Intel i925XE Chipset / 3.46GHz Pentium 4

..:: Introduction ::..

In the past few months, things have certainly changed direction in the Intel processor design houses. The latest change comes with Intel dropping the 4.0GHz version of the Pentium 4 in order to concentrate on Smithfield, their upcoming dual core processor for the desktop that is set to debut sometime in the mid to latter half of 2005. Certainly, it seems someone at Intel has seen the light and decided against advancing a processor design that has an achilles heel, it’s thermal characteristics.  

With the initial launch of the LGA 775 platforms and processors, there was a real lull in how fast the enthusiast crowd picked them up. At time of launch, there were few solid PCI-Express options, and those that were available were quite costly. There was a substantial amount of DDR-II that soon became available, and the prices seen were quite nice, though that wasn’t expected to last and it’s already making a rebound. Another problem was lack of any third arty cooling solutions. Basically, either you found an LGA 775 capable water block, or you were stuck with Intel’s standard cooler, which isn’t exactly a slouch, or something worse. As time has progressed, and more platforms and capable components have reached the market, the LGA 775 platforms have begun to garner more interest among upgrading Intel enthusiasts.

Another problem many had with the actually i925X and i915P/G platforms were their expensive premium. As we saw when these two platforms were launched, there’s only a minute difference in overall performance between the two. It certainly wasn’t enough to justify the extra costs for upgrading. As cheaper and better boards have hit the market, this too has helped business pick up for i925X and 915P/G motherboards. So, although things were slow at first as I expected, they’re beginning to pick up, and should really start to pick up when PCI-E graphics solutions become more the norm for upgrades over the several months to a year as AMD chipsets debut featuring support for PCI-E.

Today, we see the launch of Intel’s latest incarnation of i925X, the i925XE chipset. Now, as much as anyone would like to tell you, whether it be through marketing or salesmanship, i925XE is nothing more than a glorified i925X with added support for 1066MHz FSB Pentium 4’s, the first of which is also being launched today and you’ll see in action soon. The initial i925X chipset had a built in overclocking lock by our friends at Intel that made 1066MHz a nearly impossible feat even with motherboard manufacturers finding ways to circumvent the problem. Obviously, both these boards and processors will carry a heavy premium, but the real question is, as always, is it worth the extra cash? Well, before we try to answer this question for our adoring public, I’ll be re-hashing the info on the i925X and now i925XE chipsets, as well as covering the thermal specs for the 3.46GHz Extreme Edition CPU. If you already know all of these details, skip ahead to the benchmarks to find what you’re looking for.

 

 

..:: 3.46GHz EE – Thermal Design Power ::..

As we’ve seen time and time again over these past few processor releases from Intel, that three letter acronym that everyone in the computing industry dislikes, TDP, has been growing and growing. The newer “Prescott” cores have been the real culprits putting off tremendous amounts of heat. The older Northwood core Extreme Edition CPU’s have been much better, in our experiences, with keeping cool, but their time will soon come. This new 3.46GHz Extreme Edition CPU carries a TDP of 110.7W, a slight increase over the 3.40GHz EE. The Icc max for the 3.46GHz model is a cool 84.8A. To put these numbers in comparison to similar models, the 3.40EE boasts a TDP of 109.6W, while the 3.40E (550) and 3.60E (560) sport a TDP of 115.0W. The max current draw for the 3.40E and 3.60E processors is a massive 119 Amps, up 41 Amps from the 3.20 and below speed grades.

 

..:: Processor Physical Overview ::..

The processor itself is identical to the other LGA 775 processors we’ve seen to date, slightly larger than the Socket 478 Pentium 4’s featuring a slightly modified heatspreader. The heatspreader on the Socket 478 implementation goes to roughly 1mm from the edge of the package, and then drops down to the surface. On the LGA Pentium 4’s, the heatspreader goes out a similar distance, although the outer edge is lowered down to allow for the outer casing that holds the processor in place once it is mounted. On the bottom of the package, we find all 775 of the gold contacts, along with the various resistors, and capacitors that are surface mounted underneath the processor die. This package looks quite similar to the 478 solution, only larger and pin-less.

Intel EM64T Technology

..:: Introduction ::..

Intel’s Extended Memory 64 Technology, better known under the simple acronym EM64T, is Intel’s version of x86-64. The processors that feature support for EM64T are backwards compatible with 32-bit systems, and also capable of taking advantage of the added memory and computational capabilities of 64-bit processing. EM64T is built off of the IA-32 architecture, with additional registers, instructions, and otherwise enhanced instructions. Processors that feature EM64T are also capable of running under multiple operating modes, depending on various system setup aspects, i.e. 32-bit Vs. 64-bit OS. These processors are also fully compatible with existing software meant for the original IA-32 architecture.


Figure 1.

 

..:: Operating Modes – IA-32e ::..

With the addition of EM64T, Intel has added a new operating mode referred to as IA-32e. This operational mode also includes two sub-modes, compatibility & 64-bit. The IA-32e mode can only be entered if the processor is working off of a 64-bit capable operating system, such as Microsoft’s upcoming WindowsXP x64 Pro. If it were to operate under a standard 32-bit environment, i.e. WindowsXP, the processor would run in IA-32 legacy mode. In IA-32 legacy mode, the processor could be run in the three common sub-modes, those being Real Address, Protected, & Virtual 8086.

 

..:: Operating Modes – Compatibility ::..

When an EM64T capable processor operates under compatibility mode, it allows the bulk of legacy 16-bit and 32-bit applications to run without any need to be recompiled under a 64-bit environment. Under most circumstances, software developers would need to recompile their code with the latest version of Intel’s compiler, or another 64-bit capable compiler, to make their software take full advantage of the 64-bit capabilities. Those software developers who wish not to do so don’t risk having their software become incompatible due to this operational mode.

Compatibility mode is controlled by the operating system, and is controlled on the basis of code segments within an application. This is the foundation of the 32/64-bit support architecture as both 32-bit and 64-bit applications could be running at the same time. Applications run under compatibility mode have limited access to the first 4GB of linear-address space, and are only able to utilize standard IA-32 instruction prefixes and registers. The operand and address sizes are also limited to the IA-32 standards of 16-bit and 32-bit.

 

..:: Operating Modes – 64-Bit ::..

The final sub-mode of EM64T is 64-bit mode. As one would likely assume, 64-bit mode is utilized by 64-bit applications when they’re run under a 64-bit operating system. Intel has made several key changes to the IA-32 architecture to allow for these 64-bit applications, such as adding support for 64-bit linear addressing. Linear addressing is a scheme that allows access to the entirety of memory with use of a single address, usually loaded in a register or instruction. Variations of the IA-32 architecture may not offer full 64-bit linear addressing, an example being the current 600 series Pentium 4 processors which only allow for 48-bit linear addressing.

The main additions to the IA-32 architecture lie with the need for additional 64-bit registers. Intel has added eight new general purpose registers, along with eight 128-bit Streaming SIMD Extension registers. The pre-existing general purpose registers have all been widened to 64-bits as well. With these new registers comes a new opcode prefix, REX. The defaults for 64-bit mode allow for a 64-bit address size, and a 32-bit operand size. These defaults can be overridden by using the REX opcode prefix. The REX prefix allows for a 32-bit operand to be chosen when operating under 64-bit mode. It is utilized on an instruction-by-instruction basis, and therefore only called when needed. In order to allow for support of these 32-bit operands under 64-bit mode, many of the pre-existing instructions have been changed, or redefined in order to utilize the 64-bit registers and 64-bit addressing. These modified instructions are where the REX prefix is utilized.

In order to utilize the 64-bit instructions you will, of course, need a 64-bit instruction pointer. This instruction pointer is extended to 64-bits wide when operating under 64-bit mode, and contains the address of the next instruction to be executed. Without a 64-bit wide instruction pointer, the newly added instructions would be useless. Intel has also added an additional addressing mode, relative to the current address stored in the instruction pointer. This new addressing mode allows for the address stored within the instruction pointer register to be used as a base address for selecting other memory addresses. Relative addressing is typically used where previous address increments are of a known value, or where various address locations are broken up into different segments.

The remaining features that are supported under the 64-bit operating mode are the use of flat address space with single code, data, and stack space, uniform byte-register addressing, a new interrupt priority control mechanism, and of course, support for greater than 64 GB of physical address support. This is one of the, if not the most important aspect of 64-bit capable processors right now. With 64-bit bit physical address support, you’re talking about over 200 TB of memory that could potentially be addressed, far more than the current maximum of about 4GB, (some processors can currently address greater than 64 GB, but generally the maximum allowed is 4GB).

Intel 6xx & 3.73GHz Pentium 4

..:: Introduction ::..

At the time of AMD’s Athlon 64 release, Intel was seen downplaying the importance and advantages of a 64-bit setup to the average consumer, basically saying that they would supply a 64-bit capable processor when the time was right. Well, with the impending release of Microsoft’s WindowsXP x64 Professional Edition OS, it seems that now is the right time.

I doubt that anyone out there expected to see Intel launch their 64-bit processors at any other time, waiting in the background until Microsoft was ready to launch their 64-bit WindowsXP OS. Some might look at this in a conspiratorial light; in fact I wouldn’t be surprised if several of you reading this article fell into that category. We’ll leave these thoughts to others, as today we look to examine the Intel 6xx Pentium 4 “Prescott 2” line of processors, as well as the new 3.73GHz Extreme Edition, also based off of the “Prescott 2” core.

Today, we’ll be taking a look at Intel’s EM64T technology, as well as performance under the latest beta release of Microsoft’s WindowsXP x64 Pro. Intel’s EM64T processors have a few changes made to the architecture, namely with registers, the stack, and instructions. What do we think about these new processors? How do they compare versus older “Northwood” core Extreme Edition CPU’s? Read on to find out.

Intel 3.40E & 3.40C Pentium 4

..:: Introduction ::..

A little over two months have passed since the rather quiet release of Intel’s latest Pentium 4 Processor core, code-named “Prescott”. The “Prescott” chip was actually released several hours earlier than the typical Intel release times, and ended up falling on Super Bowl Sunday. With all of the build-up of bad karma following “Prescott” before launch with rumors of delays due to heat output, outrageous TDP numbers, and more, it came as no surprise that Intel would attempt to launch the “Prescott” core quietly, and this fact can be seen by the fact it was launched early, on a Sunday, on Super Bowl Sunday. The rumors about the high TDP numbers for the “Prescott” core turned out to be true, and we noted some high temperatures during our testing period with the 3.20E processor. In comparison to the 3.20C, we found the 3.20E to be a bit lacking in performance, and a little heavy on the heat output. Today, we’re going to take a look at the recently released 3.40C and 3.40E to see if the added MHz can help “Prescott” catch up some more to the “Northwood” processor. We’ll also delve further into the heat issue with some interesting temperature measurements to help illustrate it.

 

..:: 3.40C & 3.40E Official Specifications ::..

 

..:: “Prescott” Core Heat Issues ::..

As we saw with the official TDP ratings for the 3.40E and 3.40C processors, the 3.40E is going to put out a substantially larger amount of heat, much due to the increased current needed to run the processor, especially at full load, along with the increased current through the same number of power pins in the processor package, and of course, the more than doubling of transistors in the processor core. It is no secret now that “Prescott” is quite the hot chip, and that those looking for quiet cooling are going to want to stray away from the “E” processors until Intel can find a way to “solve” the heat output problems. Why do I put solve in quotations? Simply put, the amount of transistors and current needed to run the processor pose a serious problem. Intel needs to develop further manufacturing and design techniques in order to advance in the race for a cooler running chip. There is no simple overnight cure for this type of problem, so for all of you looking for Intel to come up with some magic cure than takes the TDP down a good 20W, you need to come back to reality.

For now, we’re stuck with a chip that runs quite warm, and several motherboard manufacturers have begun to take a less than positive stance against it as well. Soltek recently released their VIA PT880 based motherboard that we reviewed, and informed us that they would only officially support the “E” processors up to 3.00GHz. Their reasoning? Heat, Heat, Heat. MSI is also widely known to disallow VCore adjustment when it comes to the “E” processors due to their dynamic VID feature. I have discussed this heat issue with several first and second tier manufacturers, and nearly all of them have stated their unease with the heat levels of “Prescott.” The problem with heat is the effect it has on the core components of the board itself. The excess heat puts undue strain on capacitors, transistors, resistors, etc. Each of these electrical components reacts in unfriendly ways to increased heat, and can have their life spans shortened due to it. While we have yet to experience any instabilities due to the heat that is output by the “Prescott” processor, even with a stock Intel cooler which by the way performs quite nicely for the noise level, they cannot be ruled out for some other possible problems that could creep up with the increased heat output and strain placed on the motherboard components.

What kind of heat levels are we talking about? Well, we decided to try to get a relative temperature difference between two identically clocked “Northwood” and “Prescott” core processors to see just how much extra heat “Prescott” is putting out. In order to do this, we placed both chips in a cased situation, and put them in a bit of a torture state with no incoming or outgoing fans for airflow. While this can’t mimic the worst situation, it certainly won’t make things easy for either processor to run cool. In order to give you more accurate results as to the performance many of you will see, we chose to use the stock cooler than Intel provided for the 3.20E and 3.40E samples. While this cooler may look simplistic, it is capable of providing some excellent cooling for both the weight, and noise level. In order to acquire temperature readings, we took a thermal probe from one of our spare 5.25” drive bay devices. We cut down the excess plastic material around the tip of the thermal probe, and placed the tip of the probe against the side of the Pentium 4’s heatspreader. While this won’t be the hottest place on the chip given that by the time the heat energy has reached this point it will have lessened to some degree, it is still able to give a relative ratio of temperatures between chips. The results we achieved for our 3.40GHz models can be seen above. With an ambient temperature of 22.8C (73.0F), we found the “Northwood” core to run 8.3C (15.0F) cooler at idle, and 9.5C (17.1F) cooler under full load. These numbers easily illustrate the heat output of the “Prescott” core.

Intel Core Microarchitecture

..:: Introduction ::..

It’s been quite some time since this much excitement and enthusiasm has swirled around the release of an Intel desktop processor. Sure, advancements along the Pentium 4 development route, like the move to dual core architecture, have yielded some excitement, but nothing along the lines of what has been seen as of late.

In the past all we had to look forward to from Intel was a processor that provided a decent performance boost, and consumed more power than before. AMD’s processors on the other hand offered more performance, and used far less power. In a world of ever increasing energy costs, AMD’s offerings have become more attractive to both industry, and to the general home user. Intel set out to develop an advanced microarchitecture to address these very issues, and with that research and development came the Intel Core microarchitecture.

 

..:: Intel Core Microarchitecture ::..

Intel’s Core microarchitecture is the new basis for their mobile, desktop and server processors. Intel optimized this new microarchitecture to deliver the most performance at the least power consumption level, a.k.a. the “performance-per-watt” phrase that Intel has been throwing around for several months. The Core microarchitecture takes several of the key microarchitectural benefits developed for the Pentium M, and melds them with proven desktop microarchitectural features to provide a high performance, low power product.

In the past, Intel operated on the common notion that all that mattered in the world of performance was the operating frequency of a processor. This was the way that the vast majority of consumers evaluated a processors “performance” versus a competing product, and Intel used this to their advantage. AMD was the first to address the operating frequency myth by toting the increase in instructions executed per clock cycle with their processors. While true that IPC is vital to describing the performance of a processor, it alone does not paint a complete picture. Thus, true performance is both a function of the operating frequency and the IPC. Intel’s Core microarchitecture addresses these issues with lowered frequency, but a much improved IPC.

Now, to optimize the new microarchitecture in terms of power, one key dynamic that Intel looked at is the “dynamic capacitance required to maintain IPC efficiency”. Dynamic capacitance, as defined by Intel, is the ratio of electrostatic charge on a conductor to the potential difference between the conductors required to maintain that charge. In other words, this is the switching capacitance of the transistor setup. Power is equal to the dynamic capacitance value times the square of the operating voltage times the frequency. In order to balance the power consumption in the Core microarchitecture, the designers took all of these vital values into account to find the best balance for IPC, Voltage, Frequency, and Capacitance. Intel developed five innovations for the Core microarchitecture to address both of the power and performance goals.