web browsers. This makes the CPU Zip library tests very important to me. Compared to Intel's Core i7-980X processor, the AMD Phenom-II X6-1090T hexa-core CPU offers decent Zip-archive performance in stock form, and it becomes more comparable once overclocked.
The AES integer benchmark measures CPU performance using AES data encryption. It utilizes Vincent Rijmen, Antoon Bosselaers and Paulo Barreto's public domain C code in ECB mode and consumes 48 MB of memory.
Encryption is a technology that has come to the forefront of personal computing, and will become more common with everyday computing as the world becomes more competitive. Generally seen in the corporate enterprise segment, file encryption is steadily becoming a topic of discussion for personal computer users. So when the Intel Core i7-980X arrives and dominates the AES Encryption test by more than 1025%, it's time to sit up and take notice. Simply put, Intel's Gulftown architecture delivers enterprise-level performance to encryption tasks.
AMD Black Edition CPU Overclocking
When it comes to AMD processors, 'Black Edition' literally means overclocker-friendly. Benchmark Reviews demonstrated this in our overclocking tests with the AMD Phenom-II X4-965, which reached 4.2GHz. But that was then, two whole CPU cores ago. Six-core overclocking is all the craze these days, right?
Using the ASUS Crosshair-IV Formula ROG motherboard, overclocking was a breeze. I'm certain that AMD would have preferred we use their AMD OverDrive overclocking software for this article, but most enthusiasts prefer to go hands-on through the BIOS. It also doesn't help that we received the AMD Phenom-II X6-1090T only a few days before this article was published for product launch. So by using the ASUS Crosshair-IV Formula motherboard, it took changing only two BIOS settings to reach 4.0GHz with a 4.3GHz Turbo CORE overclock.
Benchmark Reviews will go into overclocking in more detail for our review of this AMD 890FX motherboard, but for early adopters who want a head start, I will share my secrets. AMD technicians shipped us the ASUS Crosshair-IV Formula with the AMD Phenom-II X6-1090T Black Edition processor already installed with a tweaked BIOS. Knowing that our tests needed to use 'vanilla' settings, the first step was resetting the BIOS configuration to the motherboard defaults. Next, go directly to the Extreme Tweaker menu, and set the AI Overclock Tuner to manual so you can manipulate the settings.
The first setting was simple: change the CPU Ratio (clock multiplier) to 20.0, which results in a 4.0GHz processor clock speed (20.0 x 200MHz = 4000MHz/4GHz). The second setting wasn't any more difficult, but requires CPB Control (Core Performance Boost) to be enabled. Once enabled I set the CPB CPU Ratio to 21.5, which generates a 4.3GHz Turbo CORE speed (21.5 x 200MHz = 4300MHz/4.3GHz). That's all there was to it!
But wait, there were a few pitfalls. For example, the version 0505 BIOS for the ASUS Crosshair-IV Formula motherboard allowed me to create any CPB CPU Ratio I wanted; even if that ratio was lower than the CPU Ratio. The one time I experimented with this, the system became unstable and Windows crashed. Hopefully ASUS will engineer future BIOS firmware with code that prevents using a lower CPB CPU Ratio than what the CPU Ratio is set to... or at least display a warning.
Voltages were managed by the motherboard, and each component was used the [AUTO] setting in the BIOS. In past experiments, setting the voltages to AUTO didn't always benefit the project. To my surprise, the ASUS Crosshair-IV Formula ROG motherboard took control and maintained complete stability throughout all of my benchmark testing.
I know you're still probably asking: why didn't he tweak the HyperTransport settings? The reason was just as simple as my overclocking changes: because this is a processor review. If I could change just two settings and create a 4.0/4.3GHz overclock, just imaging what the patient PC hardware enthusiast could do with a lot more time to test go/no-go settings! What makes this adventure even more enticing is that AMD mentioned that the HyperTransport 3.0 link can hit 2.6GHz on Thuben, which means there's plenty more headroom for added performance outside of the basic clock speed.
Thuban/Leo Final Thoughts
Members of the computer hardware industry have been forced to watch a real-life episode of Keeping up with the Joneses repeat itself over and over for the past several decades. More recently this behavior has become central to the standard operation procedure, and now companies use it to plan their future product strategies. For example, when Intel releases a new CPU, in no time at all AMD produces a similar product at a lower price point. In response, Intel then launches a bigger, better, faster product - to which AMD answers by repeating the process. The same scenario is true for ATI, AMD's graphics unit. If NVIDIA produces a new GPU, AMD comes back with their own at a better price point. It's become such a predictable show, that by this point every manufacturer has shared in the act of conspicuous production.
AMD has become the master of 'keeping up', and Thuban is a perfect example of their desire to deliver Intel-level products with AMD-level pricing. When Intel designed Gulftown they didn't just add two cores and call it done. The 32nm Westmere processor fabrication allowed two additional cores to be added onto a shrunken Nehalem design to form the foundation for Gulftown, which then added additional shared L3 cache to the unit for proper scaling. Thuban simply adds two more cores, and leaves everything else virtually identical to the way it was on their Phenom-II X4 series. This might sound like I'm talking down AMD's six-core Thuban series, but in reality they've just managed to create an Intel Core i7-980X Extreme Edition counterpart for only 30% of the price to consumers.
Of course, AMD didn't end Thuban with the inclusion of two new CPU cores. New to the Phenom-II series is Turbo CORE, AMD's answer to Intel Turbo Boost. AMD Turbo CORE is a hardware-based C-state performance boost that raises the clock speed of up to three processor cores when they are in a resting idle state. In theory this might sound exactly like Intel's Turbo Boost technology, but in reality they're each very different in execution. I'm summarizing here, but where Intel's TB tech can scale core clock speeds and shut down cores as needed, AMD raises the speed of three cores in unison. So what's next, HyperThreading? Probably not, since AMD has long contended that HT is counter-productive processor technology.
On an similar note, AMD has mastered is the art of consumer-friendly platform branding. The AMD Dragon platform really wasn't anything more than a press-friendly way of saying that their 790-series motherboards worked well with ATI Radeon HD 4000-series video cards and Phenom-II processors... all of which were released at approximately the same time. Combined into one synergistic 'Dragon' platform they supported the capabilities of each other, which is to say they were designed to work with each other. That sounds silly, but when you consider that Intel's been busy building incompatible platforms (P55 and X58) for the same consumer group for almost three years now, it begins to sound like a novel approach. The same concept is true for the AMD Leo platform, which combines six-core AMD Phenom-II Thuban processors with their 890-series motherboard and ATI Radeon HD 5000-series video card. The concept is so simple, it just might work.
Forward Thinking: AMD Processors
If you're sold on Intel processor's and would never stray to another brand, it still serves your best interests to cheer on AMD to produce better products. Imagine a world where only one company makes the product you depend on. Think deeply enough on the topic, and you'll realize that a strong market dominance, especially a monopoly, never serves to benefit the consumer. Value-pricing AMD Athlon/Phenom desktop processors is a winning strategy, and as consumers realize there's very little real-world difference the gap narrows, but at the end of the day PC hardware enthusiast are shrinking in number. If AMD is going to continue competing within the same market space as Intel, they'll need to expand beyond consumer-level desktop processors. The Magny-Cours AMD Opteron is an excellent example of renewed focus on the professional workstation and Enterprise server environment, but whenever Opteron products cannot compete with Intel Xeon CPUs on sheer performance, AMD should immediately fall back to the value-priced strategy.
Despite an absurd asking price, Intel has already swayed consumer opinion by producing the best-performing desktop processor available. The only way for AMD to salvage reputation and compete with this mentality is by offering a similarly powerful product to the consumer market that sells for a fraction of the comparative cost. This is likely the legacy AMD's Phenom-II X6 will inherit. Going forward, value-priced products are really AMD's only dominant strength over Intel. AMD processors have yet to match Intel's 32nm fabrication process, which makes product development a real sore spot for AMD when it comes to selling Dell and HP on the Opteron product line. As head to head performance becomes more comparable, AMD can relax discount pricing; but not any time before then.
One other area AMD could concentrate their efforts is advancing a specific technology beyond the competition. Intel's Advanced Encryption Standard (AES) Instructions Set R3, introduced with Gulftown, posts a 2000% performance gain over the previous Nehalem microarchitecture. While I'm not suggesting the AMD take on Intel's AES capability, which is pretty well untouchable judging from the Everest AES encryption results, there are other areas that serve to gain ground. One suggestion: command queuing and pre-order fetching. Since SSDs have significantly improved storage technology response time, processors should begin to find ways of sequencing data ahead of the user's request. Computer hardware needs more than speed and power; it needs purpose.
Random After-Thoughts
New and upcoming DirectX-11 software notwithstanding, video games have generally required the same graphical power as they needed over the past few years. Newer server and virtualization technology continues to refine efficiency and uses fewer CPU cycles. So essentially software is barely moving forward while hardware is making leaps and bounds. Which raises the question: to what end?
Software just hasn't been keeping up its end of the deal, and most people still use 32-bit technology (introduced back in Windows 95). it's sad but true. Games like Crysis helped give reason for advancing graphics technology, just like virtualization technology and Terminal Services helped push processor power. But the apex of software demands hasn't really changed in several years, and having massive amounts of system memory are only helpful if there's an application that requires it.
I used to overclock my CPU to get a few extra frames out of video games, or earn more work-units per day with Folding@Home. But for the past few years, just about any graphics card can fold proteins better than the CPU ever has. We now have desktop processors that can encode multimedia in a mere fraction of the time it used to take. While hardware progression hasn't hit a wall, the light at the end of the tunnel is getting very close.
I believe that If software doesn't come around soon, perhaps coming in the shape of widespread adoption towards 64-bit computing using Microsoft Windows 7, we'll soon share the same dilemma facing the automobile industry: high-horsepower engines with break-neck torque driving down roads with a 65-MPH speed limit. This has already been the case for far too long, which is why we're already sharing the other auto-industry dilemma: power efficiency.
AMD Phenom-II X6-1090T Conclusion
Although the rating and final score mentioned in this conclusion are made to be as objective as possible, please be advised that every author perceives these factors differently at various points in time. While we each do our best to ensure that all aspects of the product are considered, there are often times unforeseen market conditions and manufacturer changes which occur after publication that could render our rating obsolete. Please do not base any purchase solely on our conclusion, as it represents our product rating for the sample received which may differ from retail versions. Benchmark Reviews begins our conclusion with a short summary for each of the areas that we rate.
Benchmark Reviews begins our conclusion with a short summary for each of the areas that we rate. The first is performance, which considers how effective the AMD Phenom-II X6-1090T processor performance in operations against similar desktop CPU products. The first challenge is properly defining the competition, which by merit of price would be Intel's Core i7-920 processor, or could be the Intel Core i7-980X Extreme Edition CPU if you want to match the physical core count. If we base the competition by price (assuming AMD's assertion of a sub-$300 MSRP is legitimate), the Phenom-II X6-1090T is the hands-down winner on all fronts, as it beats the i7-920 in nearly all performance measures while also offering an unmatched price-performance ratio over the expensive i7-980X.
Aside from encryption scores, which shot Intel's i7-980X off the charts, AMD's X6-1090T delivered impressive results. The Music tests in PCMark Vantage certainly lend proof to at least one area of dominance, while the TV and Movies tests showed us that the X6-1090T could match performance with the 980X... and cost nearly $840 less. Gaming performance was moot, since most critics would agree that 1 FPS of difference is barely enough measure, and not enough to notice. In terms of real-world professional design application performance, nearly all 4-thread SPECviewperf benchmarks agreed that AMD made the best processor for their tasks.
In consideration of construction, we consider that Turbo CORE technology was added onto the existing Phenom-II architectural design, and another two CPU cores were stuffed onto the 45nm die. While a move to 32nm would have made for the perfect story, AMD still managed to offer some above and beyond 'more cores'. The one self-evident truth to AMD's quality is in the ability to overclock, which allowed enough headroom to easily dial speeds beyond 4.0 GHz without adjusting voltages. Keeping the entire package within the same 125W TDP that their quad-core series shares is impressive, especially when you consider how they both operate at similar speeds.
As of 06 December 2010, the AMD Phenom-II X6-1090T retail processor kit (HDT90ZFBK6DGR) is available at all major online retailers. Amazon and NewEgg both stock this 3.2GHz Black Edition six-core CPU for $229. Alternatively, the six-core 2.8GHz Phenom-II X6-1055T sells for $179. PC hardware enthusiasts should compare their multi-threaded needs before they buy, since the 3.4GHz Black Edition X4-965 is still available for $186.
After review of our test results, it's difficult to ignore how well the 3.2GHz six-core AMD Phenom-II X6-1090T has done in comparison to Intel counterparts. Equally impressive is how well the 3.4GHz quad-core AMD X4-965 kept up, and confirmed the power contained within AMD's Phenom-II architecture. The X6-1090T may not have always placed first in every benchmark we tested, but it occasionally offered unrivaled performance and generally finished at the top. For a product that sells for only $309.99at the time of launch, this is an impressive feat. Benchmark Reviews recommends the AMD Phenom-II X6-1090T Black Edition processor for any multi-threaded application that benefits from six physical CPU cores, especially professional design and engineering tools such as Dassault Systemes CATIA and PTC Pro/ENGINEER. For overclockers the tweaking possibilities are made more interesting by the introduction of Turbo CORE technology combined with an unlocked Black Edition processor clock multiplier.
Pros:
+ Six physical processor cores
+ Introduces AMD Turbo CORE technology
+ Excellent price-performance ratio
+ Overlocks to 4.0GHz with 4.3GHz Turbo CORE
+ Maintains AMD socket AM3 compatability
+ Outstanding media transcoding or file compression performance
+ Dominates professional engineering and design tasks
+ Unlocked clock speed multiplier
+ Maintains 125W TDP
Cons:
- Shared L3 cache is the same between four- and six-core units
- Utilizes aging 45nm fabrication process
Questions? Comments? Benchmark Reviews really wants your feedback. We invite you to leave your remarks in our Discussion Forum.
Related Articles:
|
Comments
Keep in mind that I've just spent nearly 26 hours to offer you a comprehensive hardware review on a free website. You're encouraged to purchase your own Intel i7-930 for the sole purpose of testing another processor, and let us know how that goes.
Kudos to you benchmarkreviews.
Paul Gow
Lightning PC Systems
Moderator, Techpowerup.com
The 920 has has had a product discontinuation notice issued and the 930 is the same exact price as the 920. (if you can even find the 920 anymore.. for instance microcenter no longer sells them. noone is going to order a slower processor for the same price. that makes no sense).
Every other site out there has used the 930 as the primary intel example and show that the 930 in most cases beats or is equal to the x6 1090T or worst case compare it both as a reference point.
#content.intel.pcnalert.com/dm/d.aspx/5F8AF763.../PCN109790-00.pdf
You are ignoring the fact that the 930 is faster and the same price as the 920. Therefore making the 920 obsolete as noone is going to buy a slower processor for the same price. It's really that simple. IF anyone still has inventory.. its because everyone else is buying the faster 930 chip for relativly the same price.
I think that 920s have been shown to be better overclockers in general, in 920 vs 930 matchups I've seen many sites that pick 920s as the better value and better overclocker for the price range. I have been going back and forth in building a system and the two in consideration have been the 920 and 1090 because all research has pointed to the 920 as the better cpu for overclocking even if the 930 is faster from the base.
So, I think that overall even I was kinda confused by your attitude/snark even if I did agree with you in principle. Thanks for the review, keep up the great work!
Tom's Hardware: Used the 920 and 930
Legit Reviews: Used the 930
Guru3d: Used the 920
Tweaktown: Used neither
Anandtech: Used the 920
Hardware Secrets: Used neither
Maximum PC: Used neither
PC Perspective: Used the 920
Bjorn 3d: Used neither
Hardware Canucks: Used the 920 and 930.
I could go on, but you get the picture. Of 10 major sites, only 3 included the 930 in their comparisons. 5 used the 920, and two used both, while 4 used neither. Your assertion that reviewers should purchase their own hardware to review is silly: even the sites that could easily afford to, like Anandtech, don't do that.
I think the 920 is still the option to compare along with a lynnfield CPU.
A CPU test is a CPU test, not a RAM test.
I did wish to get this motherboard here (ASUS P6X58D Premium): #newegg.com/Product/Product.aspx?Item=N82E16813131614 until I saw this processor had been released. Would this motherboard here be one that would match or surpass the one I just mentioned: ASUS Crosshair IV Formula #newegg.com/Product/Product.aspx?Item=N82E16813131644 Maybe one that is even better that is in the same price range? If someone could help out with suggestions, I would be very grateful.
Regarding power, I've found it extremely difficult to test accurately. Even with my power meters connected, the difference between idle and load is less than 20W, with a 5W margin of error. Ultimately, the motherboard will make the biggest difference.
I've recently had a discussion on the forum of MCSA students and even there, the lack of understanding was demonstrated.
I've said it before and I'll say it again now.... RAM size does not affect how quickly your CPU runs. You will get no more FPS in Crysis for having 16GB RAM instead of 4GB.
I'll repeat for the sake of getting it through to people: 4, 6, 8, 12 or 16GB RAM does not affect CPU speed. Anyone who thinks otherwise does not understand how things work.
RAM size can affect overall *system* performance, but only due to the fact that SuperFetch and file caching can reduce Windows' reliance on the paging file for temporary storage.
4GB, for now, is the sweet spot for 99% of users.
Mr. Olin? Maybe an update test with 6GB's.... -please.?
Olin I'm not sure why you questioned what I know as a IT pro but...? And I know the difference of dual, triple and quad channel platforms.(which I never mentioned, but you responded to?). I mentioned the 6GB Ram just as a standard 4bank config. I grasp how basic even bringing that out is. My bad.
Opteron: ##amd.com/us/products/server/processors/6000-series-platform/pages/6000-series-features.aspx
So with double the bandwidth requiring more sticks of memory, you're damned right it'll run faster -- not with more RAM, but with more STICKS of RAM. You could put TWICE as many sticks of HALF THE SIZE, and the memory performance will shoot through the roof.
And before you go wondering if 1090T could possibly have a 4 or even 6-channel memory controller... count the pins on it. More memory controllers = more pins. Even if it DOES have a 6-channel memory controller, 4 of those channels are disabled due to being physically not connected.
Also, Opteron's don't have triple-channel memory controllers, thus testing with 3x2GB would be utterly pointless.
Triple-channel VS dual-channel memory controller is something that is barely worth consideration when building a system, even *if* you could ignore the fact that this means deciding between CPU families. Sure, triple-channel Intel systems blast through the memory bandwidth benchmarks, but that's one hell of a synthetic benchmark and the impact on real-world performance is at best difficult to measure and at worst completely counterintuitive.
So I was curious to see any test, even one that would look at this.
If you get a hold of the 'goods' from AMD I'll say "Thanks" ahead of time.
-again great review. Also liked your Eco review, along with the review on Fudz####. Good product.
1+1+2+2 would work but, then you would be using the mobo in a config different from the Intel test, which would be using only one of the tri-chan mem ctrlrs. not that it's even relevant here, because the tests do not appear to be RAM dependent.
Well, that pretty well explains Intel's "need you fools to be suckers for punishment" attitude towards the consumer with their 980X, ..., -forget it!
i7-930 does it all and more.
...now that nvidia is outta the way, it'll be Intel's turn soon.
AMD's platforms for this kinda-money, can't be beat. Go AMD go !
;)
i7 920
asus p6t6 mobo
6g ram tripple chan
dual gtx 260 in sli
windows 7 x64
Paul Gow
Lightning PC Systems
Moderator, techpowerup.com
1. Matched pairs (or triplets): Not needed. I have a number of hands-on testimonials to the fact that singly purchased value DIMMS work great together.
2. Low-latency RAM will give you blistering performance: Benchmarks show only marginal gains. Low-latency RAM involves increased energy usage, increased heat production and, ultimately, a stressed memory controller.
3. You will benefit massively from RAM able to run at 1600, 1800 or 2000MHz: Again, benchmarks show only marginal improvements. Chipsets have decoupled DIMM clocks from the rest of the system, thus high-MHz RAM isn't a requirement for bus based overclocking (as it used to be).
I would argue that money spent on replacing e.g. 2x2GB 1333MHz Corsair Value Select DIMMs with Corsair TwinX DominatorGT (a price difference of around £ 120) would be *far* better spent on a faster GPU, CPU or HD.
The suggestion as to a better upgrade is presumptuous, as it really depends on his needs. If he is a gamer or does any encoding, I'd argue that a GPU upgrade could have a more significant impact than the memory upgrade.
The thing is, to have purchased fast RAM in the first place is of little value when compared to spending that money elsewhere. As you say, a faster GPU can perhaps be the better option. A faster GPU and cheap, Corsair Value RAM is better than a slower GPU and RAM that is advertised as 1800MHz.
I've just advised someone building an AMD system. He presented a list of components, including a Phenom X4 and Corsair Dominator GT RAM. I've told him to get 2 separate sticks of Corsair Value instead and with the money saved he can now get a Phenom X6 with a little money left over.
I don't know how I found this site (I think it was PCPerspective but don't quote me on that) but I'm really glad. I'll start dropping in regularly now. Thanks!
##lostcircuits.com/mambo//index.php?option=com_content&task=view&id=81&Itemid=1
It's a unique review with a bit of healthy discussion as well.
default OS, and default Dashboard OS, so the proceess will be transparent, but for power users, they will be able to user WARNER, or EA dashboards, X-box live and Facebook will be "cut out of the loop"
The inclusion of high-res and high-detail GPU generated scenes should be present in a CPU review only to balance and contrast against benchmark results where the lowest resolution and detail settings were used. In this way, system builders (like myself) can make accurate assumptions and calculations about overall system performance when considering each individual component. There's no point spending £ 400 on a GPU (or £ 800 on dual GPU's) if the CPU is going to bottleneck gaming performance. Likewise, teaming a 6-core CPU with a £ 20 GPU isn't going to win favour with a customer looking only for stellar gaming performance on their 45" HDTV.
first lemme tell you the softwares i use often,
Adobe After effects CS5 (64bit), Adobe Photoshop CS5 (64bit), Autodesk MAYA, Autodesk 3Ds MAX (64bit)
Now what i want is to make new rig which'll be used at home, for R & D and some freelance and i'll be using the pc for games also the usage will be 60% work and 40% Games i have to buy almost everything, my budget is $12k .I was thinking of buying:
AMD X4 965
Asus M4A79 T Deluxe
G.Skill Ripjaws DDR3 1333 2GBX2
WD Caviar Black 1 TB, SATA 6Gb/s, 64 MB Cache, 7200 RPM
ATI HD 5850
Now m confused X6 or i7 930?? and how much boost in performance?? or if someone can give me a better config at the given price range.
Thanks
For what you're doing, the AMD Phenom-II X6 is a better choice.
And howz Gigabyte GA-890FXA-UD5 in comparison to ASUS M4A79 T Deluxe
Thanks
What you don't want on a raid 0 is your DATA.
SSD's are faster than VelociRaptors in a RAID0, which are faster than a Caviar Black (even the 7200RPM, 64MB cache, SATA3 one), which are faster than most any other HD you could care to mention.
Size up how fast you want your system to go VS how much storage you really, truly need VS how safe you want your data to be VS how much you want to spend.
Generally, data is replaceable unless it's work you've created. Put that on DVD, USB drives and cloud storage so you don't lose it.
I could put together a system for you which would keep your data absolutely safe whilst being enormously fast and with a bunch of space, but it might cost the wrong side of a year's salary.
Just get the best you can for the money you've got.
Thanks
Any AMD 890FX motherboard
AMD Phenom-2 X4 processor
4 or 8GB RAM
Radeon HD 5850
700W PSU
Windows 7 Home Premium 64-bit
Any single ATI video card. Do not underestimate the performance of a 1GB HD 5670. Brand is not important.
If you can, get a Seasonic PSU with a 90%+ efficiency rating. 650W will be enough. This is by far the best PSU money can buy. Whichever PSU you get, spend at least 10% of your budget on it. Do not be tempted to get a cheap PSU, you *will* regret it.
One 64GB Corsair Nova or 60GB Corsair Force F100 for your OS and Adobe apps.
Two 1.5TB WD EARS Green drives. Set these up as a RAID1 volume. Three partitions. One no less than 8GB only for Windows pagefile. One for your relocated Users folder and one for general storage and installing games.
G.Skill 1.3v RAM. 4GB should be plenty.
Your choice of CPU is then limited by how much you have left in your budget.
The OCZ Vertex-2 LE SSD is slightly more affordable than Corsair's Force series, and both are much better than the Nova and other Indilinx SSDs.
Asus offer an 890GX motherboard with USB3 and SATA3. Two features which are really not a big deal. SATA3 isn't going to make anything faster for you and USB3 devices are overpriced. By the time you actually want USB3, you may as well buy a USB3 add-on card if your motherboard doesn't provide it.
You might find an X4 delivers plenty of performance. Dealing with large Adobe projects is more constrained by amount of RAM and speed of HD/SSD. You'd have to be doing fairly complex stuff in order to really tax an X6. Plus, buy the right X4 and you might find you can unlock an extra 2 cores!
The Nova is one of the best performing Indilinx generally available, in real-world tests. A couple of synthetic tests show skewed results but in every practical sense, it's an excellent device. It notably doesn't suffer from pausing. I'm more than happy to build them into systems. I'm *more* happier to use Sandforce based SSD's, but people do tend to compare 60GB unfavourably against 64GB, no matter the other advantages.
As for your triple monitor requirement, an Eyefinity-6 card is not a requirement. The ATI5850 can drive three displays. Their ports differ between manufacturers, but they generally have at least 1 DVI, 1 DisplayPort and 1 HDMI. Armed with the required adapters, you should have no trouble setting up your three monitors. The 5850 still carries the EyeFinity branding, so you'll get full gaming support in the drivers.
Thanks to this X6 release, consumers like myself who already own older Intel Quad Core models can wait and KNOW that since an affordable Hex Core exists by AMD, that Intel can't for too long keep releasing Quad Cores and will have to start making Hex Cores affordable. If not for AMD, INTEL would milk Quad Cores for as long as possible. AMD broke that chain, which makes me happy.
Expect to see Hex Cores become mainstream a lot sooner than later.
With Intel, the only budget option is the Atom line of systems, which are only just performant enough to scrape by in a home PC. A very inflexible solution, as there's no easy, cheap or fine-grained upgrade path.
I don't think Intel will be forced into lowering prices by AMD. Intel are already very competitive at the price points they currently occupy. AMD will remain the ones playing catch-up (think i5). This isn't a bad thing, but don't be fooled into thinking Intel run scared of AMD.
*Choosing* a processor is not about clock speed. It's about identifying your needs and them matching them to a budget.
One thing I believe (at least for me) that reviewers left behind is the virtualization aspect/capabilities of these new 6-Core CPUs. Although I have always liked both platforms (and AMD when they were ahead of this "game") I admit that I have been using Intel for the past 6 years.
Now, I want to build a system for some virtual servers and this 6-Core at an affordable price is really nice.
I really need all the threads i can get, and for less money better. The thing is, i was planning to build a render farm, but i was gonna buy some Core2Quad q9650, and then someone suggested AMD, i did some rapid research and...man, it is so cheap!!
But, after a deeper revision, i found this -correct me if i am wrong-:
The Core2Quad q9650 has 4 cores and 12MB cache.
The Phenom II x6-1090t has 6 cores, but only 9MB cache.
Isint that gonna create a bottleneck? I am no expert,so please explain to me.
And other thing, i need a cheap Motherboard for that, good performance, cheap price, has to be less than $60usd -I am gonna buy lots of them-
Thanx a lot for you attention.
However, to answer your questions: cache will not limit rendering effectiveness. If your models and textures are complex enough that you're considering a render-farm, then cache size is totally irrelevant. Overall system memory bandwidth (and latency, to some extent) will dominate, which is where Core i7 performs superbly.
I would also recommend you not to buy cheap motherboards. Overall stability and system longevity will suffer. You will get more satisfaction having fewer, but more expensive Core i7 systems than many, cheap AMD systems. If at all possible, I would actually consider having perhaps only 1 or 2 systems based around Intel's 6-core Xeon (Westmere) CPU and a Supermicro dual-CPU motherboard.
Loaded up with plenty of ECC RAM, such a system will absolutely scream through render tasks. 12 true cores in one system. Worth it? Sure!
So, why there is a change in thought its confusing buddy... hope you understand...
OK, you will see improvement using Maya and 3DS if you spend more to get an i7 system (and I don't mean an i7-930, I'm talking the expensive end of the i7 range). But... you're 40% gaming. An X6 will scream through games. Your GPU will bottleneck games, with the likes of Metro 2033 taxing even SLI Geforce 480's.
Anyone with infinite budget will be best served with a 6-core Intel Xeon CPU. Everyone else needs to consider exactly how many seconds they really need to shave off their 3D render before the price:performance ratio starts to get too steep. Perhaps I've underestimated your 3d rendering needs?
everest=
cpu queen: 42 029
photoworxx: 42 294
cpu zlib: 124 437
cpu aes: 29 239
avec une msi260gtx 20 000 3Dmark06
My next alternative system will be AMD phenom II 1090t. that is for sure.
All the best.
I am a photographer based in INDIA New Delhi.
Could anyone help me to opt/choose the right motherboard for "AMD Phenom II X6 1090T Black Edition Hexa Core Processor", because i fairly do not have any idea as which one to go for, one of the Top Stores in india suggested me to go with GIGB & ASUS...
However i just wanted to check which one is the best as you people have a good knowledge over the products.
My requirement is for a high end purpose for Extreme Gaming and Multimedia.
Your Quick response to this thread would be highly appreciated.
Regards,
DSR V
#newegg.com/Product/Product.aspx?Item=N82E16813131667