Vr-Zone had the chance to test one of the unreleased Core 2 Extreme QX6850 based on the new stepping, and the results couldn't be more encouraging.

The latest incarnation of the Kentsfield processor together with an EVGA Rev A2 vintage 680i SLI Motherboard reached 4792MHz. With just 1.55v, it ran through 3D Mark CPU Test with phase change cooling. Upping the voltage to 1.65v allowed them to score 5133MHz! "[...] looking at this, probably 3D Mark stable at 5GHz with LN2 cooling is quite easy."

News about AMD's Barcelona place it around 2 Ghz in September (probably October). The breach is widening.

Choosing your Core 2 Duo is not an easy task anymore. You have the 6xxx series with 4 MB of cache and a FSB of 1066 Mhz (not the 6300 and 6400, please). Down the ladder you may buy a 4xxx processor with 2 MB and 800 Mhz. Or you can grab an even cheaper 2xxx CPU (named Pentium) with just 1 MB and the same 800 Mhz FSB. All that without even counting one-core Celerons or four-core Core 2 Quads. And now there is a new member in the family: 1333MHz FSB processors that will be available later this summer.

Both Anandtech and xbitlabs have tested the E6750 and have reached more or less the same conclusion when it comes to overclocking:  the new G0 core stepping (expected to boast better frequency potential) gains a few hundreds Mhz more. It goes to show that Intel's 65nm process is as mature as it will get, at this point. Remember that this same year, they will be introducing the first 45nm processors.

At stock voltage with a retail Intel heatsink/fan Anandtech was able to run the system at 3.68GHz (460 x 8.0), while xbitlabs used a Zalman CNPS9700 LED and raised Vcore to 1.5V, hitting 3.69GHz. More or less the same, right? The reason is that the CPU goes up all the way until the FSB Wall stops it from going any further. The increase to a 333MHz FSB frequency may prevent some new Core 2 Duo to overclock to their limits. Those frequency multipliers are just too low. On the other hand, the Core micro-architecture was not in need of a faster bus (quad-core processors apart). Quite a mixed bag.

Three weeks ago, AMD launched the new low-power Athlon X2 BE processors. Since they were nothing more than a rebranding of the (totally unavailable) Athlon 64 X2 3800+ 35W TDP, I just thought: boring, boring.

With a slightly higher 45W TDP rating, the BE-2350 and the BE-2300, were intended to replace the old 3800+ and ensure availability. Know what? Today, you still can't find them at major e-tailers like Newegg. That's not so boring anymore.

AMD says parts are available in Korea and Japan and to expect to see these processors for sale at Newegg "in about a week or so."

Having lost the performance race, AMD continuosly promises low-power CPUs in volumes it can't deliver. They look good on paper, though.

Link: The Tech Report.

One area where AMD has been able to hold the upper hand on is the business of promoting their graphics processing units as general-purpose processors. First was the Folding@Home program last September. More recently, during the R600 launch, they spent a lot of time looking at applications for their GPU other than just graphics and added features in the GPU to aid future GPGPU projects.

Maybe not having to compete with Intel made them feel rosy, but now NVIDIA is announcing their progression into the world of high-performance computing (HPC): Tesla.

HPC is not your average piece of software, coding for multi-threaded processors is much more difficult than serial processing, which means that only tasks that are multi-threaded in nature find the bottleneck in hardware. NVIDIA's Tesla wants to be the answer for those problems. HPC applications require much more processing power because they access immense amounts of data simultaneously; for expamle: simulation of neural circuits, computational microscopes, seismic and reservoir simulation, cell phone design... There are other applications for HCP in researching electromagnetics, energy, biomedical, pharmaceutical, industrial, financial and military operations.

What does it have to do with you? Not much at the moment, but this is the beginning of a transition from clock speed to multi-core power. Consumer level programs will gradually take advantage of more than a couple of cores, and prices for solutions like Tesla will go down (right now they range from $1500 for a GPU card to $12,000 if you want a S870 server).

Link: PCPerspective.

When it comes to testing overclocked hardware, there's a plethora of programs that claim to push it to the limits, but few are so good at testing CPU, memory and chipset at the same time. Prime95 is used to find Mersenne Prime numbers, but what you should care about is it's ability to expose bad CPU cooling (bad case cooling in general) or an excess in overclocking.

The torture test runs continuously, comparing your computer's results to results that are known to be correct. Any mismatch and you've got a problem! Well, at least that's what they say. In my experience, a 99.9% stable OC will give you an error in just a few minutes. Otherwise perfectly stable machines will fall sooner or later after hours of torture. Only standar settings (or the best of OCs) under good cooling conditions will pass with flying colors.

At last, Version 25.3 of prime95 brings (among other features) the option to run multiple LL tests within a single prime95 instance. There is no longer a need to copy the ".exe" to another directory so you can run multiple instances for torturing that multi-cpu machine of yours. The dialog box now lets you choose how many CPUs or cores you want to stress.

It took them quite some time, but here it is.

Maybe this is not a reaction to Asus's C90S, or maybe it is, who knows? In any case, on September 2nd, Intel plans on reducing the price of their mobile Core 2 processors by a significant amount. Here's the table as extracted from xtreview:

The Core 2 Extreme x7900 (2.8 GHz) that you see here will be presented at that same time (along six more models) and will allow to increase the processor frequency trough a multiplier increase (similar to the T7600G already found on the Dell XPS M1710). The x7800 (2.6 GHz) will be released in July 8th.

"Reduction in mobile core 2 Duo prices will stimulate higher demand for santa rosa platform, which was presented during May this year. To the end of the fourth quarter 2007 Intel plan to bring the portion of processors core 2 Duo (Merom) with 800 MHz system bus to 80% of entire mobile production. The processors merom with 667 MHz system bus will be removed from production during March 2008."

If the price of the x7800 comes down and BIOS permits to change voltages and FSB, we could have something interesting here. Otherwise I fell inclined towards the desktop-processor-in-a-laptop aproach Asus is endorsing.
 

    *      Dual Core Processor at 1.6 GHz Clock Speed  
    *      "Conroe" Core Architecture
    *      65nm Manufaturing Technology  
    *      64k L1 Cache (Per Core), 128 kB L1 Total
    *      1 MB (Shared) L2 Cache (Full Speed)
    *      800 MHz Front Side Bus Speed  
    *      Socket-775 Form Factor Design
    *      1.35V Default Core Voltage
    *      Supports Intel 64-bit Technology
    *      Supports SSE/SSE2/SSE3/Execute Disable
    *      Supports Intel Speedstep / C1E
    *      Includes Aluminum Alloy Heatsink/Fan

Hothardware has a full review of the Pentium E2140. Yes, you heard that right, the Pentium name isn't dead. Instead, Intel will be pushing it to the value sector where the Celeron resided. For the moment, it looks like the Celeron (which has a weak following) has some future as a single-core processor, but it may be replaced soon with an older (but better known) brand.

Starting at only $90, it holds up nicely against a Core 2 Duo E6420. To achieve that, you will have to perform some overclocking, but with such a low FSB, the chip is actually crying out loud for you to do so. And taking 3D gaming out of the picture will raise it's preformance to the level of a Core 2 Duo E6600 and beyond. If you aren't into overclocking, you can still benefit from lower temperatures and near silent operation.

The only real competition this CPU has comes from AMD, with some Athlon 64 X2 processors being as cheap as $55.

Q6600 at 266$!

Despite the Athlon X2, it wasn't until last year that prices got to reasonable levels and you could actually buy a dual core CPU. Lack of competition from Intel allowed AMD to maintain ridiculously high prices for a long time. Shortly after the introduction of the Core 2, prices on competing products began to moderate and reached very good price points. Eventually, Intel also adjusted initial pricing (a little).

2005 saw the advent of the dual core, but it took a year for them to be cheap. 2006 gave birth to quad cores, but they're still quite expensive. No longer. With competition from AMD coming to this market segment before year's end, Intel is prepared to launch a preemptive attack: a severe price cut.

Dailytech has a list of prices to be made effective on July 22. If you are thinking of upgrading to Core 2 or Quad Core, maybe you fancy a souped up PC in time for Crysis, we think it would pay to wait a couple of months and take advantage of the new price points.

It wasn't clear until now if Intel could hold true to their roadmap. The Conroe die shrink (Penryn) was supposed to be launched this year, but that timeframe seemed difficult to meet. From past declarations they slated it for late Q3 or early 2008.

In a post that's a complement to a recent one, Tgdaily confirms that all is going well. 45nm technology is ready and we should see the first CPUs no later than 2007. If you don't remember what Penryn'll bring to the table (apart from lower power consumption and higher clock speeds) here's a recap:

- More cache: 6Mb per two cores.

- Fast Radix-16 Divider and Super Shuffle Engine, what this nice names mean is: higher IPC.

- Intel SSE4 instructions.

- Higher bus speeds: 1333Mhz for the desktop version, actually.

What all this'll buy you is a 10% performance increase compared to current Core 2 running at the same clock speed. Factor the SSE4 optimizations that still are on the way and we could be talking about even better improvements in several tasks (video encoding comes to mind).

What is "tick-tock"? As Intel describes it:

"The principle of cadence is based on what Intel calls the tick-tock model of silicon and microarchitecture. This model delivers a common processor architecture across all volume market segments. Each tick represents the silicon compaction beat rate, and each tick has a corresponding tock representing the design of a new microarchitecture, delivered in approximately two-year cycles."

Didn't get any of this? Tgdaily has worded it a little bit better:  

"Intel has begun describing its product introduction strategy as “tick-tock”, which promises to bring a new micro-architecture in even years and a refresh (shrink) in odd years"

This is only a new pace if you compare it to the five-year cycle of late (Pentium 4, Core 2, Athlon, Athlon64). On the other hand, there were times when minor differences qualified as new processors (K6 III or Pentium III). It all depends on what you call a "new microarchitecture" or a "refresh".

A more conservative approach that doesn't go after a big step but gets satisfied with small leaps ahead. Understandable after a failure like Netburst, isn't it?