| AMD Phenom-II X6-1100T CPU HDE00ZFBRBOX |
| Reviews - Featured Reviews: Processors | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Written by David Ramsey | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Tuesday, 07 December 2010 | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
AMD Phenom-II X6-1100T Black Edition CPU ReviewWhile enthusiasts await AMD's next-generation "Bulldozer"-based desktop CPUs, AMD continues to expand their current desktop CPU line, introducing the Phenom II X6-1100T processor. The 6-core Phenom-II X6-1100T takes over the top spot in AMD's CPU line from the 1090T, and does it at a suggested retail price of only $299. The clock speed improvements over the 1090T are minimal: both the standard and Turbo Core speeds increase only 100MHz, to 3.3GHz and 3.7GHz, respectively. Benchmark Reviews tests the new 1100T against a collection of Intel and AMD processors in gaming and computing performance. Although it's only been a few months since AMD last updated their processor line (bringing us the Phenom-II X6-1075T, among others), AMD has found the time for what's perhaps the last iteration of the Thuban architecture, bringing us the Athlon-II X3-455, the Phenom-II X2-565 Black Edition, and the subject of this review, the Phenom-II X6-1100T Black Edition HDE00ZFBK6DGR CPU. This new 6-core processor takes over the top spot from the X6-1090T Black Edition, increasing the base clock speed from 3.2 to 3.3GHz, and the Turbo Core clock speed from 3.6 to 3.7GHz. These speed tweaks are minor, but at a suggested retail price of only $299, the 1100T might just be one of the best bang-for-the-buck CPU deals today. Although many have forgotten it now, there was a time when AMD processors handily outperformed their Intel equivalents. AMD was the first company to break the "gigaHertz barrier" back in 2000 (Intel's 1GHz Pentium III shipped a few days later). A few years later, I built a system using the then-new dual-core Athlon 64 X2 processor for encoding video, and it absolutely stomped my existing dual-processor Pentium III-833 system.
But that was then, and this is now, and AMD has ceded the high end processor market to Intel, while working on their position in the low-to-mid end of the market. This has been a boon for enthusiasts, since desktop processors reached the "fast enough" mark some time ago, and money saved on a CPU can be invested in other parts of the system, often with better overall performance results. Paired with AMD's new 800 series chipsets, which offer extra PCI-E lanes for CrossfireX, SATA 6G and USB 3.0 without needing expensive add-ons like the NVIDIA NF-200, AMD's Thuban processors offer the enthusiast a way to build an affordable, yet very powerful and versatile 6-core production or gaming system. After current manufacturing techniques hit a "megaHertz wall" at about 4GHz a few years ago, both Intel and AMD have concentrated on multi-core CPUs, and it's a rare system these days that's not equipped with at least two cores. As software evolves to take advantage of the performance benefits offered by multiple native threads, we'll see the performance of multi-core systems continue to improve. By driving the cost of 6-core processors downwards, AMD's Phenom-II X6 line keeps the price of these capabilities within reach of the average enthusiast.
Manufacturer: Advanced Micro Devices Full Disclosure: The product sample used in this article has been provided by AMD. Phenom-II X6-1100T FeaturesThe following information is courtesy of Advanced Micro Devices. Turbo CORE TechnologyAMD 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
AMD Balanced Smart Cache
AMD Wide Floating Point Accelerator
HyperTransport Technology
Integrated DRAM Controller with AMD Memory Optimizer Technology
AMD Virtualization (AMD-V) Technology With Rapid Virtualization Indexing
AMD PowerNow! Technology (Cool'n'Quiet Technology)
AMD CoolCore Technology
Dual Dynamic Power Management
HDE00ZFBK6DGR Specifications
Testing and ResultsBenchmarking processors is difficult, since it's hard to isolate the CPU performance from the rest of the system. This is especially true when testing the performance of several processors across multiple platforms, because variables like the motherboard and chipset used will affect the results. In this review I'm running various synthetic and application benchmarks across three different platforms: Intel P55, Intel X58, and AMD 890FX. As one example of the differences between these platforms, the P55 and 890FX use dual-channel memory, while the X58 platform uses triple-channel memory and thus benefits from higher memory bandwidth. Still, the tests are valid in the sense that anyone using an Intel Core i7-930 will be using it in an X58 system, while Core i5-860 and AMD 965BE processors will likewise be on motherboards with their respective chipsets and architectures. Most of the benchmarks I've chosen should favor raw CPU performance, and I've tried to minimize the intra-system variables by using the same memory, hard disk, and video card for all three platforms. But be aware that it's impossible to completely isolate a processor from the rest of its system in these tests. All processors were tested at their stock clock rates, but I also tested the Phenom-II X6-1100T processor at the highest stable overclock I could achieve, 4.1GHz. This overclock was with the Turbo Core feature disabled, since I like to reach the highest speed all cores can sustain, and Turbo Core could cause the system to crash when it suddenly ramps up the CPU multiplier. Intel P55 Test Platform
Intel X58 Test Platform
AMD 890FX Test Platform
Benchmark Software
CINEBENCH R11.5 BenchmarksMaxon 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 comprises three tests: an OpenGL-based test that models a simple car chase, and single-core and multi-core versions of a CPU-bound computation using 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, and all the rendering is performed by the CPU: the graphics card is not involved except as a display device. The multi-core version of the rendering benchmark uses as many cores as the processor has, including the "virtual cores" in processors that support Hyper-Threading. The resulting "CineMark" is a dimensionless number only useful for comparisons with results generated from the same version of CINEBENCH. First, let's look at the OpenGL results.
Here we see a surprisingly even progression up the processor scale, with the AMD processors overall doing better (although not dramatically so) than the Intel processors. Of course, the processor plays a relatively minor role in this test, with the graphics card shouldering the bulk of the work.
In the single-core rendering test, the Intel 980X takes the lead with a score 7.4% better than the second-place overclocked Phenom-II X6-1100T. The standard-clocked 1100T gets third place, while the rest of the crowd brings up the rear with very small differences in the score. Readers may be surprised to see the overclocked 1100T turning in such a small improvement over its stock-clocked score, but recall that since this is a single-core test, the CPU can throttle itself up to 3.7GHz. Still, the difference between all of the CPUs tested remains small. Things change with the multi-core test: the mighty 980X's ability to spawn 12 threads puts it solidly in the lead with a score 22% higher than the second-place overclocked 1100T. The other six-core AMD processors come in next, followed by the four-core Intel CPUs and the AMD 965BE. Let's see how things look on a "Dollars per CineMark" basis, based on current Newegg prices:
The AMD 965 Black Edition wins the "dollars per CineMark" comparison for single-core performance, while the AMD Phenom-II X6-1075T wins in multi-core performance value. The Intel Core i7-980x's score is 22% higher than that of the overclocked 1100T, but it comes at more than double the "cost per CineMark." SPECviewperf 11 testsThe Standard Performance Evaluation Corporation is "...a non-profit corporation formed to establish, maintain and endorse a standardized set of relevant benchmarks that can be applied to the newest generation of high-performance computers." Their free SPECviewperf benchmark incorporates code and tests contributed by several other companies and is designed to stress computers in a reproducible way. SPECviewperf 11 was released in June 2010 and incorporates an expanded range of capabilities and tests. Note that results from previous versions of SPECviewperf cannot be compared with results from the latest version, as even benchmarks with the same name have been updated with new code and models. SPECviewperf comprises test code from several vendors of professional graphics modelling, rendering, and visualization software. Most of the tests emphasize the CPU over the graphics card, and have between 5 and 13 sub-sections. For this review I ran the Lightwave, Maya, and Seimens Teamcenter Visualization tests. LightwaveThe lightwave-01 viewset was created from traces of the graphics workloads generated by the SPECapc for Lightwave 9.6 benchmark. The models for this viewset range in size from 2.5 to 6 million vertices, with heavy use of vertex buffer objects (VBOs) mixed with immediate mode. GLSL shaders are used throughout the tests. Applications represented by the viewset include 3D character animation, architectural review, and industrial design. MayaThe maya-03 viewset was created from traces of the graphics workload generated by the SPECapc for Maya 2009 benchmark. The models used in the tests range in size from 6 to 66 million vertices, and are tested with and without vertex and fragment shaders. State changes such as those executed by the application- including matrix, material, light and line-stipple changes- are included throughout the rendering of the models. All state changes are derived from a trace of the running application. Siemens Teamcenter Visualization MockupThe tcvis-02 viewset is based on traces of the Siemens Teamcenter Visualization Mockup application (also known as VisMockup) used for visual simulation. Models range from 10 to 22 million vertices and incorporate vertex arrays and fixed-function lighting. 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 Lightwave results favor Intel. Performance scales well, with scores going up with both clock speed and number of cores. In the AMD arena, extra cores don't seem to buy you much, with the 3.4GHz AMD Phenom-II 965 Black Edition posting a better score than the 3.0GHz Phenom-II X6-1075T and even the 3.2Ghz and 3.3GHz 1090T and 1100T. Only the overclocked Phenom-II X6-1100T running at 4.1GHz beats it. Note here how the Intel 980X score is only 2.7% better than the overclocked 1100T's score. The results flip for the Maya scores, with the AMD processors pulling strongly away from the Intel processors. The $159 AMD 965 even beats the $999.99 Intel 980X, as does the overclocked 1100T. The quad-core Intel processors come in far behind the rest of the pack. The Seimens TCVIS scores are relatively even between the Intel and AMD camps. Surprisingly, the budget Core i5-750 edges ahead of the Hyper-Threading Core i7-930, and the 980X again turns in the best overall performance. The minor differences in clock speed between the hexacore AMD CPUs are actually apparent in the scores here, although not to the degree one would imagine, and the overclocked 1100T is only 10% faster in this test than it is at the stock clock speed. One thing these tests show is that some code favors multiple cores, and some code favors clock speed. SPECapc (Application Performance Characterization) tests are fundamentally different from the SPECviewperf tests. While SPECviewperf tests incorporate code from the various test programs directly into the benchmark, the SPECapc tests are separate scripts and datasets that are run against a stand-alone installation of the program being benchmarked. SPECapc group members sponsor applications and work with end-users, user groups, publications and ISVs to select and refine workloads, which consist of data sets and benchmark script files. Workloads are determined by end-users and ISVs, not SPECapc group members. These workloads will evolve over time in conjunction with end-users' needs and the increasing functionality of PCs and workstations. For this test, I ran the SPECapc "Lightwave" benchmark against a trial installation of Newtek's Lightwave 3D product. The benchmark, developed in cooperation with NewTek, provides realistic workloads that simulate a typical LightWave 3D workflow. It contains 11 datasets— ranging from 64,000 to 1.75 million polygons— representing such applications as 3D character animation, architectural review, and industrial design. Scores for individual workloads are composited under three categories: interactive, render and multitask. The benchmark puts special emphasis on processes that benefit from multi-threaded computing, such as animation, OpenGL playback, deformations, and high-end rendering that includes ray tracing, radiosity, complex textures and volumetric lighting. The test reports three scores: Animation (multitasking), Animation (interactive), and Rendering. The numeric scores represent the time it took to complete each section of the benchmark, in seconds, so lower scores are better. 
Although this test stresses system components other than the processor (the video card's OpenGL implementation, for example), it still shows obvious performance differences in the CPUs. In the Animation (Multitasking) section, we see a very nice performance scaling with frequency and number of cores in both the Intel and AMD camps, although Intel does better overall, with the Core i7-930 beating every AMD processor except the overclocked Phenom-II X6-1100T. That said, note that the very expensive Intel 980x turns in a score only 13% better than the overclocked 1100T. In the Animation (Interactive) test, the results are much more even, with the best score weirdly turned in by the cheapest, slowest CPU, the Core i5-750, and the worst score by the Core i7-930. In the Rendering test, we're back to nice performance scaling again, with the relative processor rankings almost exactly what they were in the Animation (Multitasking) benchmark. The results of this test are somewhat different than the results returned by the Lightwave code built into SPECviewperf. Although the overall trend is similar (with Intel tending to lead), a few processors switch relative performance positions. PCMark Vantage TestsPCMark Vantage is an objective hardware performance benchmark tool for PCs running 32- and 64-bit versions of Microsoft Windows Vista or Windows 7. It's 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 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
Gaming Suite*
Music Suite
* EDITOR'S NOTE: Hopefully our readers will carefully consider how relevant PCMark Vantage is as a "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) have 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). 
The TV and Movies suite concentrates on video playback and transcoding, but only uses two threads at a maximum, so hexacore processors have little advantage over quad-cores. The results scale pretty much with clock speed, and tend to favor Intel slightly. It's interesting that the 3.33GHz Core i7-980X processor beats out the 4.1GHz AMD 1100T since the other results scale closer to the clock speed differences within each camp. The Gaming benchmark relies on the hard disk and video card for over 50% of its score (see the Editor's Note above), and we're using the same HDD and video card for all platforms, so the Intel processor's decisive win in this test simply means that Vantage's gaming code is more optimized for Intel processors. Bear in mind, however, that most "real world" games will not show this difference; generally, in games, your video card matters most, followed by the clock speed (not number of cores) of your processor. The PCMark Vantage gaming test can use up to 16 threads, but very few commercial games will take full advantage of multicore processors. In the Music test, we see that more cores is definitely better, and unlike the Gaming test, these results have more real-world relevance, since multi-threading is much more common in music transcoding applications than it is in games. The Intel processors dominated this test, which is something I haven't seen in previous tests (albeit with somewhat different hardware). And I'm still having trouble coming up with an explanation for the 965 Black Edition's dominance of the AMD results, which is also something I haven't seen before...but I ran each test three times, and these are the results. Futuremark's weighing of the various system components in each test is the subject of some debate; and some of their choices (such as the Gaming test's use of a 1024x768 resolution with no anti-aliasing or texture filtering being "representative" of the "consumer experience") seem odd to me, but the TV and Movies and Music benchmarks are arguably reasonable predictors of overall system performance. AIDA64 Extreme Edition TestsAIDA64 Extreme Edition is the evolution of Lavalys' "Everest Ultimate Edition". Hungarian developer FinalWire acquired the rights to Everest in late November 2010, and renamed the product "AIDA64". The Everest product was discontinued and FinalWire is offering 1-year license keys to those with active Everest keys. AIDA64 is a full 64-bit benchmark and test suite utilizing MMX, 3DNow! and SSE instruction set extensions, and will scale up to 32 processor cores. An enhanced 64-bit System Stability Test module is also available to stress the whole system to its limits. For legacy processors all benchmarks and the System Stability Test are available in 32-bit versions as well. Additionally, AIDA64 adds new hardware to its database, including 300 solid-state drives. On top of the usual ATA auto-detect information the new SSD database enables AIDA64 to display flash memory type, controller model, physical dimensions, and data transfer performance data. AIDA64 v1.00 also implements SSD-specific SMART disk health information for Indilinx, Intel, JMicron, Samsung, and SandForce controllers. All of the benchmarks used in this test— Queen, Photoworxx, ZLib, hash, and AES— rely on basic x86 instructions, and consume very littlr system memory while also being aware of Hyper-Threading, multi-processors, and multi-core processors. Of all the tests in this review, AIDA64 is the one that best isolates the processor's performance from the rest of the system. While this is useful in that it more directly compares processor performance, readers should remember that virtually no "real world" programs will mirror these results.
Queen and Photoworxx tests are synthetic benchmarks that iterate the function many times and over-exaggerate 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. Here we see the $159 AMD 965 Black Edition handily beating the $185 Intel Core i5-750, while the $199 AMD 1075T and $229 AMD 1090T run just behind Intel's $280 Core i7-930. The overclocked 11100T beats everything except the i7-980X, which runs away from the pack with results 39% better than the next closest competitor. 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 AIDIA64 fPhotoworxx benchmark performs the following tasks on a very large RGB image:
The Intel processors dominate these results, with every Intel CPU doing substantially better than every AMD CPU. Since this test cannot effectively use more than two threads, the hexacore processors have no advantage over quad-core processors. AMD 965 Black Edition is the best performer in the AMD camp, outperforming even the overclocked 1100T. The worst performance is turned in by the stock-clocked Phenom-II X6-1100T, and the best, as usual, by the 980X. 
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 data-independent footprint that can be reduced at some cost in compression. 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. Both of these tests are much more applicable to the "real world" than the previous tests. The results of the Zip compression test scale almost linearly with processor clock speed on the quad-core Intel side, but there's some threading going on as can be seen with the Core i7-980X results, where two extra cores and an extra 530MHz or so virtually double the performance of the i7-930. Extra cores help on the AMD side, too, with all of the hexa-core CPUs handily outperforming the 3.4GHz quad-core 965 Black Edition. The AES encryption test is utterly dominated by the Core i7-980X, whose performance is 6.7x better than its closest competitor. The reason is Intel's Advanced Encryption Standard New Instructions (AES-NI), which dramatically accelerate AES code. AES-NI aside, the AMD processors uniformly perform better than the Intel processors without this feature. Other than this, we see a nice, even progression as processor clock speed and number of cores increases. Note that the very high score for the Intel 980X visually diminishes the performance differences among the other processors.
AIDA64's Hash test is new (it wasn't in Everest). The 64-bit hash benchmark measures CPU performance using the SHA1 hashing algorithm. Written in assembler code, the Hash benchmark is optimized for every popular AMD, Intel and VIA processor core variants by utilizing the appropriate MMX, MMX+/SSE, SSE2, or SSSE3 instruction set extension. The Intel results are a little uneven, with the 2.66GHz, non-Hyper Threaded Core i5-750 beating, by just a little, the 2.8GHz Hyper-Threaded Core i7-930. On the AMD side, we see a very smooth progression and clock speed and number of cores increase. In fact, the scores of the four hexa-core processors scale in perfect lock-step with their frequencies. Video Gaming TestBenchmark Reviews continually evaluates the various tests and benchmarks we use, and we have switched from Ubisoft's Far Cry 2 benchmark to CAPCOM's Street Fighter IV benchmark. Street Fighter IV uses a new, built-from-scratch graphics engine that enables CAPCOM to tune the visuals and performance to fit the needs of the game, as well as run well on lower-end hardware. Although the engine is based on DX9 capabilities, it does add soft shadows, High Dynamic Range lighting, depth of field effects, and motion blur to enhance the game experience. The game is multithreaded, with rendering, audio, and file I/O all running in different threads. The development team has also worked to maintain a relatively constant CPU load in all parts of the game so that on-screen performance does not change dramatically in different game scenarios.
I ran the Street Fighter IV benchmark at low-resolution, low settings as well as high-resolution, high settings. Low-resolution settings were 1024x768, no AA, with all other settings set to minimum; high resolution tests were run at 1920x1280 with 8xAA and all other settings maxed out. Low-resolution gaming tests make the video card less of a factor since any high-end video card like the NVIDIA GTX280 used in these tests can easily handle them; differences here are more biased towards processor horsepower. The Intel Core i5-750 brings up the rear here, but the real surprises are that the 2.66GHz, 4-core i7-930 performs identically to the 3.33GHz, 6-core 980x, and that every AMD CPU beats every Intel CPU. Yep, even the AMD 965 Black Edition trounces the Core i7-980x with a score almost 16% higher. Although the game is multithreaded, the low-resolution performance results seem to favor clock speed over the number of cores (although every processor here is at least a quad-core, so perhaps more differences would be seen comparing a dual-core to a quad-core CPU). In the high-resolution tests, as expected, the results are all similar, since the graphics card becomes the deciding factor. All the AMD processors turn in more or less identical results, with the Intel processors very close, but showing inverse performance, i.e. the Core i5-750 is best, then the Core i7-930, then the Core i7-980x. At Benchmark Reviews, we strive to keep this point in front of our readers: your processor makes relatively little difference in your gaming experience. PassMark PerformanceTest 7.0The PassMark PerformanceTest allows you to objectively benchmark a PC using a variety of different speed tests and compare the results to other computers. PassMark comprises a complete suite of tests for your computer, including CPU tests, 2D and 3D graphics tests, disk tests, memory tests, and even tests to determine the speed of your system's optical drive. PassMark tests support Hyper-Threading and systems with multiple CPUs, and allow you to save benchmark results to disk (or to export them to HTML, text, GIF, and BMP formats). Knowledgeable users can use the Advanced Testing section to alter the parameters for the disk, network, graphics, multitasking, and memory tests, and created individual, customized testing suites. But for this review I used only the built-in CPU tests, which aren't configurable. PassMark computes a "CPU Marks" score based on the scores of the individual tests:
The CPU Mark results scale as I'd expect. The Intel Core i5-750 edges ahead of the AMD Phenom-II 965 Black Edition, but only barely; the Core i7-930 is pretty much even with the AMD 1075T, and after that the AMD processors pull away from the Intel CPUs until the 980x, as usual, comes to the fore. Now let's look at the individual tests.
Integer and floating point operations are the basic things modern CPUs do. Integer operations are everything except floating point; technically, even instructions like comparisons, branches, and bit rotates are integer instructions. Floating point instructions deal specifically with floating point math operations. For example, an integer division of 3 into 7 will return "2" as the result, whereas a floating point division of 3 into 7 will return 2.333... as the result. While most program code is comprised of integer instructions, floating point instructions are important in modelling and rendering applications. Intel CPUs utterly dominate in the integer tests, with even the i5-750 beating the overclocked 1100T by more than 50%. On the floating point side of things, though, the order reverses, with AMD processors winning except for the 965 Black Edition against the i7-980x. The excellent floating point results of the AMD CPUs help explain how the AMD processors keep up in the rendering benchmarks.
The Compress and Sort CPU sub-tests are both integer-based, and the Intel processors continue to dominate, although the results are closer in the string sort.
AMD CPUs have an edge in the Encrypt portion of the test, with all AMD processors beating most Intel processors; the exceptions are the 980x being the overall performance champ and the 965 Black Edition being a few percentage points behind the i7-930. Intel dominates in the Physics test, though, with the 930 beating every AMD processor except the overclocked AMD 1100T.
The Primes test would favor the AMD camp were it not for the oddly high score of the i5-750. In the SSE test (which performs matrix rotations and multiplications, important operations in 3D transformations among other things), the pendulum swings back towards the AMD camp, with every AMD processor beating every Intel processor except the Core i7-980x, which is only narrowly edged out by the overclocked Phenom-II X6-1100T. But enough synthetic benchmarks. Let's get to something a little more real-world... Video Transcoding TestsFew consumer applications will make good use of a six-core processor, or even a four-core processor. Extra cores can give you a system that remains responsive when performing a computationally-intensive background task, but will rarely accelerate the execution of an individual program. There are several reasons for this:
Handbrake 0.94All that said, media transcoding (converting to a different format) is something that does scale well with the number of available cores, and the free and open-source Handbrake 0.94 video transcoder is an example of a program that makes full use of the computational resources available. For this test I used Handbrake 0.94 to transcode a standard-definition episode of Family Guy to the "iPhone & iPod Touch" presets, and recorded the total time (in seconds) it took to transcode the video. 
Here we finally see a real-world program that can fully utilize multi-core CPUs. The results scale nicely with both clock speed and the number of available cores. The AMD processors do themselves proud, with every six-core AMD chip beating every four-core Intel chip. Note, however, that the four-core AMD 965 Black Edition is only very slightly faster than the Core i5-750 and slower than the Core i7-930. The overclocked AMD Phenom-II X6-1100T edges just ahead of the stock-clocked Intel Core i7-980x. One might ask why the 980X didn't do better, since its Hyper-Threading feature gives it 12 virtual cores, and the answer is simply that a virtual core is an abstract construct wherein the processor tries to schedule code to make the best use of available execution units; it's not the same as a real extra core. The results depend on the task and the code, and Handbrake code just doesn't benefit much from Hyper-Threading. You can compare the bang-for-the-buck performance of these CPUs in Handbrake 0.94 by making a "dollars per reciprocal second" measurement: in other words, we're looking for the best combination of low processor price and low encoding times.
The 1075T wins this comparison, beating the overclocked AMD Phenom-II X6-1100t Black Edition. It's telling that only one of the AMD processors exceeds 400, while none of the Intel processors are under 400. In this calculation, the AMD 1075T is 3.9 times "better" than the 980X, which is reasonably close to the 980X's 5x-more-expensive price differential. The 1090T and 1075T are the easy winners in this admittedly ad hoc bang-for-the-buck measurement. x264 HD Benchmark 3.19Tech ARP's x264 HD Benchmark comprises the Avisynth video scripting engine, an x264 encoder, a sample 720P video file, and a script file that actually runs the benchmark. The script invokes four two-pass encoding runs and reports the average frames per second encoded as a result. The script file is a simple batch file, so you could edit the encoding parameters if you were interested, although your results wouldn't then be comparable to others.
The AMD hexacore processors maintain their lead over the Intel processors (other than the 980x) here, although the lead is much slimmer than it was with Handbrake, and the extra cores don't seem to improve performance as much as they do with Handbrake. In the Handbrake test, the AMD 1075T six-core processor finished in 22% less time than the four-core AMD 965, but in this test the difference is only about 2%. Of course, "seconds to encode video file" isn't directly comparable to "average frames per second", but I still expected more differences here.
For runs 3 and 4, the differences between processors are much larger. The Intel Core i5-750 and Core i7-930 manage to just squeak by the AMD 965 Black Edition, and the performance scaling with clock frequency and cores is smooth and regular. AMD Black Edition CPU OverclockingLike the 1090T Black Edition and 965 Black Edition, the 1100T has an unlocked multiplier, so while you can overclock it in the classic fashion by increasing the core frequency (which ASUS calls the "CPU Bus Frequency"), it's easier to simply increase the multiplier. Although the processor seemed to run stably at 4.2GHz (achieved with a multiplier of 21) with the voltage set to 1.5V, it wouldn't complete the SPECapc Lightwave benchmark at this setting (after three attempts), so after several tries at some dangerously high voltages, I was forced to drop the multiplier to 20.5 for a final speed of 4.1GHz with a 0.1875 overvolt. I always disable auto-overclocking features like Turbo Core when I overclock, since I'm trying to reach the highest frequency all cores can sustain and don't want a system suddenly crashing when the CPU tries to crank things up a few hundred MHz. For cooling, I used a CoolIT Systems ECO A.L.C. Eco-R120 CPU cooler. AMD's Turbo Core feature works so well that even enthusiasts may wish to consider if overclocking a Thuban-core CPU is worth the extra energy and cooling required. In several of the benchmarks in this review you'll notice very little difference between the scores returned by the stock-clocked and overclocked 1100T. The reason is that if the benchmark uses three or fewer of the six available CPU cores, the cores will be running at up to 3.7GHz, which is less than 10% slower than 4.1GHz. Since the CPU can dynamically vary the clock speed of the cores as the computational load changes, the processor will run at its most efficient settings in most cases, only cranking things up when it's needed. For applications that actually stress all six processor cores, you can expect performance to increase almost linearly with clock speed, but since few applications do this, you might just want to let Turbo Core handle this for you. The overclock certainly won't make any difference in gaming. 
If you've read Benchmark Reviews' previous coverage of the AMD Phenom-II X6-1075T Processor and the AMD Phenom-II X6-1090T processor, you'll see the 4.1GHz overclock achieved with the AMD Phenom-II X6-1100T Black Edition processor is very close to the 4.155GHz reached with the 1075T and the 4.0GHz (with Turbo Core enabled) reached with the 1090T Black Edition. At the time of this writing, the Newegg prices for these three processors are:
Now, there's never a guarantee that any given Thuban-series processor will reach the same overclocks I did (and if you're overclocking a non-Black Edition CPU like the 1075T, you'll need a good quality motherboard and RAM to hit the highest overclocks), but I'd be very surprised if you couldn't at least reach 3.9GHz or 4.0GHz on any hexacore AMD CPU. That being the case, there's no reason to spend more money for the 1100T Black Edition when its $70-cheaper sibling 1090T Black Edition is just as fast, and the $100-cheaper 1075T can be just as fast if you're using a good-quality AMD 800-series motherboard like the ASUS Crosshair IV Extreme I used in this review. Even with all CPUs running at their stock clock speeds, you'll notice very little difference in most benchmarks. Still, the whole point of overclocking (originally, anyway) was to make a less-expensive part perform like a more expensive part, and the overclocked AMD Phenom-II X6-1100T Black Edition provides most of the performance of the much more expensive Intel Core i7-980x Extreme, as I'll show in the next section where I'll explain my conclusions. AMD X6-1100T Final ThoughtsAMD has played catch-up to Intel for many years now. For a brief period in the early years of the decade, AMD was the undisputed CPU performance champ, easily trouncing Intel's "NetBurst" CPUs, but those days are long gone. Although AMD fanboys argued that their "true" multi-core CPUs were more elegant and sophisticated than Intel's initial crude attempts, which involved simply putting two separate CPU dies in a single package (and later, two dual-core dies into a single package) and forcing the cores to communicate across the front-side bus, at the end of the day the Intel CPUs were simply faster. Intel fans initially scorned the Thuban-core CPUs, especially when the publicity shots of the 6-core Thuban die show that the extra two cores seem to have almost been "tacked on": even someone who didn't know anything about processor design can see that it's obviously a four-core design at its heart. But there's one thing that a lot of people (even reviewers) seem to miss: while Intel continues to hold the performance crown, AMD is the undisputed leader in performance per dollar. And this is quite significant now that desktop processors have long since hit the "fast enough" point, where additional performance is unnoticeable in most applications. AMD's current processor lineup is much closer to Intel's in terms of performance than it has been in a while, as many of the tests in this review show: for example, the $159 AMD 965 Black Edition equals or exceeds the $184 Intel Core i5-750 in CINEBENCH, all three SPECviewperf tests, the Vantage Music test, four out of five AIDA64 tests, the Street Fighter IV low-resolution test, and is equal or slightly better in most other tests...and it also has the enthusiast cred that an unlocked multiplier means. The Intel Core i7-980x is obviously the fastest single CPU you can buy now. But what level performance do you get for the extra $700 you'd pay over an 1100T? The chart below compares the scores of the 980x and (overclocked) 1100T in each test, with the difference shown as a percentage— a positive percentage means the 1100T won, while a negative percentage means the 980x won.
On the average, on these benchmarks, the overclocked AMD 1100T Black Edition is 13% slower than the Core i7-980x. If we remove the AIDA64 AES score, the different drops to 9.5%. Admittedly, this is an average; most of the scores exhibit more of a difference than this, and a different mix of benchmarks would skew this percentage one way or the other. But this is still a reasonable estimate of the real performance difference you'll see, overall, between these CPUs. All that said, price decreases on Intel Socket 1366 processors and the introduction of several $200-and-under X58 motherboards have seriously eroded AMD's advantage in the mid-price field. Previously, building an Intel rig on an AMD budget meant going with a P55 system, but now that a Core i7-930/X58 system can be configured for about the same price as an 1100T/890FX system, the choice is less obvious. Although the 1100T's six cores make it faster than the 930 in media transcoding and other tasks that can utilize multiple cores effectively, the 930 is very overclockable, and its superior integer performance and Hyper-Threading feature will give it an advantage in many situations. The performance of your processor is only one facet of the performance of your system, and in most cases it's not the most important once you've hit the "fast enough" line. If your processor is in the AMD 965BE/Intel 930 class, your performance dollars are better spent on a good video card or SSD rather than a fancier CPU. While six cores are of little benefit to most users, if you want a hexacore CPU for media transcoding or just bragging rights, AMD gets you much of the performance of the Core i7-980x for a fraction of Intel's price. HDE00ZFBK6DGR ConclusionBenchmark tests should always be taken with a grain of salt. It's difficult to try and isolate the performance difference a single component in a computer system makes, especially when it's necessary to compare across different manufacturers and platforms. Complicating the matter is the fact that benchmarks change, a manufacturer may change the technical details of a product, and the retail price may change as well. So please use this review as just one piece of information, and do your research before making a buying decision. Considered on its own, the AMD Phenom II X6 1100T is an incredible processor. A 3.3GHz, six-core processor for $269.99 showcases how far the price-performance ratio has come in just the last year. The problem is, the 1100T can't be considered on its own because of AMD's own 1075T and 1090T. These three processors are identical (except for the multiplier and the fact that the 1075T is not a "black edition"), their stock performance is very close (less than a 10% difference, on average, between the 1075T and the 1100T) and they all overclock to 4.0-4.1GHz with good cooling. The $100 extra you'd pay for an 1100T over a 1075T can mean the difference between a Radeon 5830 and a Radeon 5870, and with that upgrade you'll see a bigger improvement in your gaming experience. The only reason I can't give this processor a golden tachometer is because AMD's other Thuban processors are just as good, and cheaper.
There are still reasons you might want to "go Intel": perhaps you need the incredible performance of the AES-NI features of their latest CPUs, or you want to run an NVIDIA SLI setup (which, due to some weird political wrangling between AMD and NVIDIA, you cannot do with current AMD-chipset motherboards). Intel continues to reign at the highest performance levels, but if you need or want hexacore goodness at a price you can actually afford, then an AMD Phenom-II X6 CPU is your baby. Pros:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Comments
It's good to see Big Green as strong as Intel again, or close enough to be called such. It's taken 6 years, but AMD is back. I bought one of the first AMD x64s (4800) back in 2004 and loved it. I gave it to my nephew and it's still going strong.
I doubt any processor maker will ever deliver the knock out blow to its competition that AMD did with the first x64s to Intel. But you never know.
When I needed more power than the old x64 4800 had, I was stuck because Intel still wasn't putting the north-bridge controller on the chip, and I hated that for performance and heat reasons, and in 2008, AMD just didn't have the power I wanted. Then the Core ixx series came out. I couldn't beleive my eyes when I saw the OC potential of those CPU's.
The Core i-series processors are incredible, and even though AMD is the cost per clock ratio winner, if you want the best dollar to performance ratio, it's hard to beat the i7 920 clocked to 3.8Ghz on air.
Also, comparing CPUs for gaming performance is like comparing race cars for seating comfort... FAIL.
What's really funny is that the Core i7 920 is so popular now that it's selling for 25 dollars more than it did two years ago, up from 270 to around 300USD. If I were buying today, I might go with AMD since I do TONS of heavy multitasking, such as having Adobe Lightroom, Photoshop, split tab browsers with 20 tabs each open, Eve online, Team Speak, Winamp streaming, video software for my outdoor cams, etc. so the more threads the better.
L2 benchmark, then come back.
I bet my Athlon-II X2-260 can get great FPS on Crysis. Oh, could it depend on GPU?
Being in the market for a CPU and Mainboard right now, I have to say that I'm looking hard at these while holding out to see what value Sandy Bridge brings to the table. I'm glad to be looking at buying parts now, and not a year ago. So much has changed,........
I think that at this point, the fight is between the 1100T, the i7 930, and perhaps the i7 860. I'll be waiting until next year though, to upgrade my current q6600 to either the Fusion or X68 platform.
sometimes i do that.
good review.
i don't give a f**k about biases. these are somehow inebitable in time, i have mines too, but: the most important thing is: NO MONOPOLY!
retards like Toeringsandthong can't understand that.
Am I reading it right that the test for the I7's are using triple channel. How about another test with the I7 running on dual channel.
Overall I like the article.:-)
It's not that I am a fan of Intel over AMD (as if I have to even prove myself, which is ridiculous - but for what it's worth I own a 1075T which, un-overclocked, is just about the best performance per dollar that money can buy for the work loads that I encounter), but I would really appreciate seeing meaningful comparisons rather than meaningless comparisons, regardless of which processor comes out on top.
You want eyeballs on your reviews and hits on your advertisements, then you need to work to have high quality reviews worth visiting. I'm not saying this review was bad, I read all the way through it because it was interesting, but you need to take my point seriously rather than trying to dodge it. Not overclocking the 980X made all of your comparisons against the overclocked 1100T meaningless, and if you think people are going to return to read more reviews if you keep it up, think again.
I didn't write this article, evidenced by the 'written by' credits listed in bold on every page. So when you repeatedly tell me what it is I should be doing, you should be careful not to make claims that I'm dodging your points.
Not sure what you mean about addressing complaints anonymously though; my name is right there. And since there are no other Bryan Ischos on the entire planet, I should be pretty easy to find (as you have noticed).
I don't know how you came to the conclusion you came to for (4). I would expect that anyone willing to overclock would be willing to overclock whatever processor they buy so I don't know why you categorize 1100T overclocking differently than 980X. And, I think the real point that should shine through here is that the 980X was the wrong processor to compare the overclocked 1100T to. You should have compared it to an overclocked i7 that is approximately the same price as the 1100T.
In terms of the amount of time it takes to benchmark these processors, I think that is a good reason to have left the 980X out of it altogether. I think that an overclocked i7 of the same cost as the AMD would have been more interesting in place of the 980X.
And - there are no shortages of reviews putting the 1100T through the standard set of benchmarks. I got here from the amdzone.com front page and the list of reviews just for the 1100T was like 20 long. I picked your site's review to read because I like the site name :) My point being, that you can "challenge" me to find an 1100T review that compared against an overclocked 980X, but challenging me to find a review that is tangibly different than yours doesn't really seem to have made your time in making the review worth it. Meaning, if what you're saying is that you don't do anything different than the other guys, then why are you bothering to repeat what has already been done over and over again? Comparing against an oc'd i7 of the same cost would have been different and interesting.
Forget about oranges to oranges: there are so many random influences on overclocking performance, that it is a risky proposition to compare OC to OC results between competing products. Yeah, you can get results, but no two people are going to get the same results, so the individual data is of little value.
The best you can really do, is to compare the one product under test at stock speeds and overclocked. Then you can see how far the reviewer was able to overclock the one processor they have, with the equipment they have, their OC knowledge and experience, and some measure of their risk tolerance. That's about all you can glean from the single-sample testing that every site performs. If you want to go out and get 100 of each processor and test them in a half dozen environments to get a true statistical understanding of the processor's overclocking potential, the world will laud your efforts and your diligence. Oh wait, that's not right. Someone else will be duty-bound to come onto your site and tell you you did it wrong. Oh, well....
Exceptions are out there, sure. The NVIDIA GTX 460 is a good example; they ALL overclocked like mad. Very consistent results were obtained across the population for that one product, but that's not the norm.
For what it's worth, here are the things that I find I most want to see in benchmark reviews that so often are left out or overlooked:
- Comparisons only between processors at similar price points
- Ideally, taking into consideration the overall cost of the processor + motherboard if possible
- Comparisons only of oc'd versus oc'd or non-oc'd versus non-oc'd parts
- Inclusion of compiler benchmarks (i.e. benchmarks of parallell compilations of very large software code bases)
- Investigation into the compiler optimizations used for 'standard' benchmarks. I always get the feeling that many benchmarks that favor Intel highly often do so in part because the compiler that was used to compile the benchmark program favors Intel. I think that the gcc compiler toolchain is fairly agnostic and usually Linux based benchmarks show a different picture than Windows based ones because of this. I really would like to see a site explore this issue further, especially because 99% of the software I run is compiled by gcc and so gcc-compiled code is what I really care about.
A site that answered some or all of these questions would definitely get my loyal readership!
* Which motherboard? There are dozens available for each chipset. X58 motherboards alone range in price from $179 to $699. You seem to think that we have access to every possible processor and motherboard, and we don't: we only have what vendors send us to review.
* I explained before why comparing multiple overclocked processors wasn't feasible given the time frame for the article. As I explained earlier I'm still unconvinced it's ever a good idea.
* Compiler benchmarks? As a programmer, I might find them interesting, but I think most of our readers could care less.
* Investigation into compiler optimizations: Now, this is a good idea, but it's beyond the scope of a simple processor comparison; it needs its own article. While 99% of the software you run may be compiled by gcc, 99% of the software most other people run isn't.
I also include the premise that people generally have a target price range that they are willing to pay for individual components of their system. Now this is a little less solid of a premise because I can certainly see people saying "I'd like to spend $200 on a processor, but if there is a processor for $300 that is twice as fast, I'd seriously consider it". However, if that person was willing to go up to $300 to pay for the twice-as-fast processor, then they'd probably also want to consider all of the other $300 processors available to see if there is an even better processor at that price point. In either case, the most useful comparison to make is for processors in the same price range, because whatever dollar figure a system builder ends up at, they would want to make sure they're getting the best deal at that price; or at the very least, that they're not passing up a much better processor should they be willing to stick with a slightly slower part for other reasons (brand loyalty, already owning compatible components, etc).
Given those two premises, I see the most useful information to someone looking to purchase a processor being comparing similarly overclocked processors (i.e. not overclocked, or overclocked with the same effort and cost put into it), at similar price points.
Now, not every article has to be about this type of value comparison - an article can certainly be about something less practical like "how far do you have to overclock an 1100T - costs be damned - to match the performance of a stock 980X", but this was not that article. This article attempted to make lots of value comparisons while violating both of the premises I set forth previously.
Hence, my conclusion that the comparisons were meaningless (for someone who is actually considering buying either processor). If you accept my premises, they were. If you are just looking for comparisons in the abstract without any value considerations, then the article was fine - except that in that case it wasted too much verbage actually making value comparisons that were not the point of the article.
We can't please everyone, and with the time constraints for most projects we aren't afforded the opportunity to dig as deep as we'd like.
If x, then meaningless. Meaningless means "has no meaning." Even if we take your premise as true, people may still get useful information out of the article, and thus your premises don't follow directly to the conclusion, and you have an invalid argument.
I along with others are still awaiting a reply, to my comments/questions, on your review of the 1075T.
Could you please spare some time to answer my queries there, thank you.
Ty,From Turkey..