Intel Core i7-3770K Ivy Bridge CPU Review

Manufacturer: Intel Corporation
Product Name: Desktop Processor
Model Number: Core i7-3770K
Price: $313.00 MSRP

Full Disclosure: Intel provided the product sample used in this article.

Barely a year after the introduction of its immensely popular Sandy Bridge CPUs, Intel brings us Ivy Bridge. It's been a busy 15 months for Intel, with the introduction of no fewer than three chipset families (P67, Z68, and Z77, and many variants thereof) and three processor families (Sandy Bridge, Sandy Bridge Extreme, and Ivy Bridge).

As it has for the past several years, Intel's sticking to their "tick-tock" processor release cycle. At roughly yearly intervals, Intel will introduced either an entire new processor architecture ("tock") or a refinement thereof ("tick"). Sandy Bridge was the tock, and Ivy Bridge is the tick. The "tick" processors are usually minor upgrades in terms of capability. For Ivy Bridge, we have the power and heat advantages of Intel's new 22nm process (as compared to Sandy Bridge's 32nm), and the improved efficiency afforded by their low-leakage "3D" transistors.

Of course, what we all want to know is how the performance and features of this new CPU compare with the Sandy Bridge line. To determine this I'll be testing the Ivy Bridge Core i7-3770K CPU against a Sandy Bridge Core i7-2600K.

Intel Core i7-3770K Features

The following information is courtesy of Intel

• 4 Cores, 8 Threads
• Intel Turbo Boost Technology 2.0
• Intel Hyper-Threading Technology
• Supports LGA1155 socket Intel X77/Z68/P67 Express Chipset-based motherboards
• Up to 8 MB Intel Smart Cache
• 2 channels of DDR3 1600 MHz
• 16 PCI Express Gen 3 Lanes
• Intel® Advanced Vector Extensions (Intel® AVX)
• Advanced Encryption Standard New Instructions (AES-NI) PCLMULQDQ Instruction
• RDRAND instruction for random number generation

Intel (SKU) Specifications

Intel's lineup of Ivy Bridge processors includes three mobile processors, five desktop processors, and four low-power desktop processors. The chart below summarizes the specifications of the five desktop CPUs, along with the Sandy Bridge-based Core i7-2660K I'll be using as a comparison.

The two desktop processors enthusiasts will find most interesting are the one we're testing, the Core i7-3770K, and the Core i5-3570K. The 3570K is a trivial 100MHz lower on the core and turbo clock speeds, and has 2M less cache, but it's still a "K"-series unlocked CPU and at $100 less than the 3770K is probably the sweet spot of the Ivy Bridge desktop line. It doesn't have Hyper-Threading, and thus presents only four cores instead of eight to the operating system, so it's the obvious replacement for the Core i5-2500K.

The 22nm process enables Intel to drop the TDP from the 2600K's 95 watts down to 77 watts, a substantial reduction. The Intel Graphics HD 4000 iGPU now has 16 "graphics execution units", up from 12 in the 2600K, but Intel says the performance should be substantially improved over that of the HD 3000. If you're a corporate IT type, you'll be interested in the i7-3770 and i5-3550, which support Intel VPro Technology, a suite of hardware security features. All new Ivy Bridge desktop CPUs have 16 PCI Express Gen 3 lanes, which provide twice the bandwidth per lane of the PCI Express Gen 2 lanes used on Sandy Bridge CPUs.

Since the Ivy Bridge desktop CPUs use the same LGA1155 package as Sandy Bridge, the chips look physically identical. Unless you read the model number you won't know which one you've got.

Let's take a look at Ivy Bridge architecture and its supporting Z77 Express chipset in the next section.

Ivy Bridge and the Z77 Express

Ivy Bridge architecture is pretty much Sandy Bridge: the fact that this CPU family is a "tick" means that Intel's more interested in process refinement rather than big architectural changes. Still, there are a few new items, and here's what Intel would like you to know about:

• Intel Rapid Start technology uses an SSD to hold system state information, enabling very rapid wake-up from a power-down state. Think of it as hibernation on steroids. This is something you'll see on Ivy Bridge-based laptops and ultrabooks.
• Intel HD4000 graphics offer up to double the performance of HD3000, as well as improved Quick Sync transcoding and 3D performance.
• The 16 PCI-E lanes provided by the CPU are now Gen 3 instead of Gen 2; this means twice the bandwidth per lane.
• Integrated USB 3.0 (four ports) via the Z77 chipset.
• Hardware based random number generation for encryption and security applications.
• Improved overclocking features with higher CPU multiplier limits, finer DDR3 frequency control, and XMP 1.3 support.

As far as actual performance improvements go, Intel's test results show raw compute performance in the range of 3%-9% better than a Core i7-2700K in most benchmarks. Intel says the biggest performance improvement will be with graphics and transcoding using the HD4000 integrated GPU, when compared to the previous-generation Intel HD3000 GPU built into the Sandy Bridge processors. The video transcoding should be a nice win if you have a program that uses this feature, but even double the performance of the HD3000 isn't anything that will excite most gamers, as we'll see in the graphics benchmarks later in this review.

While the desktop Ivy Bridge CPUs don't seem to have many real advantages over Sandy Bridge, in the ultrabook catagory we're going to see some action. Ultrabooks are hot, and Ivy Bridge mobile CPUs will bring significant power savings, which is a real win when you're running on a battery. Apple is expected to revamp its entire Macbook line to Ivy Bridge, while bringing all models to Air-like thinness, and just about every PC vendor has new Ivy Bridge-based ultrabooks coming out as well.

Intel's official "die shot" of the Ivy Bridge CPU looks almost identical to the one they showed for Sandy Bridge. Sharp-eyed observers will notice that the Processor Graphics section looks different, with what appear to be eight "cores" instead of the six visible in the Sandy Bridge die shot. This works out well with the increase in functional units from 12 in Sandy Bridge to 16 in Ivy Bridge, assuming two "functional units" per core.

Intel has introduced ten new chipsets (collectively known as Panther Point) to support their twelve new Ivy Bridge CPUs. We're interested in the top-end desktop offering, the Z77 Express. If you think the diagram below looks exactly like that for the Z68 Express chipset, you're forgiven: the differences are few and subtle. Comparing the two charts, the differences are:

• Some PCI-E lanes for Thunderbolt support
• Four native USB 3.0 ports. Finally.
• Official support of DDR3-1600
• Intel Rapid Start Technology (not directly on the chart, but included in "Responsiveness Technologies")

You might be disappointed that there are only four USB 3.0 ports, and still only two SATA 6G ports. Intel says the native USB 3.0 ports will provide better performance than provided by third-party USB 3.0 controllers, and there's a subtle extra advantage in that the on-chip ports don't require any PCI-E lanes to implement. Still, I had hoped for more from this upgrade, as I did from the X79 Express chipset.

Disappointments with the chipset aside, let's get to testing this setup.

Processor Testing Methodology

Since Ivy Bridge is the successor to Sandy Bridge, it makes sense to directly compare the two CPU families. For this review, I compared the Core i7-3770K against the Sandy Bridge based Core i7-2600K. While the 2600K is one tick below the top-end 2700K, the only difference is a fractional increase in clock speed: a base clock of 3.5GHz and a turbo of 3.9GHz as compared to the 2600K's 3.4/3.8GHz clocks. That less than a 3% boost so the benchmark results you see here should be pretty much the same as you'd see with a 2700K.

I tested the Core i7-3770K processor at both its stock clock speed and the maximum stable overclock I could attain on all cores simultaneously, which was 4.7GHz. I'll have more on this in the Overclocking section.

Each CPU was tested on an ASUS P8Z77-V Deluxe motherboard with the same hard disk, memory, CPU cooler, and graphics card, so that the CPU was the only variable.

For a graphics card, I used a reference design AMD Radeon HD5770 running at stock clock speeds. Since the Sandy Bridge and Ivy Bridge integrated GPUs share L3 cache and memory bandwidth with the processor cores, the iGPU was disabled for all benchmarks except those specifically concerned with the iGPU.

Intel Z77 Express Test Platform

• Motherboard: ASUS P8Z77-V Deluxe with BIOS 0906
• Processor: 3.5GHz Intel Core i7-3770K and 3.4GHz Intel Core i7-2600K
• System Memory: 8GB DDR3-1600 (two 4GB DIMMs) at 9-9-9-27
• Primary Drive: Seagate ST3500 500GB drive
• Graphics Adapter: AMD Radeon HD5770
• CPU cooler: Thermalright Silver Arrow

Benchmark Applications

• Operating System: Windows 7 Home Premium 64-Bit
• SiSoft Sandra Lite 2011.SPS (1780)
• AIDA64 Extreme Edition v2.30.1900
• Futuremark PCMark 7
• Futuremark 3DMark 11 (for DX11 testing)
• Futuremark 3DMark Vantage (for DX10 testing)
• Aliens vs. Predator benchmark (for DX11 testing)
• Unigine Heaven 3.0
• Maxon CINEBENCH R11.5 64-Bit
• Street Fighter IV benchmark
• PassMark PerformanceTest 7.0b1021
• x264Bench HD 3.0, including AMD-supplied variants using new FX instructions
• SPECviewperf-11:
  • Lightwave 9.6
  • Autodesk Maya 2009
  • Siemens Teamcenter Visualization Mockup
• SPECapc LightWave 3D v9.6
• Handbrake 0.96 video transcoding
• Blender 3D rendering
• POV-Ray 3D rendering

Let's start the benchmarking with an AIDA64 run in the next section.

AIDA64 Extreme Edition Tests

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 little 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.

The 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.

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, so quad-core processors with Hyper-Threading have no real advantage. The AIDIA64 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

In the Queen test, the 3770K turns in a score only 3.5% higher than the 3600K, and 5.5% higher in the PhotoWorxx test (at stock speeds).

In the ZLIB test, the Ivy Bridge is 6.8% better, while in the Hash test it's 14.4% better.

In the AES test it's only 3.3% better. Oddly the Ivy Bridge score drops slightly in this test when it's overclocked. We've seen the Sandy Bridge's insensitivity to overclock in this test in our reviews of the Core i7-2600K and the Core i7-3960X Extreme CPUs. AMD CPUs, on the other hand, dramatically improve their scores in this test when overclocked.

Let's move on to the PCMark 7benchmark.

PCMark 7 Tests

PCMark 7 is Futuremark's successor to PCMark Vantage. The full suite of tests comprises seven different sequences with more than 25 sub-tests that exercise your system's abilities in storage, computation, image and video manipulation, web browsing and gaming. It was developed with input from the designers, engineers and product managers at AMD, Compal, Dell, Hitachi GST, HP, Intel, NVIDIA, Samsung, Seagate, Western Digital and many other well-known companies.

For this benchmark I chose the PCMark test, which provides a number indicating total system performance, as well as the Productivity, Creativity, and Computation test suites.

Productivity Test

The Productivity test is a collection of workloads that measure system performance in typical productivity scenarios. Individual workloads include loading web pages and using home office applications. At the end of the benchmark run the system is given a Productivity test score. The Productivity test consists of:

• Storage
• Windows Defender
• Starting applications
• Web browsing and decrypting
• Productivity
• Data decryption
• Text editing

Creativity Test

The Creativity test contains a collection of workloads to measure the system performance in typical creativity scenarios. Individual tests include viewing, editing, transcoding and storing photos and videos. At the end of the benchmark run the system is given a Creativity test score.

• Storage
• importing pictures
• video editing
• Image manipulation
• Video transcoding - high quality

Computation Test

The Computation test contains a collection of workloads that isolate the computation performance of the system. At the end of the benchmark run the system is given a Computation test score.

• Video transcoding - downscaling
• Video transcoding - high quality
• Image manipulation

It's important to note that since PCMark 7 was designed as a system test, the scores are dependent on the configuration of the entire system being tested, including things like the memory, hard disk, and graphics cards used: it's not an isolated CPU test like most of the other benchmarks I'm using in this review. However, since all other hardware (motherboard, video card, memory, hard disk, etc.) was identical, with only the CPUs being changed, any performance differences here can be attributed to differences in CPU performance.

At stock clocks, the 3770K has a 10% advantage in PCMark and Productivity, 5% in Creativity, and just under 1% in Computation. The overclock didn't help much except in Computation, where it boosted the 3770K's score by 24%.

In the next section I run everyone's favorite benchmark: CINEBENCH!

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 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 photo-realistic 3D scene, "No Keyframes", the viral animation by AixSponza. This scene makes use of various algorithms to stress all available processor cores, and all 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.

I'm particulaly fond of CINEBENCH's multi-core rendering test, which highlights the performance advantages of more cores and Hyper Threading in an easy-to-appreciate visual manner. In multi-core rendering, the 3770K ekes out a mere 6% win, but its lead increases to 27% when overclocked. In the single-core rendering test, the percentages are 7.7% and 28.4%, respectively.

CPU-Dependent 3D Gaming

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 multi-threaded, 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 its lowest resolution (640x480) will all graphical features turned down to the minimum possible settings. This makes the video card much less of a factor in the results, biasing towards processor performance. We see the same pattern we've seen in other tests, with the new Ivy Bridge CPU turning in a stock-clocked score just under 10% better than the stock-clocked 2600K.

PassMark PerformanceTest 7.0

The 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. The CPU tests comprise a number of different metrics. The first three I'll look at are integer performance, floating point performance, and a benchmark that finds prime numbers.

In the integer performance test, there's virtually no advantage to the 3770K, but it posts surprisingly strong results in the floating point test, beating the 2600K by an amazing 67%!

SSE stands for "Streaming SIMD Extensions", and are instructions that handle multiple chuncks of data per instruction (SIMD = Single Instruction Multiple Data). SSE instructions work on single-precision floating point data and are typically used in graphical computations. Again, the 3770K is no better than the 2600K at stock clock speeds, but improves dramatically when overclocked.

Interestingly, the Ivy Bridge CPU posts slightly lower scores in the Compression and String benchmarks than does the older Sandy Bridge CPU.

Handbrake Media Encoding

It's a truism that consumer-level computer performance reached the "fast enough" point years ago, where increases in system performance don't make thing any faster for most people. Web browsing, e-mail, word processing, and even most games won't benefit dramatically from a super-fast CPU. There are some exceptions, though, and media encoding is one of them: transcoding video, especially high-definition video, can bring the strongest system to its knees. Fortunately, media transcoding is one of those things that (depending on the design of the code, of course) that scales really well with both clock speed and the number of cores, so the more you have of both, the better your results will be.

The free and open-source Handbrake 0.96 video transcoder is an example of a program that makes full use of the computational resources available. For this test I used Handbrake 0.95 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.

The Core i7-3770K comes in at 7% better than the 2600K at stock clocks, and improves its score by another 16% when overclocked.

x264 HD Benchmark 3.19

Tech 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.

Posting virtually rhe same scores as the 2600K at stock clock speeds, the 3770K does very well when overclocked.

Again, we see the 3770K beaten at stock clock speeds in three of the four runs by its older sibling. Weird.

HD4000: Quick Sync Transcoding

While gamers sneer (justifiably) at the performance of integrated GPUs in general and Intel's "HD Graphics" in particular, Sandy Bridge CPUs introduced a new feature: Quick Sync video transcoding. Portions of the iGPU were optimized for video transcoding tasks, and programs that took advantage of this feature could transcode video faster than all but the very best discrete video cards.

Quick Sync was slow to take off, since the original P67 desktop chipset didn't support it: the iGPU of these new processors was disabled in P67 based desktop motherboards. If you had a system based on the H67 chipset, you could use Quick Sync, but were prevented from overclocking your system. Intel finally removed these artificial restrictions with the Z68 chipset, which allows systems to both overclock and use the integrated GPU. Lucid Logix' Virtu GPU virtualization software added the cherry to the sundae by enabling users to fine-tune how the Sandy Bridge iGPU was used in conjunction with a separate graphics card.

With the Ivy Bridge CPUs, Intel has moved from the HD 3000 graphics to HD 4000. The specifications of this new iGPU are spotty: we know that it has 16 "execution units" as opposed to the HD 30f00's 12, but not much more than that. Still Intel says users will see a dramatic performance increase, so I decided to check out Quick Sync performance on the 3770K to see how it compared.

This chart shows the results from transcoding a video (the same Family Guy episode I use for my Handbrake testing) to the "iPad" presets in Arcsoft Media Converter. The output video is an MP4 at 1280x720 (720p) resolution with a bit rate of 4Mbps. I ran the transcode five times: first using the CPU coding of the 2600K and 3770K, then using the Radeon HD5770 card, then using the iGPUs of the 2600K and 3770K. It is important to note that each "code path" through the transcoding process is different and that the output video for each version, although it adheres to the settings made in the program, is not pixel-per-pixel identical.

In this test, at least, we see that although Quick Sync is dramatically faster than pure CPU-based transcoding and the extra boost provided by the HD5770 card, the HD4000 iGPU is not significantly faster than the HD3000. The difference is a mere 13%.

But maybe we'll see more impressive results from video performance tests in the next section.

HD4000: DX10 Performance

The HD 4000 iGPU built into the Ivy Bridge processors is the first Intel integrated GPU to support DX11. The Sandy Bridge's HD3000 iGPU only supported up to DX10.1, so for my first graphics tests I stuck with DX10 software. I compared the HD3000, HD4000, and Radeon HD5770 GPUs using Unigine's Heaven 3.0 benchmark set to DX10 graphics, with shaders set to "medium", no anti-aliasing, 4x anisotropic filtering, and a resolution of 1680x1050. This benchmark takes you on a fly-through exploration of a beautifully-rendered city floating in the clouds.

I also used Futuremark's PCMark Vantage, selecting the "Jane Nash" and "New Calico" benchmarks. The former tracks a 60s-era female spy as she's discovered in the evil villian's lair and must run for her life, stealing a jet-powered flying boat to escape, while the latter shows a giant space battleship hovering outside an asteroid belt, dispatching a fleet of space fighters which mercilessly bombard the planet below. I set PCMark Vantage to the "Entry Level" presets with "optimal textures" and the same 1680x1050 resolution used in the Heaven benchmark.

As these results show, neither iGPU is suitable for serious gaming at this resolution. And in the case of PCMark Vantage, "entry level" settings mean precisely that: water is a static, unmoving texture, even as boats plow through it; cloth doesn't drape, and so forth...it's a pretty bare-bones experience. While the HD4000 does substantially better than the HD3000 in these tests, outperforming its older incarnation by 45%, 70%, and 49%, respectively, it's still nothing that any gamer would consider even close to playable. Even casual gamers would be well advised to pony up the $100 or so for a Radeon HD5770 video card, or something else in that class, as it makes the difference between playability and non-playability.

HD4000: DX11 Performance

AMD's Fusion APUs have offered DX11 support for about a year now, and their graphics performance of this pair has been surprisingly good (sadly I did not have a Fusion APU available to test against). The HD4000 iGPU built into the Intel Core i7-3770K, however, is Intel's first DX11 offering. Since the HD3000 doesn't support DX11, the only thing I could compare it against was the Radeon HD5770 video card.

For the DX11 testing I used the Heaven 3.0 benchmark again, except set to DX11 and with moderate tessellation (other settings remained the same). The Aliens vs. Predator was set to medium textures, low shadows, 8x anisotropic filtering, no anti-aliasing, SSAO and shadow sampling off, and hardware tessellation, also at 1680x1050.

Here we see the HD4000 making a surprisingly good showing against the Radeon HD5770, 27% slower in the Heaven benchmark and 34% slower in Aliens vs. Predator. Next, I fired up Futuremark's 3DMark 11 benchmark and ran the four basic graphics tests. GT1 and GT2 render underwater scenes with miniature submarines moving among the elements of an undersea facility, with multiple light sources playing through the cloudy water, while GT3 and GT4 render very complex jungle scenes with many shadows and tessellated textures. For this test I selected the "Performance Level" presets and set the resolution to the same 1680x1050 I've used on the other tests.

Neither the HD4000 nor the HD5770 do very well here, but the Radeon is much faster than the HD4000 is this more strenuous test.

The best thing I can say about the HD4000 is that it's definitely quicker, actually much quicker, than the HD3000, but it still falls far short of usability for any type of 3D action gaming.

SPECviewperf 11 tests

The 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 modeling, 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. Results are reported as abstract scores, with higher being better.

Lightwave

The 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.

Maya

The 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 Mockup

The 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 SPECviewperf suite is a good example of a real-world test of applications that would normally be the province of a high-end workstation: the individual tests comprise code and models from real applications, running scripts that do real work. Again we see that Ivy Bridge performance improvements over Sandy Bridge are quite marginal.

SPECapc Lightwave

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 and 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.

I've found the SPECapc Lightwave 3D test to be an excellent indicator of overclock stability. In many cases, overclocked systems that will make it through every other benchmark here will crash in this test.

Bear in mind that what this benchmark does is use scripts to control a stand-alone instance of Lightwave, so in that sense it's more indicative of real-world performance than the embedded Lightwave code in SPECviewperf. Again the performance difference between the two CPUs is almost nonexistent at stock clock speeds.

Blender

Blender is an open-source, free content creation suite of 3D modeling, rendering, and animation capabilities. Originally released in 2002, it's available in versions for Mac OS X, Windows, Linux, and several Unix distributions. It supports rigid and soft-body objects and can handle the draping and animation of cloth, as well as the rendering and animation of smoke, water, and general particle handling.

Our Blender test renders multiple frames of an animation of a rotating chunk of ice, with translucency and reflections. Rendering of this model uses ray-tracing algorithms and the program reports the rendering time for each of the animation's 25 frames. The results are a summation of the rendering times for all frames and the lower the score, the better.

Blender is limited to a maximum of 8 threads, so these four-core Hyper-Threaded CPUs are a good match for it. Overclocking the Core i7-3770K provides about a 15% gain in performance.

POV-Ray

The Persistence of Vision ray tracer is a free, open source 3D modeling program that uses ray-tracing algorithms to generate realistic three dimensional images. Ray tracing is very computationally intensive, and the POV-Ray program has a handy built-in benchmark to let you check the performance of your system. AMD wins this round, posting stock-clocked results that are 26% better than Intel, although the FX-8150's lead narrows to a mere percentage point when both processors are overclocked.

Not much to say here. The Core i7-2600K and Core i7-3770K are virtually neck and neck, their scores separated by about 3%. Overclocking provides a 21% performance advantage.

Memory Bandwidth

Since the Core i7-3770K uses the same architecture as Sandy Bridge, the memory controller is probably the same. But nothing reveals the truth like a good memory bandwidth benchmark, in this case SiSoft Sandra Lite. Remember, this test uses the same memory and the same motherboard; only the CPUs have changed.

Perhaps unsurprisingly, the memory bandwidth scores are all within 3% of each other, which is within the margin of error for this test. Intel says that Ivy Bridge does allow the potential for much higher memory frequencies than Sandy Bridge, although as we've shown in our memory reviews, the real-world performance impact of high speed memory is minimal.

Core i7-3770K Overclocking

I admit I came to this section of the review with high hopes: Sandy Bridge CPUs are almost insanely overclockable, with late steppings of the 2600K and 2700K reaching 5GHz on good air cooling in many cases. Surely Ivy Bridge, with its lower voltage, 22nm process, and low-leakage transistors, would be even better! Since this is a "K"-model CPU, overclocking is a simple as bumping the multiplier and adjusting the voltage as needed. As with Sandy Bridge, you can set different turbo limits that will be used when 1, 2, 3, or all 4 cores are loaded; I prefer to try to hit the highest stable overclock I can with all four cores loaded.

If you were expecting the same thing I was, prepare to be disappointed. Even with a Thermalright Silver Arrow cooler, the best air cooler I've ever tested, the highest overclock I was able to reach on all cores under load was only 4.7GHz. I didn't have to increase the core voltage to achieve this.

At 4.8GHz, the system would boot into Windows and run most benchmarks, but even though I didn't increase the voltage, individual core temperatures spiked past 100 degrees Celcius and the CPU would throttle itself until temperatures came back down, so 4.7GHz was the best I could do. The 21% overclock I achieved was completely stable, although some core temperatures would still reach into the 90s, and it's not a configuration I'd suggest running for long periods of time, as those temperatures (while apparently below the throttling limit) are still quite high.

The chart below compares the normalized stock benchmark scores with the overclocked benchmarked scores.

An 18% average improvement from a 21% overclock is pretty good. However, it's likely that a late-stepping 2600K or 2700K could be overclocked to beat the overclocked 3770K, even if by only a small margin. We can hope that later steppings of the Ivy Bridge processor will yield better overclocks as Intel refines the 22nm fabrication process.

Ivy Bridge Final Thoughts

As a "tick"-cycle CPU, the Intel Core i7-3770K represents a refinement of the Sandy Bridge architecture rather than a whole new direction, and if our performance expectations were too high, perhaps it's Intel's fault for bragging about their 3D transistors and 22nm production process since May of 2011. Intel's said that these process advancements can be tuned to provide more performance in the same power envelope, or lower power in the same performance envelope, and with the initial Ivy Bridge CPUs, it's obvious they've decided that the latter is the way to go...at least for the time being. The performance difference between the stock-clocked 2600K and 3770K CPUs was less than 5% when averaged across all these benchmarks, as the table below shows:

Four percent isn't too exciting, and bear in mind this was in comparison with a Core i7-2600K rather than the incrementally faster 2700K. It's also likely that when overclocking, a current Sandy Bridge CPU would actually be faster.

Video performance fared better, although we can only directly compare DX10 performance and transcoding:

Just looking at gaming performance improvement, it's 55%, so HD4000 really does offer a significant performance boost to gamers...but even so, a three-year-old midrange video card like the Radeon HD5770 trounces it completely. Although impressive from a percentage improvement point of view, HD4000 graphics remains irrelevant to most gamers, Intel's promises of "Great Mainstream 3D Gaming" notwithstanding.

While Ivy Bridge power savings won't mean much for us desktop users-- a Core i7-3770K running at full load 24/7 for a year would save its user $20.51 compared to a Core i7-2700K assuming 13 cents per kilowatt-hour-- but that's a "desktop centric" point of view. Laptop computers passed desktop computers in unit sales in 2005, and the tablet/smartphone trend has taken an even bigger chunk out of the desktop market. The fact is that portable devices and servers in data centers drive the CPU market these days, and that's why Intel has emphasized low power and thermals over performance in Ivy Bridge. Ivy Bridge CPUs will let your spiify new Ultrabook run longer and save large companies huge amounts of money on data center power costs.

So while Ivy Bridge desktop CPUs are only an incremental advance over Sandy Bridge, the desktop isn't the market they were designed for. Still, consider that Ivy Bridge CPUs are still faster and cheaper than Sandy Bridge, which was already an amazingly fast processor. The tiny fraction of desktop users who need more computational power than Ivy Bridge can provide are already using either Sandy Bridge Extreme or multi-CPU Xeon systems.

I really would have liked to see more PCI-E lanes, but it's obvious that Intel's decided that 16 (CPU) + 8 (chipset) is all this market segment needs, and folks needing more can simply move to Socket LGA2011, or perhaps an AMD 990FX system. Still, vendors are stepping up to the plate with PLX-enabled motherboards that can compensate to some extent for the dearth of lanes.

Matched with a Cougar Point motherboard, Ivy Bridge considered as a package offers a nice set of advantages over Sandy Bridge: slightly better performance, better graphics, lower power draw, more versatile PCI-E allocation, native USB 3.0 ports, and so on. The fact that the Core i7-3770K's MSRP is slightly below existing Sandy Bridge CPUs is a bonus. If you're building a new system, there's no reason to go with a Sandy Bridge/Z68 setup unless you can find one at a good price. Conversely, if you already have a 2500K or higher Sandy Bridge system, there's no real reason to upgrade to Ivy Bridge.

Intel Core i7-3770K Conclusion

Ivy Bridge is finally here, and if its performance improvements over Sandy Bridge are underwhelming, remember that Sandy Bridge set a very high bar. Without Sandy Bridge to compare it to, we'd be lauding Ivy Bridge performance and overclocking to the skies.

Ivy Bridge was perhaps one of Intel's worst-kept secrets, what with their bragging about their 22nm, "3D" transistors for the last few months. Unknown production delays led to Z77 Express based motherboards being available for weeks before the CPUs they were designed for, and that didn't help things either.

Ivy Bridge represents a better value than Sandy Bridge, since it's both cheaper and faster. We can complain that the magnitude of these improvements may be rather small, but the truth is that Sandy Bridge CPUs represented such a huge advance over the previous generation Intel and current AMD processors that there's no way Ivy Bridge could have been that much better. Also, keep in mind that new mainstream CPUs simply aren't designed with desktop systems as a priority any more. Ivy Bridge's biggest new feature is its improved performance per watt.

Intel says the Core i7-3770K processor will sell for $312...but that's Intel's price for a tray of 1,000 CPUs. Still, Intel's "1000 piece" price typically predicts retail prices pretty well, although you can expect some early price-gouging for the individual CPUs from the usual suspects, as we saw with AMD's Bulldozer CPUs and Intel's own Sandy Bridge Extreme when they were introduced.

The real impact of the Ivy Bridge architecture will be felt in the mobile device and server market, but once the price settles down, this top-end Ivy Bridge processor will be a very good value, since it's both faster and cheaper than its Sandy Bridge forebear. The only weakness in this CPU compared to Sandy Bridge is its overclocking performance, which will hopefully improve...but right now you can probably get a little more performance from an overclocked Sandy Bridge CPU, and that's just enough to keep the Core i7-3770K from taking the gold.

Pros:

+ Lower power than Sandy Bridge, although this isn't significant for desktop platforms
+ Faster and cheaper (if not by much) than Sandy Bridge CPUs
+ Intel HD4000 iGPU significantly faster than HD3000
+ Can be used in Z68-series motherboards (with vendor BIOS support)
+ Still the best performance in a mainstream consumer CPU

Cons:

Ratings:

• Performance: 9.5
• Construction: 9.00
• Overclock: 8.0
• Functionality: 8.50
• Value: 9.0

Final Score: 8.8 out of 10.

Quality Recognition: Benchmark Reviews Silver Tachometer Award.

Benchmark Reviews invites you to leave constructive feedback below, or ask questions in our Discussion Forum.

Related Articles:

• 4TB Seagate Desktop HDD ST4000DM000
• SilverStone SST-EC03 USB 3.0 PCI-E Card
• OCZ RevoDrive 3 X2 PCI-E Solid State Drive
• OCZ Core Series SATA-II SSD OCZSSD2-1C64G
• Hiper HPU-5K880 and HPU-5B680 Type R Mk II PSU's
• Intel Core i7-870 Processor BX80605I7870
• Logitech G600 MMO Gaming Mouse
• HIS Radeon HD6870 IceQ-X Turbo-X Video Card
• G.SKILL 2133MHz DDR3 Pi-Series Memory Kit
• Radeon HD 5770 CrossFireX Performance Scaling