Intel Sandy Bridge-E Core i7-3960X Review: High End CPU roundup

[siopao1984]

---

Intel Sandy Bridge-E Core i7-3960X Review: High End CPU roundup by VR-Zone.com

With the announcement of the new Core i7-3960X by Intel, the current performance throne holder (i7-990X) was replaced, not surprisingly, by yet another Intel CPU. We took some time to compile a detailed comparison of all major current top desktop X86 CPUs by both Intel and AMD, to show you the real performance gains when moving from the old to the brand new.

[siopao1984]

Introduction

We often compare cutting-edge new CPUs to their brethren from the previous generation. After that, we rarely go back to the old models, simply because it doesn’t have too much of a practical value, but also because the new models are better than the old ones as a rule.

Furthermore, as far as CPUs are concerned, we often talk about various abstract subjects such as bottlenecks – a term we often encounter when testing the strongest graphics cards, for instance. It’s occurred to us recently that we often don’t compare the old and new in a sufficient number of tests, perhaps because of our desire to dive into the new technology as much as possible, which may leave out some of the much needed objectivity. This might cause some confusion amongst PC users and a wrongful logic when assembling a new PC or upgrading the old one. Are new CPU generations really all that better than the old ones? Do they bring all the marketed increases in performance, or are those just percentages occurring in particular programs under particular conditions? Ultimately, is it all about presenting you what you want to see rather than what it is?

During conversations with reputed hardware manufacturers, for instance, we’ve encountered the general attitude that Sandy Bridge CPUs are around 30% faster than Nehalem/Westmere-based ones. Many articles on the net confirm the same allegations. However, as time went by, we’ve established that this sort of prejudice is just wrong in many ways, as the situations where that percentage is reached come in a very limited number. In regard with that situation, this review is meant to give an overlook of the strongest CPUs made by AMD and Intel so far.


Test Setup

For the needs of this test, we obtained highest end CPU's for each generation, both from AMD and Intel. Intel was represented with Westmere Core i7 CPU (Core i7 990x) and Sandy Bridge - Core i7 2700K and Core i7 3960x. AMD was represented with their strongest CPUs: AMD FX-8150 (Bulldozer) and AMD Phenom II X6 1100T (Thuban).



After obtaining a total of five CPUs (three from Intel and two from AMD), we had to decide on the motherboards to be used. The negative aspect of Intel’s policy to change the socket with each new CPU generation came to full effect yet again – three motherboards and three OS installations were required just for their part of the test.



In order to do the test we obtained following motherboards for Intel processors: ASUS Rampage IV Extreme (X79 chipset), ASUS Sabertooth X58 (X58 chipset) and ASUS P8P67 WS Revolution (P67 chipset). AMD only required a single 990FX motherboard for both CPU's- ASUS Crosshair V Formula (990FX chipset). These five systems were tested in two scenarios. The first entailed the CPU set to its default clock, while the second one was with all CPUs set to exactly 4.5 GHz, thereby measuring clock-for-clock performance. The only exception to this was Phenom II X6 1100T, which had a lower OC margin, so it had to be excluded from the second round. In order to make testing more realistic, we overclocked through changing the multiplier exclusively, while the FSB remained locked at its default value. Furthermore, the RAM clock was constantly at 1600 MHz with 9-9-9-24 1T latencies. All these precautions isolated the CPU as much as possible. The chipset factor is something you just can’t avoid in this sort of test, but we’re fairly certain that its impact on overall CPU performance is irrelevant. As the number of RAM channels has a lot to do with the architecture of each CPU, we used the best available settings each time, ranging from dual- to quad-channel memory, the latter being used by Sandy Bridge E models.

For all the test for round up we used Kingston HyperX 1600 CL9 4 x 4GB kit, and just for special, latency test we used Kingston HyperX T1 2133CL9 4x2GB kit. Having in mind both performance and capacity two Kingston HyperX 240GB SSD were used. Trusty and more then capable, PSU Cooler Master UCP 900W was used to power up all these monsters. We used two different cooling solutions; primarily it was Cooler Master V10 for all the sockets beside socket LGA 2011. For that one we used Intel liquid cooling solution (RTS 2011LC) that came together with Core i7 3960x CPU. Only logical choice for VGA was ASUS GTX580 as a most powerful single GPU graphic card.

Tests were split into two categories: typical benchmarks of the CPU, memory and system in general on one hand, and various games on the other. All in all, over 300 tests were done in total, and there was a lot to deduce as well.

[siopao1984]

A bit about CPUs



As for AMD, the older cream-of-the-crop CPU carries the title Phenom II X6 1100T. It’s a hexa-core CPU made in the “old” 45 nm production process. It has 3 MB of L2 and 6 MB of L3 cache, works at a default clock of 3.3 GHz, and reaches up to 3.7 GHz in Turbo Core mode. It is also the cheapest of all models in this roundup, which makes it an attractive buy. The second AMD product is, you’ve guessed it, FX-8150, which again offers the most cores for the given amount of money. AMD FX-8150 has the Bulldozer chip, made in the new 32 nm lithography, and working at 3.6 GHz by default and 4.2 GHz in TurboCore mode. Speaking of cache memory, it should be said that, due to the significant number of cores FX-8150 has, it had to have 16 MB of cache. This cache is evenly distributed between L2 and L3 – 8 MB each. The specificity of its architecture is a very controversial subject; although it looks fantastic “on paper”, the general impression of Bulldozer isn’t too good, and that view is shared by media and disappointed customers alike. True, the world may not be ready for AMD’s fledgling just yet, and things are bound to go up as time passes, but that doesn’t mean all that much at the present moment, unfortunately.

On the other side of the ring, Intel has seen little but praise ever since the first Nehalem came out. Sandy Bridge reinforced their position as the provider of successful and popular architectures. We picked the very expensive Core i7 990X Extreme Edition as the old model. This hexa-core Nehalem had a price of over 1000€ when in first came out, and it hasn’t reduced all that much up to today. With the aid of hyperthreading, the total number of parallel threads it takes is 12. The default clock of this CPU is 3.46 GHz, and Turbo Boost stretches this number up to 3.73 GHz. Unlike AMD, the oldest model on Intel’s side is made in 32 nm. The large cache of 12 MB, called Smart Cache by Intel, provides all six cores with ample breathing room.

Next up is Core i7 2700K, which is a stronger version of the enormously popular 2600K, one of the best CPUs in history in terms of the price/performance ratio for anyone looking to build a really strong system. It’s based on Sandy Bridge architecture, and used to be the strongest model based on this core until recently, but it was still only a quad-core CPU. After this CPU generation came out, it became clear that Intel wasn’t planning to stick around with the then-current socket for too long, as the new Extreme series required a new one. This is why the LGA2011 platform was created, in order to round up the product gamma and create clear segmentation in Intel’s lines.



Intel Core i7 3960X

As a reminder, the original Sandy Bridge CPUs have four cores each, with i5 revisions containing 6 MB of L3 cache, while i7 have the whole 8 MB as well as hyperthreading. So far, the desktop segment’s cream of the crop has been Intel’s undisputed i7 990X, based on the old Nehalem microarchitecture codenamed Gulftown. The time is nigh for a generation shift, and Intel’s socket strategy in this iteration gets a new CPU platform. Intel’s new beast is using the brand new LGA 2011 socket.



This basically means that we now have two Sandy Bridge platforms: LGA 1155, using a dual-channel memory controller from inside the CPU, and LGA 2011, which has double that bandwidth due to a quad-channel controller. In other words, if you want maximum performance, make sure you place four modules in your motherboard. This sort of memory bandwidth is mostly helpful with the execution of HPC program code, as this sort of applications are replete with vector calculations and are extremely memory-intensive in terms of bandwidth for the most part.

With a large number of cores, narrow memory bandwidth may easily present a bottleneck, which makes quad-channel access the logical option. Sandy Bridge E, which means the entire LGA 2011 platform, is actually a server platform adjusted to desktop standards. In its very essence, Sandy Bridge E is Intel’s Xeon E5, raised to a much higher frequency, and with two of its eight cores shut off. Each Sandy Bridge E CPU has the cores, L3 cache and the GPU connected to the ring bus. This series’ L3 cache is split into 2.5 MB segments – a single ring port contains one core with 2.5 MB of L3 cache, which amounts to 20 MB of L3 cache for a total of eight cores. Shutting off two cores has consequently provided a CPU with six cores and 15 MB of L3 cache. The standard desktop version uses 2 MB chunks (2600K has 4 x 2 MB of L3 cache).



This new i7 3960X is a top model, with a price of around 1000$ on the US market. Its brother, i7 3930K, has an unlocked multiplier just as well, but a smaller L3 cache of “only” 12 MB; needless to say, its price is much lower, almost half that of the stronger version. There’s also a quad-core i7 3820 in the making, with 10 MB L3 cache, but its price remains a mystery for now, while its performance is expected to be somewhat better than the current LGA 1155 model 2700K.

Core i7 3960X has a maximum Turbo Boost frequency of 3.9 GHz, just like i7 2700K. However, its default clock is 3.3 GHz for all cores, which is 300 MHz more compared to the default clock of i7 2700K. For comparison’s sake, i7 980X, also “ticking” at 3.3 GHz, or more precisely, 3.33 GHz, has a Turbo Boost frequency higher by only 133 MHz, so it works at 3.33-3.46 GHz in most cases.

Conceptually, this is an improved Nehalem/Westmere, but the microarchitecture itself has a few very important changes that have taken place, impacting both consumption and performance. Here are the most important ones:

Decoded microoperations cache (similar to trace cache in Netburst architecture)
Physical register file
Double 128-bit load
Increased buffer and on-the-fly instruction size – better Out Of Order execution
Improved memory controller that officially supports DDR3-1600 memory, here as quad-channel
256-bit ring bus architecture connecting L3 cache, the cores and the GPU system controller
Turbo Boost 2.0
GPU/IGP has also been left out in Sandy Bridge E, as it was initially designed for server platforms.

As for other features, everything that you know about the existing gamma of LGA 1155 CPUs applies here as well. These improvements were made in order to additionally speed up execution of traditional code without recompiling.

The new product is also complemented with the new set of SIMD instructions – AVX. This is short of Advanced Vector Extensions. This isn’t an extension of the previous SSE instruction set, though; it’s an entirely new format altogether. AVX uses extended 256-bit registers to work with 256-bit vectors. In applications that use AVX, a performance improvement of up to 50% can be observed compared to standard code.



X79 chipset

The X79 Express chipset doesn’t contain anything essentially new, except for the larger number of PCI Express lines and PCI-E 3.0 support, perhaps. Surprisingly enough, the new chipset doesn’t support USB 3.0, although the technology has been around for a while. The chipset isn’t exactly reputed for coolness, so active cooling has to be implemented. For instance, ASUS’ Rampage IV Extreme motherboard, the one that was used for testing purposes, has a fan on the PCH. PCH, or Platform Controller Hub, is the new name for what was previously known as the Northbridge.



X79 chipset’s internal codename is Patsburg. Unlike the X58 chipset, X79 isn’t using QPI links for CPU communication, instead relying on DMI2 interconnection. DMI (Direct Media Interface) is used to connect the CPU to the PCH, and essentially works as a PCI Express controller. X79’s DMI2 uses exactly the same communication protocol as the PCI Express bus, with four lines supporting speeds of up to 5 GT/s each. This is more than enough for any multi-GPU system to communicate with the RAM.

Similarly to QPI and X58, X79 also has an I/O bus used by peripherals, PCI-E and the rest of the system to communicate with the memory, with the difference being only in the fact that the PCI Express interface is now closer to the CPU core. The memory controller is situated inside the CPU itself, so all communication takes place indirectly through it. The main difference in system architecture between Nehalem and Sandy Bridge CPUs is that the DMI controller is integrated into the CPU in Sandy Bridge, while X58 and Nehalem had DMI located in a separate chip; as a result, X79 only has a single chip, while X58 has two: X58 IOH and the ICH10R Southbridge.

BCLK overclocking is limited, just like with P67 and Z68 chipsets. LGA 1155 CPUs with Sandy Bridge architecture have brought integrated clock generator inside the chipset. Since there’s no external clock generator anymore, raising the BCLK value also raises the frequency of all other components in the system, such as PCI-E and SATA links. This makes it much more difficult to do the familiar BCLK-tied overclocking. However, K-suffix CPUs have an unlocked multiplier, and the same goes for the tested Extreme Edition marked with X, which makes overclocking that much simpler.

At factory-default settings, i7 3960X uses much more power than its previous-generation counterpart, i7 980X, i.e. i7 990X, and the consumption measured was higher than AMD’s flagship FX-8150 CPU too. In overclocked mode, the new CPU fares a bit better. Therefore, from the consumption aspect, Sandy Bridge E is definitely not an eco-friendly solution. Performance is around 16% better clock-for-clock compared to i7 980X, which is a respectable advancement. Certain applications give Core i7 3960X an advantage higher than 200% compared to AMD’s FX-8150, based on Bulldozer architecture, and around 40-45% compared to i7 980X.

[siopao1984]

ASUS Rampage IV Extreme (X79) Motherboard

ASUS Rampage IV Extreme’s cooling system is very well made, which kept this model suitably cooled at all times during testing. The aluminium profiles lie on the components via thermal tapes, and we particularly liked the fact that there’s a backplate behind the power section components, which makes the heat spread even more efficiently. ASUS installed a turbine fan on one of the coolers, which keeps the entire model perfectly cool at full speed. In silent mode, Rampage IV Extreme was warm, but no more than that. Although ASUS is quite confident in the cooler’s performance, it’s still bound to attract major amounts of dust, so pay attention to that.



This model’s layout is excellent; we have no negative remarks. The motherboard has four third-gen PCI-E x16 slots. When all four graphics cards are present, the first two slots work in x16 mode, while the second two work in x16. No, it’s not a typing error – this motherboard actually supports all four slots in x16 mode. We can’t recall seeing that feature so far, which makes this motherboard a quad-GPU Mecca. There is even a fifth PCI-E slot which works in x8 mode, as well as a normal x1 one.



The chipset and DIMM slots have excellent voltage filtering, while the CPU is served by ten high-quality capacitors closely cooperating with the digital power unit. We’re talking about the second generation of ASUS’ Extreme Engine Digi+ subsystem for CPU voltage calibration and control. It should also be mentioned that this motherboard has two CPU power connectors: an 8-pin and a 4-pin one. This will make extreme overclockers rejoice, and we’re sure that world records will be brought down in flames using this very motherboard.



Another excellent feature is that ASUS has enabled users to keep their old LGA1366 coolers via a special bracket contained in the bundle, so if you already have one, you won’t have to buy a new one. Excellent move!



The upper right corner of the motherboard contains the Power and Reset buttons, as well as the two-character diagnostics display. Physical voltage measuring points are present throughout the motherboard as well, which is another argument in this motherboard’s reach for the overclocking community. It should also be mentioned that there are physical microswitches on the PCB that can turn off whichever PCI-E x16 slot you want to.



The motherboard has four SATA II and four SATA III ports. Two out of four SATA III ports are in command of ASMedia’s controller, which does its job well. All connectors for use with the enclosure (power, reset, power LED, HDD LED etc.) are neatly marked. The DIMM slot brackets are removed from the slots closer to the graphics card, so your RAM won’t physically come into conflict with your primary graphics card in the top slot, whatever its length. USB 3.0 was to be expected on the PCB itself from such a high-profile motherboard, and our expectations were met. Fan connectors are numerous and scattered around too, so that you never have to look for one for too long. Realtek’s ALC898 (7.1) audio chip goes for one of the best integrated audio solutions, which has been tested and tried numerous times. The motherboard only has one Gigabit LAN adapter, but considering the sheer number of things present on the I/O panel, we can’t really consider this a drawback. Of course, the star of the I/O panel is the always handy Bluetooth module in 2.1 + EDR revision.



ASUS’ VGA Hotwire technology is intended for true overclocking connoisseurs. The PCB contains a set of pins right next to the Power and Reset buttons. These are made for connectors that can be placed here on one side and welded onto the graphics card on the other. What you get is top-quality voltage control of the graphics card via the already familiar OC Key system. To the point and highly usable – good work, ASUS!

The EFI BIOS is made in the black “ROG” fashion. It’s replete with overclock options, and everything can be adjusted to the smallest detail. In short, this motherboard has everything you could ever want from a motherboard.

The CPU’s BCLK worked completely stably at 230 MHz without an increase in any voltage value. In extreme conditions (overclocking session with liquid nitrogen), the BCLK of our Core i7 3960X EE CPU would have probably worked at 15-20 MHz more than that. Kingston’s 2133 MHz CL9 RAM that we used was able to work stably at its default frequency and a voltage of 1.65 V, which is stunning for a quad-channel kit. A stable overclock of 4.5 GHz was achieved with HT on and a voltage of 1.37 V, while turning off the HT brought us up to almost 4.6 GHz at the same voltage. All these voltages are 100% safe for everyday use, so we couldn’t be any happier with Rampage IV Extreme. Of course, higher voltages will gain you even higher clocks, but since we were stuck with Intel’s not-so-efficient water-cooling system, we could only look for voltage drops at higher voltages, not really test the entire platform thoroughly.



All in all, ASUS’ Rampage IV Extreme is one of the best motherboards that we’ve encountered lately, and definitely the best if we restrain ourselves to Intel’s X79 chipset. It’s really a top product in every way, targeting gamers and overclockers. The only thing that could bother you is the rather steep price of around 360€, but we’re talking about the fastest desktop platform money can buy, so the price is adequate to the concept.

[siopao1984]

Futuremark - 3DMark 11, 3DMark Vantage and PCMark 7

3DMark 11






3DMark Vantage






PCMark 7



Futuremark benchmarks gave very interesting results, primarily 3DMark, both versions of which prefer Intel over AMD. In their case, 3DMark programs knew well how to recognize additional cores, so both the old and the new Core i7 got very high marks. The difference between the two is almost non-existent at default clocks. On the other hand, when we bumped up both CPUs to 4.5 GHz, the new architecture gained significant advantage in 3DMark 11, giving Sandy Bridge a margin of 20% over Nehalem. Vantage wasn’t able to establish such a big difference, giving 10% more to Core i7 3960X compared to 990X. PCMark

11 have proven to be a benchmark which distinguishes platforms very little, so all systems scored within 10% of each other. As for AMD, 3DMark of any version gave almost no advantage to FX over the older Phenom, despite its having two more cores and a clock higher by almost 300 MHz, which is a bit disappointing.

[siopao1984]

Cinebench R11.5



Cinebench has shown yet again how good it is at making use of Intel’s CPUs, which performed remarkably. Regardless of the test type (OpenGL or CPU-heavy tests), Intel’s second-gen CPU generation has proven to be fantastic. Even 990X paled in comparison with 2700K in the OpenGL test. The single-core test was particularly inclined towards the Sandy Bridge architecture, as the difference in performance was up to 50% compared to other architectures. AMD has left us underwhelmed, on the other hand, with the new Bulldozer scoring lower than 1100T, using the old Thuban core, in every single subtest. Weak optimization is the cancer here, but that’s hardly an excuse, just an explanation, however bizarre it may seem, as both CPUs came from under the same roof.

SiSoft Sandra 2011 Professional



















SiSoft Sandra 2011 is the following synthetic benchmark which proved many points. All CPU-focused tests gave a huge advantage to hexa-core versions of Core i7 over the quad-core 2700K. Another interesting thing is that the weaker Sandy Bridge model came dangerously close to its stronger brethren when overclocked. AMD’s Bulldozer is allegedly not for use with this benchmark, as the latter has still not been optimized for it, although we can’t see the base of these claims, as FX-8150 fared very good in comparison with Intel’s and AMD’s own models alike.

[siopao1984]

7-Zip, AIDA64 and x264 Video Compression









As far as memory tests are concerned, and by that we mean the likes of 7-Zip, but also AIDA64, it was clear to us that all hexa-core Intel CPUs have an advantage due to the triple-channel memory access being used. However, the benefit of the latter hasn’t really shined through too confidently, with the best score in WinRAR being achieved by FX-8150, which still uses the standard dual-channel access.

As for video compression, the situation is very interesting. This is one of the tests that simulate real-world scenarios as much as possible, and Nehalem and Sandy Bridge outperformed AMD’s CPUs as expected. However, FX-8150 shows significant improvements over 1100T, and manages to keep up with 2700K, which is much more expensive. An additional surprise was presented at 4.5 GHz – 2700K still wasn’t able to leave FX-8150 behind!

[siopao1984]

Gaming I: Street Fighter 4, Left 4 Dead 2 and Civilization V







Games such as Street Fighter 4 and Left 4 Dead 2, which are pretty outdated in the graphics segment and work well on virtually any modern card, make up the first group. Their performance is largely dependent on the CPU with a graphics card such as GTX 580. It’s true that it’s only a matter of which three-digit number you’ll get as the end result, but this makes it that much easier to see the difference between the CPUs. The differences are large, as you can see, but all Core i7 CPUs fall within 5% of each other. Not that this bears any weight in the performance of the actual game, though, as it’ll work just as well with a Core i3. Incredibly enough, the previously seen situation with AMD reoccurs, placing Phenom II X6 1100T over FX-8150, which is disappointing, despite the academic nature of this test.

The second game type is represented by Civilization V, the major part of which consists of CPU calculations. This type of games is uncommon today, and it’s usually strategies with complex AI, such as the Total War series, which means that the graphics section isn’t very demanding, but the CPU part can be, especially in large battles, where the number of units on screen at the same time reaches gargantuan proportions. If we had to pick a single model to play this game on, other than the expectedly superior and just as expensive Intel’s strongest model, it’d be 1100T, which managed to keep up with 2700K, and again, overpower FX-8150.

[siopao1984]

Gaming II: Battlefield 3 and DiRT 3





The last and most common case are games that are very graphics-intensive: Battlefield 3 and Dirt 3. This is by far the most realistic scenario in terms of usability, and showcases the real situation best. If you have a look at the charts, you’ll see that practically every system basically yielded the same performance. 10% seemed like a huge gap, and differences were academic for the most part. The strongest single-GPU graphics card that we used gives exactly the same performance with a 150€ CPU as it does with a 1000€ CPU. Bear this in mind if games are what you’re worried about.

[siopao1984]

Power Efficiency



The last aspect is the energy efficiency. As a rule, newer models mean better efficiency, although Bulldozer has done well to prove us wrong. Two extra cores justify it to a certain extent, but insufficiently so to convince us entirely. AMD has stepped on the line with power consumption, and that has to be said. Intel has managed to improve efficiency, on the other hand, despite the fact that Nehalem was manufactured in 32 nm just as well as new CPUs. Ingenuity is always to be applauded.