Original Link: http://www.anandtech.com/show/2468
Intel's 45nm Dual-Core E8500: The Best Just Got Betterby Kris Boughton on March 5, 2008 3:00 AM EST
- Posted in
What's the next best thing to an Intel 45nm quad-core processor? Why, a 45nm dual-core, of course. At least that's what Intel seems to being saying lately. While we tend to agree, there are certainly more than a few important considerations to take into account when deciding just which CPU is best suited for your intended uses. Choosing a CPU can be as personal an experience as buying a new car. While you know what you want, it really comes down to what you need, what you can afford, and more importantly, what makes sense. Although the four-core model easily overclocked to 4GHz or higher on air alone certainly does sound sexy, the brown sub-compact in the corner of the lot may be just what you're looking for. Don't worry though; either way Intel has an answer for you….
Intel's "tick-tock" strategy gives us a very early glimpse at the future of micro processing. If all goes well, Moore's Law should be as true in 2010 as it is today…
Amid rumors of manufacturing problems, the next step in the continuation of Intel's accelerated "tick-tock" strategy - which pledges process-technology shrinks of existing designs and the introduction of entirely new core architectures on an alternating two-year cycle - comes the release of a new line of 45nm dual-core processors, codenamed Wolfdale. Built on the familiar Core 2 architecture, these processors feature a few notable changes with some rather large implications for the overclocking crowd, all of which we will discuss in more detail later. For starters, advancements in process technology have allowed Intel to shrink the size of the transistors used in these CPUs from last-generation's 65nm down to 45nm, allowing for a ~50% reduction in die size for an equivalent design.
The changes don't end there; a few core processing modification have been made, making Wolfdale a little faster clock-for-clock than Conroe. These changes include but are not limited to: a new divider technique called Radix 16 that nearly doubles the speed of operations involving operand division, the introduction of 47 new Intel Streaming SIMD Extensions 4 (SSE4) instructions (many perfect for HD video production and decoding), and a unique 128-bit wide Super Shuffle Engine that significantly improves performance for all SSE-related instructions (i.e. content creation, image manipulation, and video encoding). Unfortunately, it will take some time for software developers to catch up with most of these innovations, but eventually we should see more and more programs and games that show the true power of these feature sets.
The layout of discrete components on the bottom of any Intel CPU is an easy way to quickly determine which product series you hold in your hands. This is what a 45nm E8000-series dual-core looks like.
Finally, the L2 cache size has been substantially increased. The E8000-series processors will feature up to 6MB of shared L2 cache, up from 4MB per core pair. However, the larger L2 cache comes with a move from the previous low-latency 4MB 8-way association scheme to a more complicated 24-way associated cache when using 6MB, adding precious nanoseconds to each data fetch. The larger cache is technically "better", but the higher latencies will in some cases negate the benefit, so this is not a clear improvement in 100% of cases. There has been no formal word yet from Intel as to whether this trade-off was a result of the use of the larger cache or if it was an intended design change.
All 45nm dual-core Intel CPUs will operate at a default bus speed of 333MHz (1333 quad-pumped), which is needed in order to give the included store forward technology and intelligent prefetch algorithms the fast memory access they desire. These background processes, combined with the use of the large L2 cache, are instrumental in Intel's recent success in hiding most of the traditional memory access latency experienced with many older designs. Although memory access operations are still slower than desired, more cache means these processes are able to look farther ahead, fetch more data into the L2, and increase the chances that an incorrect branch assumption will not result in a costly data stall. The move to an integrated memory controller (IMC), like that in Nehalem planned for a Q4 2008 release, will largely invalidate the necessity of these super-scalar caches.
The Intel Core 2 Duo E8500 processor promises to be the fastest, most energy efficient dual-core CPU ever designed for the PC.
We have noted in previous articles what an amazing difference Intel's new high-K 45nm process has made in the improved switching efficiency and the reduction in leakage current of these tiny transistors. Our results with the Core 2 Extreme QX9650 were nothing short of impressive. First impressions left us feeling as though Intel had either made a mistake in calculating the processor's thermal characteristics, or more likely they decided to conservatively rate these new quad-cores relative to the older 65nm quad-core CPUs. In any case, the improvement was real and measurable.
Drawing upon that same success, the E8000-series of dual-core processors shows great promise when applied in situations that demand unrivaled performance and/or energy-efficient operation. While there is no doubt that the E8500 will excel when subjected to even the most intense processing loads, underclocked and undervolted it's hard to find a better suited Home Theater PC (HTPC) processor. For this reason alone we predict the E8200 (2.66GHz) and E8300 (2.83GHz) processors will become some of the most popular choices ever when it comes to building your next HTPC.
What's more, Intel has decided to stay with the classic LGA775 package for at least one more round, meaning there is a good chance an upgrade to one of these new 45nm processors could be easier than you originally thought it would be. Past upgrades have required the purchase of an entirely new motherboard due to modifications to the Voltage Regulation Module (VRM) specifications, dictating board-level hardware changes needed for new processor compatibility. Not so with Wolfdale; a simple micro-code BIOS update from your motherboard vendor is often all that is necessary to add official support for these CPUs. After that, it's only a matter of removing the old processor and installing the new one, and you can begin enjoying all the benefits an E8000-series processor has to offer.
E8000 Lineup and Early Overclocking Results
Intel plans to introduce no fewer than four new 45nm dual-core processors as part of its Q1 2008 Wolfdale launch. At the time of publication, the following models are scheduled for immediate availability:
- E8200, 2.66GHz, 6MB shared L2 cache, 1333MHz FSB, maximum 8.0x multiplier
- E8300, 2.83GHz, 6MB shared L2 cache, 1333MHz FSB, maximum 8.5x multiplier
- E8400, 3.00GHz, 6MB shared L2 cache, 1333MHz FSB, maximum 9.0x multiplier
- E8500, 3.16GHz, 6MB shared L2 cache, 1333MHz FSB, maximum 9.5x multiplier
Estimated street prices, although unconfirmed and subject to change, are expected to be around $299 for the E8500, $249 for the E8400 and $163 for the E8200. As you can see, each processor features a maximum multiplier - there has been no formal mention of an "Extreme Edition" dual-core processor at this time. Additionally, rumors of the impending release of an E8600 processor (presumably running at 3.33GHz with a 10x multiplier) go unconfirmed. It it becomes available, the E8600 may very well be the CPU to own as it would allow for operation at 10x400 (4GHz), a very good place to be when it comes to tuning in maximum memory performance.
Unlike the Conroe release, all processors will make use of the full 6MB shared L2 cache offered on the top-end E8500 model. Undoubtedly, a 45nm Celeron or Pentium line (or E5000/E7000 - choose your favorite naming scheme) will eventually make their way to retail. We expect these to come in at 3MB and/or 1.5MB of shared L2 cache. Based on what we have seen when it comes to 65nm Pentium E2000 and Core 2 E4000 chips, when they do arrive the 45nm variants will offer tremendous value and an amazing price/performance ratio.
Designed to run at an already fast 3.16GHz, this E8500 is just starting to stretch its legs and show its true potential with some water-cooling TLC.
We were able to overclock our E8500 sample all the way to 4.5GHz with water-cooling; what's more, we were able to demonstrate complete stability at these speeds running many hours of Prime95, a popular tool for stress-testing systems. Most X6800/X6850 owners will attest to this amazing achievement - the average overclock for top-bin 65nm CPUs falls somewhere near the 3.8 ~ 4.0GHz mark. Results such as these combined with Wolfdale's modest clock-for-clock advantage over Conroe show the prospect of 15% or more processing power when overclocking. Even though the official frequencies may not have changed significantly (yet) - a new maximum of 3.16GHz, up from 3.00GHz - this increase in overclocking headroom makes the E8500 a marvel to behold.
Our top (unstable) overclock on water is nothing short of impressive. Although we were unable to benchmark up here, future steppings may change that. The ability to POST and load Windows at 4.8GHz on water promises more to come….
A quick maximum-frequency run on water indicates the proverbial sky's the limit when it comes to overclocking the E8500. The maximum achievable frequency had more to do with our nerves than anything else. Given the voltage, our E8500 was more than happy to continue scaling higher. However, we eventually said enough is enough and called it quits - that point came when we were subjecting our poor 45nm CPU to over 1.6V, a level that could possibly require you hand over your credit card number in exchange for another CPU in no time flat. Quite simply, we believe any voltage over 1.45V is asking for trouble with 45nm processors and our conversations with Intel to date have all but confirmed our suspicions.
The E8500 is no slouch when it comes to chasing a high FSB. Of course, all of this will be for naught when Intel releases their next-generation Nehalem architecture.
Here's the obligatory high-FSB screenshot, for those that care. In case you missed it, we recently had an in-depth article on why high FSB overclocking might not really be the best approach to take when dialing in maximum system performance. Suffice it to say, our fascination with these displays of CPU or motherboard worthiness is rapidly waning. However, we are also not so stubborn as to not acknowledge the importance of high FSB potential when it comes to pushing processors with low multipliers. Our recommendation is straightforward, however: buy the model with the highest multiplier that you can possibly afford. Most motherboards (and systems) are far happier running 9x490 than 8x550.
So Dual-Cores are no Longer Extreme?
It may be hard to believe, but the quad-core concept just celebrated its first birthday. Launched in late 2006, this anniversary also signifies the introduction of a rather significant adjustment to Intel's eternally-evolving marketing strategy. For the first time ever, Intel has decided not to produce a dual-core Extreme Edition variant of their leading quad-core product offering. That means there are currently no plans to manufacture a 45nm dual-core CPU featuring an unlocked multiplier (or as Intel likes to put it, with "overspeed protection removed"). Until now, this made choosing the right processor easy: those that lacked the means (or the need) for a quad-core could feel content in knowing they would not be expected give up having an unlocked multiplier should they decide to go with the dual-core in lieu of quad. Now, anyone that wants to enjoy the operational freedom that comes with having a fully adjustable multiplier with a 45nm processor will have to pony-up the dough for a QX9650 (or QX9770) or go without.
We recognize this change for what it really is - a bold move when it comes to fulfilling the needs of enthusiasts worldwide, considering how a vast majority of today's games and applications still favor systems with fewer high-speed cores over those with more cores at lower frequencies. Intel's decision to supply processors with unlocked multipliers under an "Extreme Edition" branding became an essential ingredient in the creation of all future roadmaps. Eventually these unique processors became the basis for a new class of computing platforms, one that embodied a shift in marketing philosophy. Rather than focus solely on serving the large OEMs, Intel also recognized the direct importance of the enthusiast community. We could argue that when it came to winning the admiration and approval of overclockers, enthusiasts, and power users alike, no other single common product change could have garnered the same overwhelming success.
Our love affair with the quad-core began not too long ago, starting with the release of Intel's QX6700 Extreme Edition processor. Ever since then Intel has been aggressive in their campaign to promote these processors to users that demand unrivaled performance and the absolute maximum amount of jaw-dropping, raw processing power possible from a single-socket desktop solution. Quickly following their 2.66GHz quad-core offering was the QX6800 processor, a revolutionary release in its own right in that it marked the first time users could purchase a processor with four cores that operated at the same frequency as the current top dual-core bin - at the time the 2.93GHz X6800. From there only a small default FSB speed bump from 266Mhz (1066 quad-pump) to 333Mhz (1333 quad-pumped) and a stepping change from B3 to G0 was all that was needed to justify the creation of the QX6850, which ran at a slightly higher speed of 3.0Ghz (9x333). Again, the X6850 matched the QX6850 in every way but one, that being that it had two fewer cores.
Writing multithreaded code that makes efficient use of four or more cores is a daunting task - to date few applications and even fewer game developers are able to boast of this accomplishment. Given this, is it that hard to admit that perhaps we've all been a little guilty of demanding too much, too soon from our favorite software vendors? It should not be surprising then to learn then that many of today's ultimate gaming machines make use of "lesser" dual-core CPUs in place of their quad-core counterparts. With most titles able to take advantage of only two cores at a time, optimum gaming performance (read: maximum FPS) is often achieved by running a dual-core CPU at a greater frequency than is attainable using even the best quad-core processors.
Because dual-cores can often be coaxed to run at a higher, final stable speed then quad-core CPUs - which also consume significantly more power - most modern games have been engineered to make use of no more than two threads simultaneously executing in parallel. These games thus benefit from the additional overclocking headroom of dual-core CPUs. Meanwhile, in the case of the quad-core processor, approximately half of the processing resources sit idle while the code executes on any two of the four slower cores.
If you're not an overclocker, aside from the obvious processor count increase from two to four cores, there is little difference between Intel's top-end dual-core E8500 and their QX9650 Extreme Edition quad-core CPU. Each is fabricated based on exactly the same underlying 45nm, second-generation Core 2 architecture. Both interface with their host motherboard's MCH at an equivalent quad-pumped FSB speed of 1333MHz. And technically speaking, on a by-core basis, each must contend for the same amount of shared Level 2 cache (6MB per die). The only real difference is their core operating frequencies - the E8500 at 3.16GHz (9.5x333) and the QX9650 at 3.00Ghz (9x333). Because of the raw speed advantage, if the target application or game only makes use of two cores then the E8500 ends up being the better choice.
This isn't to say that the quad-core CPU is left without the existence of a proper application - far from it. Programs that heavily rely on the impressive parallel processing capabilities of a quad-core processor can realize up to nearly double the per-clock performance. This is especially true of tasks that lend themselves to the use of multiple program instances. For example, consider an encoding program that makes use of only two cores. Running two instances, and simultaneously encoding two files, would effectively load all four cores. Of course, this assumes there is a work queue in which the next available job can be drawn from, without which no benefit could be realized. There are certainly applications where more cores is almost always better; whether you use those applications on a regular basis is the real question.
"Accurate" Temperature Monitoring?
In the past, internal CPU temperatures were sensed using a single on-die diode connected to an external measurement circuit, which allowed for an easy means of monitoring and reporting "actual" processor temperatures in near real-time. Many motherboard manufacturers took advantage of this capability by interfacing the appropriate processor pins/pads to an onboard controller, such as one of any of the popular Super I/O chips available from Winbond. Super I/O chips typically control most if not all of the standard motherboard input/output traffic associated with common interfaces including floppy drives, PS/2 mice and keyboards, high-speed programmable serial communications ports (UARTs), and SPP/EPP/ECP-enabled parallel ports. Using either a legacy ISA bus interface or a newer LPC (low pin-count) interface, the Super I/O also monitors several critical PC hardware parameters like power supply voltages, temperatures, and fan speeds.
This method of monitoring CPU temperature functioned satisfactorily up until Intel conducted their first process shrink to 65nm. The reduction in circuit size influenced some of the temperature-sensing diode's operating characteristics enough that no amount of corrective calibration effort would be able to ensure sufficient accuracy over the entire reporting range. From this point on Intel engineers knew they would need something better. From this came the design we see effectively utilized in every CPU produced by Intel today, starting with Yonah - one of the first 65nm processors and a precursor to the now wildly-successful Core 2 architecture.
The new design, called a Digital Thermal Sensor (DTS), no longer relied on the use of an external biasing circuit where power conditioning tolerances and slight variances in sense line impedances can introduce rather large signaling errors. Because of this, many of the reporting discrepancies noted using the older monitoring methods were all but eliminated. Instead of relying on each motherboard manufacturer to design and implement this external interface, Intel made it possible for core temperatures to be retrieved easily, all without the need for any specialized hardware. This was accomplished through the development and documentation of a standard method for reading these values directly from a single model specific registers (MSR) and then computing actual temperatures by applying a simple transformation formula. This way the complicated process of measuring these values would be well hidden from the vendor.
A few quick lines of code (excluding the custom device driver required) is all that is needed to quickly retrieve and report values encoded in an MSR for each core.
The transformation formula we spoke of is actually exceedingly simple to implement. Instead of storing the absolute temperature for each core, the MSR is designed to essentially count down the margin to the core's maximum thermal limit, often incorrectly referred to as "Tjunction" (or junction temperature). When this value reaches zero, the core temperature has reached its Tjunction set point. Therefore, calculating the actual temperature should be as easy as subtracting the remaining margin (stored in the MSR) from the processor's known Tjunction value. There is a problem however: Intel has never published Tjunction values for any CPU other than the mobile models. The reason for this is simple. Since mobile processors lack an integrated heat spreader (IHS), it is not possible to establish a thermal specification with respect to its maximum case temperature ("Tcase"), normally measured from an embedded probe located top, dead-center in the IHS.
Thus, calculating actual core temperatures requires two separate data points, only one of which is readily available from the MSR. The other, Tjunction, must be known ahead of time. Early implementations of the process used to determine a processor's particular Tjunction value by isolating a single status bit from a different MSR that was used flag whether the part in question was engineered to a maximum Tjunction set point of either 85ºC or 100ºC. Because Merom - the mobile Core 2 version of Conroe - used one of these two values, it was somehow decided that the desktop products, built on the same process, must also share these set points. Unfortunately, it turns out this is not the case.
All of these prominent means for monitoring Intel CPU core temperatures are based on assumed maximum Tjunction setpoints which cannot be verified.
More than a few programs have been released over the last few years, each claiming to accurately report these DTS values in real-time. The truth is that none can be fully trusted as the Tjunction values utilized in these transformations may not always be correct. Moreover, Intel representatives have informed us that these as-of-yet unpublished Tjunction values may actually vary from model to model - sometimes even between different steppings - and that the temperature response curves may not be entirely accurate across the whole reporting range. Since all of today's monitoring programs have come to incorrectly assume that Tjunction values are a function of the processor family/stepping only, we have no choice but to call everything we thought we had come to know into question. Until Intel decides to publish these values on a per-model basis, the best these DTS readings can do for us is give a relative indication of each core's remaining thermal margin, whatever that may be.
Determining a Processor Warranty Period
Like most electrical parts, a CPU's design lifetime is often measured in hours, or more specifically the product's mean time to failure (MTTF), which is simply the reciprocal of the failure rate. Failure rate can be defined as the frequency at which a system or component fails over time. A lower failure rate, or conversely a higher MTTF, suggests the product on average will continue to function for a longer time before experiencing a problem that either limits its useful application or altogether prevents further use. In the semiconductor industry, MTTF is often used in place of mean time between failures (MTBF), a common method of conveying product reliability with hard drive. MTBF suggests the item is capable of repair following failure, which is often not the case when it comes to discrete electrical components such as CPUs.
A particular processor product line's continuous failure rate, as modeled over time, is largely a function of operating temperature - that is, elevated operating temperatures lead to decreased product lifetimes. Which means, for a given target product lifetime, it is possible to derive with a certain degree of confidence - after accounting for all the worst-case end-of-life minimum reliability margins - a maximum rated operating temperature that produces no more than the acceptable number of product failures over a period of time. What's more, although none of Intel's processor lifetimes are expressly published, we can only assume the goal is somewhere near the three-year mark, which just so happens to correspond well with the three-year limited warranty provided to the original purchaser of any Intel boxed processor. (That last part was a joke; this is no accident.)
When it comes to semiconductors, there are three primary types of failures that can put an end to your CPU's life. The first, and probably the more well-known of the three, is called a hard failure, which can be said to have occurred whenever a single overstress incident can be identified as the primary root cause of the product's end of life. Examples of this type of failure would be the processor's exposure to an especially high core voltage, a recent period of operation at exceedingly elevated temperatures, or perhaps even a tragic ESD event. In any case, blame for the failure can be (or most often obviously should be) traced back and attributed to a known cause.
This is different from the second failure mode, known as a latent failure, in which a hidden defect or previously unnoted spot of damage from an earlier overstress event eventually brings about the component's untimely demise. These types of failure can lay dormant for many years, sometimes even for the life of the product, unless they are "coaxed" into existence. Several severity factors can be used to determine whether a latent failure will ultimately result in a complete system failure, one of those being the product's operating environment following the "injury." It is accepted that components subjected to a harsher operating environment will on average reach their end of life sooner than those not stressed nearly as hard. This particular failure mode is sometimes difficult to identify as without the proper post-mortem analysis it can be next to impossible to determine whether the product reached end-of-life due to a random failure or something more.
The third and final failure type, an early failure, is commonly referred to as "infant mortality". These failures occur soon after initial use, usually without any warning. What can be done about these seemingly unavoidable early failures? They are certainly not attributable to random failures, so by definition it should be possible to identify them and remove them via screening. One way to detect and remove these failures from the unit population is with a period of product testing known as "burn-in." Burn-in is the process by which the product is subjected to a battery of tests and periods of operation that sufficiently exercise the product to the point where these early failures can be caught prior to packaging for sale.
In the case of Intel CPUs, this process may even be conducted at elevated temperatures and voltages, known as "heat soaking." Once the product passes these initial inspections it is trustworthy enough to enter the market for continuous duty within rated specifications. Although this process can remove some of the weaker, more failure prone products from the retail pool - some of which might have very well gone on to lead a normal existence - the idea is that by identifying them earlier, fewer will come back as costly RMA requests. There's also the fact that large numbers of in-use failures can have a significant negative impact on the public perception of a company's ability to supply reliable products. Case in point: the mere mention of errata usually has most consumers up in arms before they are even aware of the applicability.
The graphic above illustrates how the observed failure rate is influenced by the removal of early failures. Because of this nearly every in-use failure can be credited as unavoidable and random in nature. By establishing environmental specifications and usage requirements that ensure near worry-free operation for the product's first three years of use, a "warranty grace period" can be offered. This removes all doubt as to whether the failure occurred because of a random event or the start of its eventual wear-out, where degradation starts to play a role in the observed failures.
Implementing process manufacturing, assembly, and testing advancements that lower the probability of random failures is a big part of improving any product's ultimate reliability rate. By carefully analyzing the operating characteristics of each new batch of processors, Intel is able to determine what works in achieving this goal and what doesn't. Changes that bring about significant improvements in product reliability - enough to offset the cost of a change to the manufacturing process - are sometimes implemented as a stepping change. The additional margin created by the change is often exploitable with respect to realizing greater overclocking potential. Let's discuss this further and see exactly what it means.
The Truth About Processor "Degradation"
Degradation - the process by which a CPU loses the ability to maintain an equivalent overclock, often sustainable through the use of increased core voltage levels - is usually regarded as a form of ongoing failure. This is much like saying your life is nothing more than your continual march towards death. While some might find this analogy rather poignant philosophically speaking, technically speaking it's a horrible way of modeling the life-cycle of a CPU. Consider this: silicon quality is often measured as a CPU's ability to reach and maintain a desired stable switching frequency all while requiring no more than the maximum specified process voltage (plus margin). If the voltage required to reach those speeds is a function of the CPU's remaining useful life, then why would each processor come with the same three-year warranty?
The answer is quite simple really. Each processor, regardless of silicon quality, is capable of sustained error-free operation while functioning within the bounds of the specified environmental tolerances (temperature, voltage, etc.), for a period of no less than the warranted lifetime when no more performance is demanded of it than its rated frequency will allow. In other words, rather than limit the useful lifetime of each processor, and to allow for a consistent warranty policy, processors are binned based on the highest achievable speed while applying no more than the process's maximum allowable voltage. When we get right down to it, this is the key to overclocking - running CPUs in excess of their rated specifications regardless of reliability guidelines.
As soon as you concede that overclocking by definition reduces the useful lifetime of any CPU, it becomes easier to justify its more extreme application. It also goes a long way to understanding why Intel has a strict "no overclocking" policy when it comes to retaining the product warranty. Too many people believe overclocking is "safe" as long as they don't increase their processor core voltage - not true. Frequency increases drive higher load temperatures, which reduces useful life. Conversely, better cooling may be a sound investment for those that are looking for longer, unfailing operation as this should provide more positive margin for an extended period of time.
The graph above shows three curves. The middle line models the minimum required voltage needed for a processor to continuously run at 100% load for the period shown along the x-axis. During this time, the processor is subjected to its specified maximum core voltage and is never overclocked. Additionally, all of the worst-case considerations come together and our E8500 operates at its absolute maximum sustained Tcase temperature of 72.4ºC. Three years later, we would expect the CPU to have "degraded" to the point where slightly more core voltage is needed for stable operation - as shown above, a little less than 1.15V, up from 1.125V.
Including Vdroop and Voffset, an average 45nm dual-core processor with a VID of 1.25000 should see a final load voltage of about 1.21V. Shown as the dashed green line near the middle of the graph, this represents the actual CPU supply voltage (Vcore). Keep in mind that the trend line represents the minimum voltage required for continued stable operation, so as long as it stays below the actual supply voltage line (middle green line) the CPU will function properly. The lower green line is approximately 5% below the actual supply voltage, and represents an example of an offset that might be used to ensure a positive voltage margin is maintained.
The intersection point of the middle line (minimum required voltage) and the middle green line (actual supply voltage) predicts the point in time when the CPU should "fail," although an increase in supply voltage should allow for longer operation. Also, note how the middle line passes through the lower green line, representing the desired margin to stability at the three-year point, marking the end of warranty. The red line demonstrates the effect running the processor above the maximum thermal specification has on rated product lifetime - we can see the accelerated degradation caused by the higher operating temperatures. The blue line is an example of how lowering the average CPU temperature can lead to increased product longevity.
Because end of life failures are usually caused by a loss of positive voltage margin (excessive wear/degradation) we can establish a very real correlation between the increased/decreased probability of these types of failures and the operating environment experienced by the processor(s) in question. Here we see the effect a harsher operating environment has on observed failure rate due to the new end of life failure rate curve. By running the CPU outside of prescribed operating limits, we are no longer able to positively attribute any failure near the end of warranty to any known cause. Furthermore, because Intel is unable to make a distinction in failure type for each individual case of warranty failure when overclocking or improper use is suspected, policy is established which prohibits overclocking of any kind if warranty coverage is desired.
So what does all of this mean? So far we have learned that of the three basic failure types, failures due to degradation (i.e. wearing out) are in most cases directly influenced by the means and manner in which the processor is operated. Clearly, the user plays a considerable role in the creation and maintenance of a suitable operating environment. This includes the use of high-quality cooling solutions and pastes, the liberal use of fans to provide adequate case ventilation, and finally proper climate control of the surrounding areas. We have also learned that Intel has established easy to follow guidelines when it comes to ensuring the longevity of your investment.
Those that choose to ignore these recommendations and/or exceed any specification do so at their own peril. This is not meant to insinuate that doing so will necessarily cause immediate, irreparable damage or product failure. Rather, every decision made during the course of overclocking has a real and measureable "consequence." For some, there may be little reason to worry as concern for product life may not be a priority. On the other hand, perhaps precautions will be taken in order to accommodate the higher voltages like the use of water-cooling or phase-change cooling. In any case, the underlying principles are the same - overclocking is never without risk. And just like life, taking calculated risks can sometimes be the right choice.
Initial Thoughts and Recommendations
Although the release of Core 2 Duo E8000-series Wolfdale processors will never be as groundbreaking as the Conroe launch during the summer of 2006, Intel has done a fine job of building upon their already enormously successful Core 2 platform. Reductions in power consumption, improvements in energy efficiency (especially at idle), and the amazing overclocking capacity of these processor make them downright irresistible. Now that we've had a good chance to experience all that this new 45nm process has to offer, we will be hard pressed to buy anything else. Without a doubt, Intel has re-proclaimed their dominance in the marketplace of technology.
One of the most popular uses for processors of this caliber is 3D gaming. If excellent performance is already possible when running at default speeds, just imagine the stunning frame rates displayed when overclocking to 4GHz and beyond on air-cooling alone. (Ed: Assuming you have the GPUs to back that up, naturally.) Never before has achieving these levels of overclocks been so easy. However, don't become tempted by the incredible range of core voltage selections your premium motherboard offers; it's important not to lose sight of the bigger picture.
These processors are built on a new 45nm High-K process that invariably makes them predisposed to accelerated degradation when subjected to the same voltages used with last-generation's 65nm offerings. Although we certainly support overclocking as an easy and inexpensive means of improving overall system performance, we also advocate the appropriate use of self-restraint when it comes to choosing a final CPU voltage. Pushing 0.1V more Vcore through a processor for that last 50MHz does not make a lot of sense when you think about it.
More than a couple Penryn dies will comfortably fit on a single 300mm wafer used in the manufacturing of these processors.
Perhaps even more exciting than the prospect of assembling a game rig is the potential these processors possess for use in the HTPC arena. Some of our initial tinkering has allowed us to realize the hope of creating a complete home entertainment solution capable of full-load operation at less than 90W power draw from the wall. With the recent availability and - dare we say it? - affordability of Blu-ray drives in the US, the prospect of putting together an all-in-one multimedia powerhouse that runs both cool and silent is finally becoming a reality. Of the models soon to hit the shelves, the E8200 at 2.66GHz or E8300 at 2.83GHz are sure to be winners when looking for processors suited for just these types of low-power applications. Couple this with an Intel G35 chipset with integrated graphics, Clear View Technology, and onboard HD audio over HDMI and you have all the makings of a serious HTPC.
Intel has also worked hard to make all of this performance affordable. Many US retailers now stock the 65nm Q6600 quad-core CPU at less than $200, which places it squarely in the 45nm dual-core price range - something to think about as you make your next purchasing decision. However, if it comes down to the choice between a 65nm and 45nm CPU we would pick the latter every time - they are just that good. The only question now is exactly when Intel will decide to start shipping in volume.