Intel Core i7-3820 Sandy Bridge-E CPU Review

[MAC]

When Intel’s high-end LGA2011 platform was launched in November, we were very impressed with what the company had managed to create. However, we lamented the fact that only the pricey Core i7-3930K and even pricier Core i7-3960X models were scheduled for imminent release. The only information we had regarding the availability of the more budget-friendly Core i7-3820 was sometime in Q1 2012.

Now that launch was a little over 3 months ago, and little did we know at the time but the i7-3930K has had really spotty availability thus far. This lack of affordable processors has obviously hindered the mass market appeal of the platform as a whole. Thankfully, this situation should change shortly. The Core i7-3820 has finally been released, and it is starting to show up in the retail channel...albeit at a glacial pace.

This is good news, because while we are a fan of the two high-end six-core models we always knew that what would make or break the LGA2011 platform is the i7-3820, the spiritual successor to the widely successful i7-920. While both chips share a quad-core/eight-thread design and an attractive sub-$300 price tag, the similarities end there. The i7-3820 is a 32nm part with a 3.6GHz default clock and Turbo capabilities up to 3.9GHz. It features 10MB of L3 cache and a new beefed up quad-channel DDR3-1600 memory interface which is theoretically capable of 51.2GB/s of bandwidth. The integrated PCI-E controller has also been revamped and in coordination with the new X79 Express chipset can supply up to 40 PCI-E 2.0 lanes (with PCI-E 3.0 compatibility) to the PEG slots. While this chip is not fully unlocked like the i7-3930K and i7-3960X are, it offers more potential frequency headroom then you will ever require using air or water-cooling. Furthermore, since this platform was fundamentally designed with overclocking in mind, you will be able to tap into that extra performance easier than ever before.

So if you are in the market for a new brand new system or just a sizeable upgrade, and are willing to give the LGA2011 platform a fresh look, this new processor is definitely a good place to start.

 

[MAC]

Sandy Bridge-E: Intel Core i7-3820

Sandy Bridge-E: Intel Core i7-3820

Sandy Bridge E / Sandy Bridge / Gulftown / Zambezi - Click on image to enlarge​

As you should know by now, Sandy Bridge-E’s naming scheme is greatly simplified when compared to other processor families. While the Sandy Bridge LGA1155 models are part of the Core i3/i5/i7 2000 series, which consists of numerous models with different core counts, threads counts, IGP types, and supported technologies. It’s a bit of a mess. Thankfully, Sandy Bridge-E (SB-E) has been given the 3000 series moniker, and they are all Core i7 models, which simply put means that they all support Hyper-Threading (HT) and Turbo Boost technology. As you will see below, they don’t all have the same core/thread count though.

As we stated in the intro, arguably the most interesting chip of the series is the Core i7-3820. Intel is clever, and with this model number they are clearly hoping to evoke the undying love that most enthusiasts had and continue to have for the i7-920. As illustrated above, the Core i7-3820 is a 32nm quad-core/eight-thread processor with a 3.6GHz default clock and a Turbo Boost mode that tops out at 3.9GHz in single and dual-threaded workloads. This default clock speed advantage over its higher-end siblings is a little bit deceptive though, since during our time with the i7-3960X we never saw it dip below 3.6GHz, as long as Turbo Boost was enabled. Accompanying its four cores is 10MB of L3 cache, an impressive 2.5MB per core, and new beefed up memory controller that features a quad-channel DDR3-1600 interface which is theoretically capable of 51.2GB/s of bandwidth.

Core count aside, one of the main differences is that this model is only partially unlocked. What this means is that there are supposed to be 6 bins available above the highest Turbo Boost mode, which would mean that the CPU multiplier is capped at 45X. However, during our time with the processor, and on two different motherboards, 43X was the highest selectable mulitplier. While this might sound terrible, it is not really a big deal and here’s why: 43 X 100MHz (the stock base clock) x 1.25 (gear ratio) equals almost 5.4GHz. In all likelyhood, this multiplier cap is not going to affect anyone’s overclocking fun...unless they are dabbling with sub-zero cooling or yields improve substantially.

On a side note, the Core i7-3820 is actually based on a native quad-core die. Intel's modular die design allows it to easily add or substract parts, so it is quite easy for them to manufacture a seperate die instead of simply using six-core rejects. The fact that the i7-3820 is based on the different M1 stepping instead of C1 or C2 like the higher-end models also lends some credence to this assertion.

Click on image to enlarge​

Although our i7-3820 is obviously an engineering sample, it is fundamentally the same as any retail boxed chip that you can buy at your favourite e-tailer, except for what's laser-etched onto the integrated heatspreader.
If you thought LGA1366 processors were big, wait until you have one of these in your hand. These new SB-E chips are downright enormous, which is understandable given the fact the package has to fit 2011 contact points.

Based on the digits on the heatspreader, we can determine that this sample was manufactured in the 39th week of 2011, which is just four weeks later than our i7-3960X sample. By the way, notice that little black dot on the bottom-right corner of the HIS? Intel has drilled holes in all the SB-E chips in order for enthusiasts to place a thermal probe in there. This is further proof that Intel really geared the LGA2011 platform towards overclockers.

Click on image to enlarge​

As mentioned above, we never saw the default 3.3Ghz clock speed. At idle, the chip would drop down to 1.2GHz, and under load it would alternate between 3.6-3.7-3.8GHz depending on the workload. It would very sparingly hit 3.9GHz even in single or dual-threaded workloads, and it was not just a bios issue since we experienced the same situation on two different motherboards. We are happy to report that the new version of CPU-Z finally supports Sandy Bridge-E, and as you can see in the screenshots the core voltage ranged from 0.864V to 1.260V, which is a bit lower than what we saw with our i7-3960X.

 

[MAC]

Feature Test: Turbo Boost 2.0

Feature Test: Turbo Boost 2.0

Since we don't have a fancy Intel-provided graphic to demonstrate how Turbo Boost functions on the i7-3820, here is a wordy explanation of how it works. Turbo Boost is a performance enhancing feature that automatically unlocks additional speed bins (multipliers) and allows the processor to self-overclock based on thermal conditions and workload. For example, if the Power Control Unit (PCU) senses that only one core is active and the other three are in an idle state, it will use the unused power and thermal headroom to overclock that single active core to ensure superior single-threaded performance. Conversely, if you are running a multi-threaded application, the PCU will measure the thermal headroom and if the processor is running cool enough it will overclock all six cores. On the Core i7-3820 processors, Turbo Boost can provide a 300MHz frequency boost when 1 is loaded, 200MHz when two cores are in use, and 100MHz in applications that utilize 3 or 4 cores.

Turbo Boost Off - Click on image to enlarge - Turbo Boost On​

Although the i7-3820's version of Turbo Boost is nowhere as aggressive as on the i7-3960X, it really doesn't need to be. With a very high default clock of 3.6GHz, there wasn't much headroom for many additional Turbo Boost bins. Furthermore, we don't think Intel wanted it's baby SB-E chip to overshadow the flagship i7-3960X in any area.

It should be noted that although the i7-3820 is capable of hitting 3.9GHz in single-threaded workloads, we almost never saw it, whereas it was consistently achieved on the flagship six-core part. We noticed this same behaviour on two seperate motherboards with the latest available bioses, so it's a little unusual.

To check out the performance gains that Turbo Boost can provide on this part, we selected a nice mix of benchmarks with both light and multi-threaded workloads.

As you can see, the overall performance gains are quite minimal no matter what type of workload. Having said that, the i7-3820 does have a high default clock speed, and we would always choose that over a more aggressive form of Turbo Boost and a lower default clock speed.

 

[MAC]

Test Setups & Methodology

Test Setups & Methodology

For this review, we have prepared four different test setups, representing all the popular platforms at the moment, as well as most of the best-selling processors. As much as possible, the four test setups feature identical components, memory timings, drivers, etc. Aside from manually selecting memory frequencies and timings, every option in the BIOS was at its default setting.

Intel Core i7 LGA2011 Test Setup​

AMD Llano FM1 Test Setup​

AMD Zambezi AM3+ Test Setup​

AMD Phenom II AM3 Test Setup​

Intel Core i5/i7 LGA1155 Test Setup​

Intel Core i3/i5/i7 LGA1156 Test Setup​

Intel Core i7 LGA1366 Test Setup​

For all of the benchmarks, appropriate lengths are taken to ensure an equal comparison through methodical setup, installation, and testing. The following outlines our testing methodology:

A) Windows is installed using a full format.

B) Chipset drivers and accessory hardware drivers (audio, network, GPU) are installed.

C)To ensure consistent results, a few tweaks are applied to Windows 7 and the NVIDIA control panel:

• UAC – Disabled
• Indexing – Disabled
• Superfetch – Disabled
• System Protection/Restore – Disabled
• Problem & Error Reporting – Disabled
• Remote Desktop/Assistance - Disabled
• Windows Security Center Alerts – Disabled
• Windows Defender – Disabled
• Screensaver – Disabled
• Power Plan – High Performance
• V-Sync – Off

D) Windows updates are then completed installing all available updates

E) All programs are installed and then updated, followed by a defragment.

F) Benchmarks are each run three to eight times, and unless otherwise stated, the results are then averaged..

Here is a full list of the applications that we utilized in our benchmarking suite:

• 3DMark06 Professional v1.2.0
• 3DMark Vantage Professional Edition v1.1.0
• 3DMark11 Professional Edition v1.0.2
• 7-Zip 9.22 beta 64-bit
• AIDA64 Extreme Edition v1.85.1641 Beta / v2.00.1719 Beta
• Cinebench R10 64-bit
• Cinebench R11.529 64-bit
• Civilization V 1.0.1.383
• Crysis v1.2.1 64-bit
• Crysis 2 v1.9 + DX11 Pack + HiRes Texture Pack
• Deep-Fritz 12
• DiRT 3 v1.2.0
• Far Cry 2 v1.03
• HyperPI 0.99b
• Lame Front-End 1.0 (LAME 3.97 32-bit codec)
• Left 4 Dead 2 v2.0.8.9
• LuxMark v1.0
• MaxxMEM² - PreView v1.90
• PCMark 7 Professional Edition v1.0.4
• Photoshop CS4 64-bit
• POV-Ray v3.7 RC3 64-bit
• SPECviewperf 11
• Street Fighter IV Benchmark V1.0.0.1
• Team Fortress 2 v1.1.7.6
• TrueCrypt 7.1
• Valve Particle Simulation Benchmark v1.0.0.0
• WinRAR 4.0.1 64-bit
• World in Conflict Demo v1.0.0.0
• wPRIME version 2.05
• x264 HD Benchmark 4.0
• X3: Terran Conflict Demo v1.0

That is about all you need to know methodology wise, so let's get to the good stuff!

 

[MAC]

Synthetic Benchmarks: AIDA64 / MaxxMEM² / SiSoft

Synthetic Benchmarks: AIDA64 / MaxxMEM² / SiSoft

AIDA64 Extreme Edition 1.85 - CPU & FPU Benchmarks

AIDA64 Extreme Edition 1.85 - Cache Benchmark

AIDA64 Extreme Edition 1.85 - Memory Benchmarks

MaxxMEM² - Memory Benchmarks

Sisoft Sandra 2011.SP5 - Memory Benchmarks

Sisoft Sandra 2011.SP5 - Cache Benchmarks

​

 

[MAC]

Synthetic Benchmarks: SuperPI 32M / wPRIME 1024M

Synthetic Benchmarks: SuperPI 32M / wPRIME 1024M

SuperPi Mod v1.5

When running the SuperPI 32MB benchmark, we are calculating Pi to 32 million digits and timing the process. Obviously more CPU power helps in this intense calculation, but the memory sub-system also plays an important role, as does the operating system. We are running one instance of SuperPi via the HyperPi 0.99b interface. This is therefore a single-thread workload.

wPRIME 2.03

wPrime is a leading multithreaded benchmark for x86 processors that tests your processor performance by calculating square roots with a recursive call of Newton's method for estimating functions, with f(x)=x2-k, where k is the number we're sqrting, until Sgn(f(x)/f'(x)) does not equal that of the previous iteration, starting with an estimation of k/2. It then uses an iterative calling of the estimation method a set amount of times to increase the accuracy of the results. It then confirms that n(k)2=k to ensure the calculation was correct. It repeats this for all numbers from 1 to the requested maximum. This is a highly multi-threaded workload.

 

[MAC]

System Benchmarks: Cinebench R11.5 / Deep Fritz 12

System Benchmarks: Cinebench R11.5 / Deep Fritz 12

Cinebench R11.5

Cinebench R11.5 64-bit
Test1: CPU Image Render
Comparison: Generated Score

The latest benchmark from MAXON, Cinebench R11.5 makes use of all your system's processing power to render a photorealistic 3D scene using various different algorithms to stress all available processor cores. The test scene contains approximately 2,000 objects containing more than 300,000 total polygons and uses sharp and blurred reflections, area lights and shadows, procedural shaders, antialiasing, and much more. This particular benchmarking can measure systems with up to 64 processor threads. The result is given in points (pts). The higher the number, the faster your processor.

Deep Fritz 12 - Chess Benchmark

 

[MAC]

System Benchmarks: Photoshop CS4 / POV-Ray 3.7 RC3

System Benchmarks: Photoshop CS4 / POV-Ray 3.7 RC3

Photoshop CS4 64-bit

For the image editing portion of this review, we will use Photoshop CS4 in coordination with Driver Heaven’s Photoshop Benchmark V3, which is an excellent test of CPU power and memory bandwidth. This is a scripted benchmark that individually applies 15 different filters to a 109MB JPEG, and uses Photoshop’s built-in timing feature to provide a result at each test stage. Then it’s simply a matter of adding up the 15 results to reach the final figure.

POV-Ray 3.7 RC3

 

[MAC]

System Benchmarks: Lame Front-End / x264 HD 4.0

System Benchmarks: Lame Front-End / x264 HD 4.0

Lame Front End

Lame Front End v1.0 is a single-threaded application, which means that it only utilizes a single processor core. This will obviously limit performance but it will allow us to gauge a processor's single-threaded performance as well as test any turbo feature that it might have. We will be encoding a WAV rip of Santana’s Supernatural album and converting it to MP3 using the highest fidelity VBR 0 quality preset.

x264 HD Benchmark v4.0 x264

x264 HD Benchmark v4.0 x264
Test: MPEG-2 HD 720P Video Clip Conversion to x264
DVD Video Length: 30 Seconds
Comparison: FPS of Second Pass

x264 is quickly becoming the new codec of choice for encoding a growing number of H.264/MPEG-4 AVC videos. Think of it as the new Divx of HD and you can understand why we felt it critical to include. Tech Arp's recent development of the x264 HD Benchmark takes a 30 second HD video clip and encodes it into the x264 codec with the intention of little to no quality loss. The test is measured using the average frames per second achieved during encoding, which scales with processor speed and efficiency. The benchmark also allows the use of multi-core processors so it gives a very accurate depiction of what to expect when using encoding application on a typical full length video.

 

[MAC]

System Benchmarks: TrueCrypt / LuxMark

System Benchmarks: TrueCrypt / LuxMark

TrueCrypt 7.1

LuxMark v1.0