Intel DX79SI & Core i7-3960X

..:: Sandy Bridge-E and X79 ::..

Ahh, it’s that time again. Several weeks after a product launch, when the string of products hit the channels and start to make their way into the loving arms of enthusiasts around the world. The Intel X79 chipset is the latest edition of this event, and we’ll be taking a look at it today in Intel’s DX79SI “Siler” motherboard. I can remember back several years when Intel motherboards were consistently written off as stable, but lackluster as best. They were the boring mid-size sedan to others’ high-end sports coupe. This point was hard to argue, but soon things began to change for the better. We started seeing motherboard products that allowed for ever increasing flexibility. Are they at the pinnacle of “tweakability”? No, that post is still reserved for other third party manufacturers. However, with the steady change of pace, we are seeing better and better products for enthusiasts coming out of the Intel labs. Today, we’ll examine the DX79SI “Siler” motherboard with the latest production BIOS and drivers to see just where Intel is headed this time.

I always like to give the manufacturers some time after product launch to work out a few bugs with BIOS updates, driver updates, etc. before conducting a full evaluation, and that holds true today. Both the DX79SI and Core i7-3960X have been on the market for roughly three months now, though during some of that time competing boards were still stocking up in the market. Now that we have several options available, we can use these as a rough comparison in value and features. To give you an idea of the performance of the DX79SI and Core i7-3960X pairing, we will pit these two against our ASUS SABERTOOTH X58 and Core i7-965X. This board and processor combo reached new performance levels when they were tested, and will serve as a baseline for comparison with the X79 platform.

Intel Sandy Bridge – Core i7-2600K, Core i5-2500K

..:: Sandy Bridge Arrives ::..

It’s that time of the year again, CES is happening and naturally we have new products from Intel to toy with. If you’re new to the performance computing world, you may or may not have heard of the “Tick-Tock” philosophy that Intel employs for deploying new processors. The “Tick” events revolve primarily around a process improvement, i.e. Penryn was 45nm and Westmere 32nm, while a “Tock” event means we have a new microarchitecture gracing our presence. Intel’s Sandy Bridge is a “Tock” event, and is an improvement upon the wildly successful Nehalem microarchitecture. When Nehalem debuted, it was a boon for Intel placing them in the performance lead with little competition from AMD. Intel’s Core Series is still the predominant processor, and the most sought after series for MBReview readers. Today, we’ll be evaluating the Sandy Bridge microarchitecture, as well as comparing the new Core i7-2600K and Core i5-2500K against some Gen 1 Core processors. This will only be the first of several reviews over the next few days and weeks covering the new processors, and the new chipsets that come along for the ride.

One of the key weaknesses of the existing Westmere processors is the fact that those with integrated graphics require a specialized packaging technique. These processors use two individual dies, each based off of a different manufacturing process. If you remember back to the initial days of dual core processors, this is also how Intel chose to develop the first products. There was much argument over whether or not these were truly “dual core” processors as they were not one monolithic die. All arguments aside, this dual-die issue is addressed in Sandy Bridge processors. As the 32nm process has matured, Intel is now able to package both the integrated graphics core and processor core into a single monolithic die. Another key weakness, if you can call it that, is that the Nehalem / Westmere products simply can’t be bumped up in frequency much more, if at all. Given all the thermal and manufacturing constraints, the only other way to provide a boost in performance is to either move to a smaller manufacturing process and bump frequency, or improve the microarchitecture of the processor. That improved microarchitecture, and bumped frequency level, is Sandy Bridge.

..:: Sandy Bridge Overview ::..

Perhaps the most annoying buzzword of the past few years has to be “integration” and it’s associated variants, however, this concept has also brought some of the most innovative and popular products to date to the world. We’ve seen a steady progression over time of key functionality that formerly resided in the Northbridge being moved directly onto the processor die. This offers a slew of advantages with both cost and performance, and Sandy Bridge brings this effort to a head. Sandy Bridge processors will finally fully integrate the memory controller, PCI Express controller and now the graphics core. What’s next in the integration train? Only time will tell.

Intel Core i7-980X

..:: Intel Core i7-980X – Six Cores of Fun ::..


A mere two months ago, we had our first experience with a 32nm processor based off of the “Westmere” process, “Clarkdale”. At the time, we examined performance of the LGA1156 based Core i5-661, featuring the new AES-NI instruction set along with an integrated graphics die. We were impressed with the performance of the processor, especially the AES capabilities thanks to the new instruction set. The one thing we were left wondering was, how will the new instructions and die shrink effect the upcoming hexacore processors based off of the “Westmere” process. Today, we can finally reveal our take on the performance of the hexacore Core i7-980X processor, codenamed “Gulftown”.

..:: Intel Core i7-980X – A Closer Look ::..

The Core i7-980X processor is, of course, based off of the 32nm “Westmere” process technology, using high-k + metal gate transistors. The die boasts a massive 1.17 billion transistors, and yet comes in at a mere 248mm2. To put this into perspective, the “Bloomfield” processors that currently sit atop Intel’s high performance product line featured 731 million transistors in a die size of 263mm2. Thanks to the die shrink from 45nm down to 32nm, Intel has managed to squeeze in the extra transistors into a chip that’s actually smaller than “Bloomfield”, an impressive feat.

The Core i7-980X also boasts the same feature set as the “Clarkdale” processor, sans the integrated graphics die. The Core i7-980X sports Turbo Boost, Hyper-Threading, Triple-Channel DDR3, Extreme Memory Profiles, QPI, 12MB of Smart Cache and the new AES-NI instruction set. The Core i7-980X is also drop-in compatible with virtually all existing X58 based motherboards. All that should be needed for most is a BIOS update to add support for the i7-980X.

The Core i7-980X shares several similarities with the i7-975. The Turbo Boost settings for the i7-980X allow for a maximum frequency of 3.6GHz, the same as the i7-975, however the i7-980X is limited to 3.46GHz when running in multicore Turbo. Both processors are specified to have a 130W TDP, and a default frequency of 3.33GHz. Surprisingly, the 1ku pricing for each is also $999. Many were speculating initially that the i7-980X would have a much higher quantity price, but Intel has chosen to offer it at the same price as the existing i7-975. This isn’t to say you’ll see identical prices at the retail level. We’ve already heard of initial reports well in excess of $1,000 USD.

..:: Intel Core i7-980X – A Performance Heatsink…From Intel? ::..

The retail boxed versions of the i7-980X will also feature a new thermal solution from Intel, dubbed the DBX-B. Unlike previous thermal solutions from Intel, this new heatsink resembles numerous other heatpipe based solutions in the commercial market. Intel’s previous thermal solutions all featured similar circular designs and were controlled by the motherboard. This new heatsink allows for two settings, quiet and performance. This setting is controlled by a switch on the unit. Currently, Intel has no plans to market this heatsink as a separate product, but if the feedback is good enough, they are open to different options. We’ll be examining the performance of this new heatsink in comparison to a Thermalright Ultra 120 in our upcoming article, “Intel Core i7-980X Overclocking”.

..:: Intel Core i7-980X – Summary ::..

As you will see in the subsequent pages of benchmark results, Intel has a real winner with the Core i7-980X. Not only are we seeing performance gains of 25%, 35% and 50%+ from this new six core behemoth, we are getting it at the same price (in 1ku quantities) as the existing Core i7-975. This essentially renders the i7-975 obsolete as long as the retail market value of the i7-980X is the same, or competitive. This would also justify keeping the i7 namesake, as this will essentially replace the i7-975 at the top of the performance spectrum. We’ll have to wait a short while to see just how much of a price premium this new chip goes for, but if you were considering an i7-975, wait and see what happens because picking up an i7-975 now could prove foolish.

It’s extremely rare to see such a substantial performance boost that does not result in a price boost over the previous product. Especially when you’re talking 25-50% performance gains in multi-core, multi-threaded applications. Now, is the Core i7-980X for everyone? Of course not. Not many of us can readily drop $1,000 USD and up on a new processor. If you’re in the commercial market, or specialize in computationally intensive applications such as CAD / CAM / CAE, photo editing, video editing, etc. then this may be a valuable investment to increase productivity. For the average user, it really doesn’t add much to the current mix. Under normal user circumstances, you would be hard pressed to see a noticeable difference between the i7-980X and something far cheaper like the i7-920.

What we’ll be waiting for in the future is the debut of quad core processors based off of the “Westmere” 32nm process. These chips will bring all of the performance benefits such as the new AES-NI instructions, but should come in at a more reasonable price. Intel will not be creating a special four core die, rather using the same six core die with two permanently disabled. This should allow for a fairly quick roll out of quad core processors, and at a reasonable price considering how the i7-980X is positioned. If all else holds true to the past, within a year we should begin to see the roll out of cost effective six core processors as well. That is the day we will all be awaiting. Until then, keep the dream alive.

Now that you’re familiar with the Core i7-980X and our thoughts,continue on with the benchmark results and see just what this new chip has to offer. As a baseline, we’ll be comparing it against the current performance champs from Intel, the i7-965 and i7-975, along with the popular i7-920.

Intel Core i5-661 – Nehalem for the Mainstream

..:: H55 Chipset – Nehalem Hits the Mainstream ::..


Well, we’ve now entered a new year and we’re starting things off with a bang. Today, Intel has released the NDA on information and performance surrounding their new chipsets and the new Westmere processors. We’ll be waiting a few more months for the six-core version of Westmere targeted at the Extreme user, but with this release Intel has completed the circle bringing the Nehalem architecture to all mainstream users. Westmere doesn’t merely apply a die shrink and move on, not in the least. Intel has added several new instructions for AES encryption that vastly improve the performance and bandwidth of encryption calculations and applications. Did we also mention the option for a multi-chip package with integrated graphics? We’ll be examining Westmere’s performance with the Intel Core i5-661. We’ll see how this compares against the Core i5-750 and Core i7-870 all on the DH55TC “Tom Cove” motherboard. Before we delve into the performance of this new setup, let’s take a quick look at what Westmere has to offer.

..:: Westmere – More than a Die Shrink ::..


The benefits of a die shrink are well known to any enthusiast. The shift to 32nm is no different than what we have seen in the past. An impressive feat of the Westmere design is that we are now seeing TDP’s of 73W for a Westmere only solution, and 87W for a combined Westmere / IGP. This is vastly improved from the prior 45nm designs for the Core i7 and i5’s in then LGA1366 and LGA1156 package. We also see higher clock rates available at these much reduced TDPs. Does it get any better than a faster, cooler processor? I can’t think of anything I love better, so long as we don’t see a punishing performance hit to get there.

A second benefit of Westmere is the continued integration of what formerly consisted of the (G)MCH into the actual processor die. Integration is and has been a buzz word for years now, but with Westmere we can finally do away with the (G)MCH. Due to the integration of PCI-Express, the memory controller, and the option of a multi-chip package with the graphics die, motherboards no longer need a three chip solution. The remains of the (G)MCH and ICH have now been integrated into a single die, known now as the Platform Controller Hub, or PCH. This allows for a substantial space savings on the motherboard, meaning more compact solutions, or even better, boards with more features.

A multi-chip package is not a new development in the industry. You may remember the initial dual core processors consisted of two independent dies. This is the case now with the graphics die and the Westmere die on the Core i5-661. While Westmere is built off the 32nm process, the graphics die utilizes the 45nm process. Using a multi-chip package design allows for a simpler and faster integration of both dies into one package.  Will we see a single die solution in the future? I would bet on it.

Intel DP55KG / Core i5 750 / Core i7 870

..:: Introduction ::..

In November 2008, Intel launched the Core i7 processor based on a new microarchitecture codenamed Nehalem. This long-awaited microarchitecture brought dominant performance for the Core i7 900 series, but with that performance came at a steep price relative to Intel’s own Core 2 processors. X58 motherboards with prices in the range of $300-$400 USD are far from uncommon. This was certainly a contributing factor to a core component of the market simply being priced out of Bloomfield and X58. Intel recognized this fact, and after some delay has released the Lynnfield Core i5 and i7 800 series processors, as well as a new mid-range chipset dubbed P55. The real question is, have Lynnfield and the P55 chipset come to the rescue of the mainstream market?

It’s well known now that the Lynnfield core has had some features removed, i.e. QPI Links, however Intel has also added features such as an integrated PCI Express 2.0 controller directly in the die. This feature of Lynnfield should help to negate some of the loss of the QPI. After all, one primary use of QPI was previously to link the processor to the X58 IOH. Now that the IOH has been integrated directly into the die, we’ll benefit from lower latencies.

Another change that comes with Lynnfield is the move to supporting only dual channel DDR3 vs. triple channel DDR3 offered by Bloomfield and the X58 chipset. How much of a difference will this truly make in the end? I would initially suspect that there will not be much performance degradation in the shift from dual channel to triple channel when it comes to real world applications. We’ll soon see if this does indeed turn out to be the case.

An astonishing feat of the Lynnfield core is that it turns out it is larger than Bloomfield, yet we’re still seeing these sell at a discounted rate from the Bloomfield processors.Couple this fact with relatively cheap P55 solutions in comparison to their X58 brethren and you have what could be an excellent and affordable solution.

Today, we’ll be taking our look at Intel’s own P55 motherboard, the DP55KG, codenamed “Kingsberg”. This board boasts some impressive features, and as you’ll soon see have an added extra component or two. It’s always fun to work with board that still have debug ports and all of the engineer leftovers. Before I delve into the performance of the board, let’s take a tour around the PCB to see just what Intel has to offer with the DP55KG.

Intel 3.20GHz “Prescott” Pentium IV

..:: Introduction ::..

When the original Pentium 4 microprocessor debuted several years ago, it was their first x86 microprocessor built from the ground up to debut since the days of the Pentium Pro. From the times of the Pentium Pro onward to that of the Pentium III, all of the processor generations had been built on an ever improved P6 architecture. With the end of the Pentium III and the beginning of the Pentium 4, Intel moved on towards their NetBurst architecture scheme. At first, many did not particularly care for some of the NetBurst architecture’s features, those that many enthusiasts viewed as the weak points. The main aspect that was viewed as an achilles heel for the Pentium 4’s microarchitecture was the length of the pipeline, twenty stages, double that of the Pentium III. Well, “Prescott” keeps on trucking along this line of deeper piplines as it features a massive thirty one stage pipeline! 

 

As some of you may, or may not know, in the world of superscalar microprocessor architecture, one way to increase the clock frequency capabilities of a given microarchitecture is to increase the pipeline length. By increasing the length of the pipeline, executions can be broken up into separate, smaller steps which allow for easier scaling of the architecture, and therefore overall speed (MHz) of the microprocessor. The downsides to increasing the pipeline length are that the processor has a lessened IPC, or instructions per clock, capability. AMD, Intel’s main competitor has used their IPC rating as a way to set apart their offerings from the likes of Intel. Even AMD’s latest microprocessors based off of their x86-64 microarchitecture do not feature a pipeline as deep as that of Intel’s Pentium 4. The main downside to a deeper pipeline deals with branch misprediction. When a branch is mispredicted, the entire pipeline must then be flushed, and all of the data that was flowing through must restart from scratch. This induces performance killing penalties that can hamper the true capability of the processor. As pipelines are deepened, there is an ever increasing need for a more accurate branch prediction algorithm to avoid such problems, and we’ll soon see how Intel has tackled the problem of branch prediction.

As word of Intel’s “Prescott” core grew, many began to believe that “Northwood” was the end of the road for the Pentium 4, and “Prescott” would bring in the beginning generations of the Pentium V, or next generation processor depending on Intel’s naming scheme. After details of the core enhancements began to become publicly available as time went on, it was seen that “Prescott” would not bring an end to the Pentium 4 line, rather it would be yet another tweaked version of the core, with enhancements made to the microarchitecture, along with several “standard” enhancements such as an increase in the size of the on-die L2 cache. Many, including myself, initially believed that the 3.20GHz mark would be the end of the Pentium 4 “Northwood” core, although today, along with the release of several “Prescott” and “EE” processors, Intel has released a 3.40GHz “Northwood” core processor. In our review of the 3.20GHz “Northwood,” we found that Intel did seem to have enough frequency room to work with to create a 3.40GHz version, although we did not believe that they would do so, and obviously were incorrect. Today, we will be examining the various enhancements that Intel has made to with the “Prescott” core, and bring to rest some of the rumors about performance, microarchitecture changes, and more that have been floating around the internet.

Intel Pentium Extreme Edition 840

..:: Introduction ::..

The race to have the first dual core processors on the market has come to an end. Both AMD and Intel have launched their respective dual core processors, AMD with the Athlon 64 X2 series, and Intel sporting the Pentium D’s and Extreme Edition. AMD was the first to disclose plans for a dual core processor, and Intel soon followed after the well documented problems of reaching the 4.0GHz barrier with the “Prescott” core due to tremendous power requirements. The debate over who has the first real dual core processor is still going on with AMD throwing the latest punch at Intel’s solution to dual core with the D and Extreme Edition CPU’s.

The Extreme Edition 840 was released to the press in late April for a rather astounding action, an official preview with performance benchmarks. This was an unheard of event from Intel who normally keeps performance under wraps until the very second the NDA’s lift. AMD soon followed with sanctioned previews of their Athlon 64 X2 chips, and the race was on.

With Intel’s dual core processors, there’s one minor difference between the Pentium D’s and the Extreme Edition processor, and that is core support for Hyper-Threading. The Pentium D processors lack support for Hyper-Threading Technology, unlike the Extreme Edition series. Both these processors will operate off of an 800MHz FSB for the time being, though I’d hope to see a move to 1066MHz sometime in the “near” future.

Intel Extreme Edition 955

..:: Introduction ::..

It seems like it was just yesterday that we were sitting here discussing Intel’s move to the 90nm manufacturing process and the “Prescott” core, yet today we have the first chip in our hands built using the latest 65nm process, the “Presler” Extreme Edition 955. The move to 65nm comes after a rather problematic transition to 90nm. With the “Prescott” core, Intel adopted a huge 31 stage processing pipeline, up from the 20 stages that were previously seen across the Pentium 4 line.

Surprisingly, the “Prescott” core fared well against the 20 stage “Northwood” on a clock-for-clock basis due to some of the architectural changes. The Achilles heel of both the “Prescott”, and eventually the “Prescott 2”, core was and still is, heat. These cores pump out tremendous amounts of heat to go along with their light dimming power consumption. Well, maybe not quite to that level, but nevertheless under both idle and load conditions these cores gobbled up power.

The 31 stage pipeline can be thanked somewhat for causing the increase in power consumption, as well as heat. One of the main trade-offs of a high frequency, deeply pipelined processor is that it’s undoubtedly going to gobble up power. In the future “Conroe” will cut the current pipeline length to 14 stages, something that will certainly help Intel with power consumption.

With the later versions of the 90nm “Prescott” core, Intel took on the issue of power consumption by implementing thermal controls and dynamic clocking into the Pentium 4 line. We’ve seen this in action and it works like a charm, taking our 3.4GHz processor down to 2.8GHz when under lower load conditions. This technology helped lower the power consumption, but alone it wasn’t enough to make up the huge difference between the Pentium 4 and its leading competition from AMD.

With the advent of the “Smithfield” core, we saw similar heat and power consumption issues arise now that we had not one, but two processing cores running side by side. With the addition of EIST on the Extreme Edition and Pentium D models beyond the 820, the processors would dynamically lower the multiplier to 14x when the system was running under a low load state. If full processing power was required, then they’d quickly ramp back up to their full multiplier setting allowing for full utilization. The problem of power consumption will be better addressed in the future with “Conroe” variants and the un-named architecture.

For now, we’re seeing the launch of the first 65nm cores, code-named “Cedar Mill” and “Presler”. The “Cedar Mill” processors will be launched in the future and are quite simply the single core versions of the “Presler” dual core processors. With the move to 65nm, we should, in theory, see at least a decent drop in power consumption between the “Smithfield” and “Presler” cores. “Presler” and “Smithfield” differ in that “Presler” features two separate cores that are interconnected, while with “Smithfield” both of the processing cores were on one piece of silicon. Architecturally, they are one in the same, with a little shrinkage.

Intel Pentium D 820

..:: Introduction ::..

It’s been a few weeks now since we took our first look at a dual core processor in Intel’s Extreme Edition 840. We found that in the case of multi-threaded applications, these new dual core solutions will provide a very nice performance boost. The problem with the Extreme Edition 840, however, was that for the price and current application selection, it simply wasn’t a logical solution for the standard user. Why throw down so much money on a processor that has little software support for the main futures, 64-bit processing, and dual cores. Intel’s answer to this is the standard Pentium D series which we’ll be examining today.

The Pentium D attacks the dual core market on one important front, the one usually occupied by AMD, affordability. When AMD debuted their X2 series of processors, it didn’t make too many new friends other that the solid AMD enthusiast base due to the costs. Intel debuted the Pentium D’s priced only slightly above your standard Pentium 4 processors. Intel did this to position the processors for market penetration, an aspect they’re still dealing with today. To date, saturation of both Intel and AMD dual core processors has been slower than expected.

Another advantage of the Pentium D series is that, excluding support for Hyper-Threading, they’re no different than you’re Extreme Edition 840. This is another reason the Extreme Edition 840 is a hard sell. Hyper-Threading adds a nice boost to performance in a dual core setup, more so than it did for a single core CPU, but it isn’t enough to justify the multi-hundred dollar difference between the Pentium D. Today, we’ll be taking a look at Intel’s Pentium D 820 processor. This processor clocks in at 2.80GHz and could be what you’re looking for.

Intel LGA775 3.40EE & 3.60E (560)

..:: Introduction ::..

If we were able to take a step back in time, I’m sure that many of you reading this article would remember back when Intel proclaimed that processors in socket form were dead in the water. Intel, along with AMD, then progressed forward with slot mounted processors, which didn’t exactly last all that long on the grand scheme of things. Sure, we had Slot A, Slot I and II, but those only lasted through a generation and a half of processors. If you take care to remember, both companies moved back to socket implementations by the time the 1.00GHz mark had rolled around, AMD with their well-loved Socket 462, and Intel with their Socket 370, and then Socket 423/478. Lately, it seems as though Intel and AMD have switched places when it comes to the socket merry-go-round that consumers must deal with. Not long ago, Intel had the 370, 423, and 478 processor sockets, a fact that wasn’t looked well upon by many in our enthusiast community. AMD on the other hand had their Socket A, and, well, Socket A. Since then, Intel has stuck it out with the Socket 478 processor until today, where we’ll be debuting the latest LGA 775 socket while AMD has gained a higher precedence in the community with their Opteron and Athlon 64 processor lines, and have also adopted 754, 940, and now 939 pin sockets. Funny how things work out isn’t it?

Intel’s new LGA, or Land Grid Array, 775 processor socket takes a step away from traditional implementations in that the package no longer features pins, rather the bottom of the LGA 775 processors only have small gold contacts. With the LGA package, Intel has moved the pins into the bottom portion of the processor socket, something that will make installation of the processor easier in that there is no need to watch for bent pins on the package…although it will make it more difficult as well. You no longer need to worry about bent or damaged pins on the processor, rather now you have to worry twice as much about bent pins within the processor socket itself. We’ve heard some horror stories about the frailty of these pins, and from working first hand with LGA 775 motherboards, I can say that these stories likely can be taken with more than a grain of salt. In order to properly install the processor, you need to vertically drop it in the socket. If you angle the processor too much, we have found a slight angle to be alright, you risk bending some of the tiny pins and then smiling with delight at your now useless motherboard.

LGA 775 isn’t all bad, in fact it comes along with quite a few positive aspects. The processor can now deal with a higher amount of current flowing into it due to and increased number of power pins. This helps take the incredible current load off of many of the other power pins, and can help lower the amount of leakage, and also reduce some heat. Heat has been the primary issue with Intel’s latest processors, especially Prescott. Herein lies yet another positive aspect of LGA 775. With the plastic retention mechanisms used on 478 platforms, as heatsinks began to grow in mass due to increased thermal requirements, the motherboards would begin to bow due to the mechanical stress placed upon them. The heaviest heatsinks where mounted through the motherboard, as this helped relieve and disperse some of the stress. With the LGA 775 socket, Intel has allowed for similar implementations to be used as a standard. Many found the 478 heatsinks to offer incredibly easy installation, and with this new socket, Intel has made this task even easier.

Lastly, I know I’m not the only one out there who has went to remove their heavy copper heatsink and had the processor end up pulling right out of the socket with it. This has surely lead to some scary moments for consumers out there as well. With the new socket, Intel has fixed this problem. When the socket is unlocked, there is a small casing with a square portion removed from it that rotates upward. Once the processor is installed, and the socket locked down, this outer casing holds down the processor by the outer edges removing all possibility of damage due to a heatsink swap.