The long road to Conroe

Posted by Scott M. Fulton, III

Santa Clara (CA) - History may credit AMD almost as much as Intel with the creation of the Woodcrest, Conroe, and Merom architectures, the latter two of which Intel released today. In 2004, AMD effectively threw down a very sizeable gauntlet, challenging Intel on every conceivable front, including retail outlets and courtrooms, to innovate or perish. In good time, Intel responded, with a marketing program and technology initiative that many felt was contrived, inconclusive, lacking vigor.

Intel's Israel design team spelled out for the company the extent of what it had to do to meet AMD's challenge: The NetBurst architecture was a physical and thermal dead end. Intel had to stop, repurpose itself, and build the most convincing low-power, high-performance architecture in history, from scratch. It would be an impossible task for any other company. Whether Intel has actually succeeded will depend on the next few months of extensive user testing. But we've seen the power of Intel's new launch vehicle, and the capability for it to meet and probably exceed anything AMD ever anticipated, appears to be there.

Here, from the pages of Tom's Hardware Guide and TG Daily, is how we've covered the last few years of events leading up to today's historic launch:

26 March 2004: Intel Conroe means goodbye to Pentium 4

The first word of Conroe's existence came over two years ago, with this bulletin carried on what was then called the Tom's Hard News page:

March 26, 2004 - 11:07 EST

Intel Conroe means goodbye to Pentium 4...A fresh roadmap from Hiroshige Goto on PC Watch mentions Conroe as a codename for the first time. Conroe is the replacement next generation desktop chip, which builds in all sorts of super features.

21 February 2005: Intel launches Pentium 4 600-series

The fact that Intel was lagging behind AMD technologically started to become apparent in early 2005, when the company finally took steps to make x86 architecture available on a 64 bit platform. At one time - in defense of the need for Itanium architecture - the company had said a 100% compatible 64 bit x86 architecture would be impossible. AMD saw that as a gauntlet thrown down with its name attached. It took that challenge and ran with it.

Intel today released its first 64 bit capable desktop processor. The new Pentium 4 6xx-series adds more performance as well as Enhanced SpeedStep, a feature that throttles clock speed and can reduce power consumption of the chip.

...

The big news of the introduction is the arrival of EM64T. About two years after AMD's first 64 bit desktop processors, Intel now follows with a very similar extension that allows users to run 32 bit and 64 bit applications on the chip without taking performance hits in the 32 bit space as it was the case with Intel's IA-64 approach. Later this year, Intel will include EM64T across it entire product line, including Celeron D chips and specially marked Pentium 4 5xx versions and clock speeds of at least 3.0 GHz.

3 March 2005: Intel aims to decrease power consumption of desktop processors to 60 watts

Here was the big problem with NetBurst: You couldn't ramp up the gigahertz without creating a toaster oven.

In a recent article, we reported that Intel is pushing the power envelope with its desktop dual-core processors to a new record level of 130 watts, raising concerns of additional cost for users to control heat and acoustics levels of a mainstream PC. As we learned at IDF, Intel's chip development will be turning the corner: The company intends to dramatically decrease power consumption to less than half of the the upcoming Pentium Extreme Edition 840.


The multicore duel begins

30 March 2005: Dueling Multicores: Intel and AMD fight for the future

At the beginning of 2005, Intel found itself having to integrate multi-core, 64 bit, and low-power technologies all at the same time, and it didn't yet have a clear roadmap to do so. So as a sort of stopgap measure, Intel introduced hyperthreading (HT) - a way to make a single core handle the workload of two cores. It did increase performance, and it was a fairly popular technology - the trouble was, it didn't lead customers where Intel wanted them to go.

Intel is now faced with the problem of whether HT truly is a stepping stone to its own approach to dual-core. Its challenges are these:

  • Despite Intel's good intentions to accelerate existing single-threaded software, the company is asking software developers to utilize compiler switches to help optimize their use of hyperthreading. If performance gains average only about 15%, however, there may not be enough incentive.
  • The fact that AMD uses shorter instruction pipelines - which means fewer available resources for distribution among multiple threads - had been touted by Intel as a design deficiency. But with AMD demonstrating a working x86 ("x64") dual-core Opteron, and touting benchmarks showing that Opteron outperforming Intel packages with HT turned on, AMD is characterizing AMD64 as not needing hyperthreading.
  • HT does not lead consumers on a direct technological path to Itanium 2 technology; in fact, a clear technological barrier exists between HT and Itanium 2's EPIC approaches to parallelism. Conceivably, both approaches may be technically capable of co-existing - in other words, theoretically, EPIC code can run in an HT CPU. But once thread division is made at one level, by either software or hardware acting as an agent, further division by the other agent becomes far more difficult, if not altogether unnecessary. So whether pairing EPIC with HT would lead to negligible performance gains, or perhaps even performance losses, is unclear at this point. Further hindering Intel's decisions are reports that have already been published elsewhere - whether through a real Intel leak or a misunderstanding by reporters - that HT will be integrated into future versions of Itanium 2. At the time this article went to press, however, multiple Intel sources declined comment on this specific issue.
25 May 2005: Dothan over NetBurst: Is the Pentium 4 a dead end?

The Pentium M - a mobile processor, not a desktop or server CPU - became the forerunner of all of Intel's Core Microarchitecture. It solved the problem of power consumption when applied to a mobile platform, so why couldn't it do the same thing for a desktop platform? The team at Tom's Hardware Guide tackled that problem, and saw the truth perhaps before anyone else did:

Read the full story ...

Let us try to sum up the insights we have gained during the course of this little project...With the help of a simple socket adapter card and a BIOS upgrade, certain mainboards using Intel's 865/875 chipsets can be upgraded to use a Pentium M instead of a Pentium 4. Such a system offers up-to-date performance paired with low power requirements.

Additionally, we were able to raise the FSB from 133 to 160 MHz without any trouble at all. The result was that our 2.13GHz Pentium M 770 ended up running at 2.56 GHz! At this clock speed, our two year old platform was able to beat the processor heavyweights Athlon 64 FX and Intel Pentium 4 Extreme Edition in all 3D games!

8 August 2005: Intel to trim power consumption of desktop CPUs in half by H2 2006

It was just under a year ago that Intel's real battle plan for the code-name it had leaked out over a year earlier, would finally come together. Unlike Microsoft, Intel knew that the second half of the year meant July, not December...or the following June.

Intel promised us in earlier conversations that its engineers are working on reducing the power consumption of its desktop processors. But it was unclear when this will happen. Documents seen by Tom's Hardware Guide now indicate that the new processor architecture code-named "Conroe" and scheduled for the second half of 2006 will deliver on this promise. If we believe our sources, then Intel is targeting a power consumption of about 60 to 70 watts per processor - or 30 to 35 watts per core.


Conroe is for real

12 August 2005: Intel to unveil Pentium 4 successor at IDF

By the time Intel made the news official, Tom's Hardware Guide readers were already familiar with most of the facts.

The current architecture named "NetBurst" was launched with the Pentium 4 ("Willamette") in 2000 and is quickly nearing the end of its life. With power consumption at record levels, Intel has no room to increase clock speeds and performance levels in the current Pentium 4 500/600 and the 90 nm Pentium D 800. The 65 nm part Pentium D 900 ("Presler"), the last NetBurst Pentium, will launch in Q1 2006 with slight increases in clock speeds.

16 August 2005: Intel to sample Conroe processor in Q1 2006

For the first time in years on the part of any technology company, high-level customers would be treated to a product launch schedule that was not only met, but beaten.

With Yonah and Presler almost out the door, Intel's development focus in the consumer and business segment begins to shift to the next generation desktop processor architecture. The company already operates bootable systems based on Conroe processors and will initiate the sampling phase in early 2006, Tom's Hardware Guide has learned.

About one week before Intel plans to reveal more details about the future of its desktop products, news reached our offices that Intel may be much further along in the development process of its next generation desktop architecture than commonly believed.

According to sources, Conroe is up and running with multiple operating systems in Intel's validation labs and has entered the debugging process. Assuming that Conroe will run in a similar 9-month validation time frame as the current Pentium D 800 (Smithfield), Intel may target a September/October 2006 introduction date.

25 August 2005: In the heat of a 'dual-core duel,' AMD responds to Intel's Conroe

As was becoming common practice at or around the time of IDF, AMD was launching a full-court press assault on the Intel news, daring that company to make good on its pledges, and fully ready to capitalize on the slightest misstep. Intel, it turned out, was betting on this strategy from AMD, and therefore gave AMD less new information to chew on than many had expected.

Intel's move - announced Tuesday at IDF - away from its own NetBurst architecture and towards shorter pipelines, is seen by some as another strategic victory for AMD, which was first to the finish line with 64 bit x64 architecture. At the same time, not having Intel's big pipelines to aim at ends up giving AMD one less Intel target to shoot down. So when questioned about Intel's having come around to seeing things AMD's way, [AMD server and workstation product manager Brent] Kirby changed the subject quickly, like a presidential debater reorienting a panelist's question. "Even if [Intel is] moving to a shorter pipeline," he said, "they're still in a trap with the front-side bus. That front-side bus is going to haunt them for as long as they keep that. They're going to need to make some efforts to take care of that issue."

In his first keynote address to IDF as CEO of Intel, Paul Otellini announced his company is recalibrating its performance indicators around a "performance-per-watt" scale, which may replace megahertz as Intel's key power factor. Kirby countered by saying his company's Opteron processors already meet Intel's power targets, having originally been launched at 89 watts, and with successive HE and EE units launched for 55 and 30 watts, respectively.

4 December 2005: Top-secret Intel processor plans revealed

Come December, we knew for certain that Merom - the mobile technology inspired by Yonah, which became Pentium M - would be the crux of Intel's future efforts. What would save power and increase performance in small spaces, would do the same for larger ones.

Read the full story ...

All chips based on the Merom design will use 65 nm technology and are expected to outperform the competition both in raw speed and raw speed per Watt.

...The introduction of the Merom design will be a turning point in Intel's product policy, because it will be the backbone for all processor families that go into the desktop, the mobile or the enterprise space. In contrast, the desktop and enterprise markets are provided with Pentium 4 and Pentium D NetBurst architecture processors while the mobility CPUs are derived from the more efficient Pentium M design. At this point we should also mention that all processors currently shipping out of AMD's Fab 35 facility in Dresden, Germany, are already based on one single processor design. Still AMD has not yet been able to transition to either a 65 nm production process or 300 mm wafer manufacturing.


The empire strikes back

13 March 2006: IDF Spring 2006: Will Intel's Core Microarchitecture close the technology gap?

To make Core Microarchitecture deliver on its performance promises, it had to be able to split its instruction sequences out of order in such a way that it didn't physically generate heat in doing so. This was a monumental task, the scope of which was finally revealed at last March's IDF:

Read the full story ...

Core Micro Architecture is an out-of-order design with which individual instructions are scheduled and staggered in a 14-stage pipeline. In order to increase instruction efficiency, Intel focused on improving the flexible instruction execution. While that sounds easy, it conflicts with the requirements of IA machines to have a clean memory ordering for the sake of adhering to program semantics. One easy example is that store operations need to be completed prior to loading data, because you would want to access the current (latest) dataset.

Executing more instructions at the same time was also achieved within the three ALUs (Arithmetical Logical Unit), which can process SSE instructions in a single cycle (128 bit wide SSE). In addition to that, L2 cache improvements, thanks to the shared design as well as new prefetchers that work on the basis of memory disambiguation (prefetch data that is not going to be modified by other queued instructions), help to feed the pipeline more efficiently.

7 May 2006: Intel's Conroe, Merom get a name: Core 2 Duo

In another break with the past, Intel didn't make up a new word for its official brand names. Besides, after "Pentium," could the company really risk "Sexium?"

Read the full story ...

The new processor name builds on the Core brand, which was introduced with the current "Core Duo" mobile processor generation in January of this year. According to the company "Core 2" is a sign for a second generation of "Core" technology, which may be a bit confusing, as the current "Core Solo" and "Core Duo" processors are built on technology that has been carried over from previous "Pentium M" processors and the new "Core 2" generation is built on a completely new architecture (which is named "Core" as well.)

14 July 2006: Tom's Hardware: Core 2 Duo smokes AMD's Athlon 64 X2

The first sign of vindication for Intel's strategy came just two weeks ago, with the confirmation from Tom's test engineers that Core 2 Duo had absolutely reclaimed the performance crown from AMD, and not by a hair or half a car length either.

Read the full story ...

Performance jumps drastically in some scenarios: For example, the chip breaks the 3000-point barrier in the CPU test of 3DMark06 for the very first time, approaches 9000 points in the CPU test of PCMark05 and almost 32,000 points in the SiSoft Sandra 2007 CPU Test - 58% more than AMD's fastest processor at 2.8 GHz.

Especially interesting is the way how Intel achieves this new level of processor performance. Tom's Hardware found that its Conroe system consumed less power than a comparable AMD system and up to 30% less power than a Pentium EE 965-based computer. The 18% clock speed increase of the overclocked version resulted in a relatively modest 7% increase in overall system power consumption.

AMD's challenge has been met, and in indisputable terms has been exceeded. The fact that AMD didn't meet Intel's performance challenge response day-and-date, in the manner of its all-out assaults at IDF, leaves some observers puzzled. But even the worst-case scenarios from analysts characterize AMD as down, but not out. AMD is a company whose architecture is designed to respond to change, and its response could be huge. If it isn't, just that fact alone will itself be huge. At any rate, this duel is far from over.

Core 2 Duo in detail: Past, present and future (22 images) ...

Complete coverage:
TG Daily interviews Intel: "Core is changing the game"
Intel is back: Core 2 Duo launches
Up to $16,000: Core 2 Duo computers flood the Net
Official: Intel releases Core 2 Duo, Core 2 Extreme
Tom's Hardware: Core 2 Duo smokes AMD's Athlon 64 X2
Intel to launch Merom, Conroe on Thursday
Four AMD dual-core prices now at or near Intel price/performance curve
Technology Background: Will Intel's Core Architecture Close the Technology Gap? (Tom's Hardware)