We’ve had time to get to grips with the amazing performance of Intel’s Core i5 series: the latest addition to what Intel calls the “high mainstream” segment. Unfortunately, the new chips have been priced just out of reach, and a lot of potential buyers have turned to AMD and their value-centric Athlon II range. While cheap quad-cores may have been AMD’s biggest selling point since the arrival of Intel’s Nehalem architecture, they’re pretty-much beated by Intel’s new chips, codenamed Clarkdale.

It's amazing what water-cooling can get you.

Clarkdale is nothing new, really, since all the details have been circulating online for months before its release. However, it heralds a new approach by Intel (and AMD, eventually) to bring all the important bits of a PC together on one chip:  the system-on-chip design, which integrates the CPU, GPU, and memory controller. We saw the beginning of this approach when Intel merged the memory controller with their Core i7 chips back in 2008, surpassing AMD’s design that has been out for years.

Clarkdale now adds the GPU to the mix, but leaves it on a separate 45nm-derived die, along with the memory controller. Some have blamed the decision to have two dies on one SOC for Clarkdale’s high memory latency, but its performance is rather good regardless. We’ll see this design being eschewed next year in favour of the new Sandy Bridge processors, which will integrate everything into the CPU on a 28nm process. But how does Clarkdale perform?

In most scenarios, we see the Core i3 530 overtake AMD’s Athlon II. It also beats Intel’s famous Q6600 by quite a margin, and hangs out with Intel’s Q9400 with ease. But wait… those are quad-cores, right? This is where we see the immediate benefits of the move to the 32nm process: where a low-end dual-core chip surpasses a quad-core in games. It’s confusing but rather marvellous for those who intend to upgrade, because this represents better value for money than Intel’s i5 750. Care to have your mind blown further? The new Clarkdale-based Core i5 661 comes within striking distance of the i7 920. It’s that good.

Image sourced from Anandtech. Click to read their comprehensive review.

But that’s not all: power consumption is down, encryption performance (for the paranoid) is improved as this is now a hardware-accelerated feature, and the onboard GPU is actually pretty decent. The X4500 chip was never any good for gaming or watching 1080p movies, but Clarkdale’s graphics performance often surpasses Nvidia’s 9400 and ATI’s HD4200 designs. It’s enough to run Batman: Arkham Asylum, which is very impressive. But the goodness doesn’t stop there.

Image sourced from Anandtech. Click to read their comprehensive review.

An increase of 0.1 volts allows you to overclock the 530 to 4.3Ghz on a stock Intel cooler. What’s more impressive is that on stock voltage, it can reach 3.8Ghz; water-cooling allows a maximum of 4.9Ghz, albeit with a huge voltage boost.

So, what can we take from this? It’s safe to say that Intel has stopped dragging its feet in the market, and will finally dominate the budget segment again thanks to Clarkdale and its mobile version, codenamed Arrandale.

Can anyone believe that the successful, and potent, LGA1366 is now one year old? Since 2008, we’ve seen five new chips for the LGA1366 platform, as well as the introduction of the i5 series of CPUs and the new LGA1156 socket. It seems as though Intel’s market research paid off: things are selling well and AMD is having a tough time finding market share and buyers, just the way Intel thinks it should be.

For users who first migrated to the Core i7, they had the choice of the i7 920 or 940, and the 965 Extreme which features an unlocked multiplier. The i7 920 is still considered the best performing mainstream chip on the market, and has sold well. Unfortunately, for those who have moved or are already on the LGA1366 socket, upgrades are tough because there’s nothing to upgrade to. Well, come the second quarter of 2010, there might just be something to look forward to.

Intel has been making a lot of fuss about its new family of processors, codenamed Gulftown. Developed on the new 32nm process, which Intel has been working on since they taped out the 45nm current-generation processors, and manufactured by TSMC, Gulftown will be the most overclockable, power-sipping, and quad-core crushing chip to date.

Absolutely beautiful, don't you think? And look at the core voltage.

Boasting six cores and 12 threads, 12MB of L3 cache and a full set of features including SMT, Turbo Mode, and support for Virtualization, the Core i9 is quite possibly the most extravagant processor ever created. Its debut is to be on the LGA1366 socket, and it is exclusively for LGA1366 owners. Intel currently has no plans for a six-core chip on the LGA1156 socket, and this further differentiates the two sockets and goes to show that Intel wants LGA1156 to fall into the budget and mainstream markets. Meanwhile, AMD fans have the benefit of using the six-core Rana on both AM2 and AM3 boards.

Not everyone will be able to afford an i9 at first. It is going to be marketed as an Extreme Edition processor with an unlocked multiplier, which just about means it’ll probably be around R12,000. It has been suggested that the first model will launch with a 2.4 GHz clock speed, and, clock for clock, it should be a good deal faster than the i7 920. Already some people have gotten hold of preview samples, and some have boasted overclocks of 6Ghz+, an astounding achievement with LN2.

Will it be worth upgrading to this new chip? I don’t really know. Intel has placed an NDA on any reviews or benchmark scores, but already the chip breaks all previous 3DMark Vantage records. But in any case, current Core i7 920 and up owners shouldn’t need to upgrade; their rigs are probably still performing admirably. For overclockers and people with too much money, the Core i9 would probably be a good buy, but supply is guaranteed to be short due to its current Extreme Edition status.

But will it run Crysis? Oh boy, you have no idea.

It used to be that selecting a processor was easy. AMD had their simple processor lineups, namely the Athlon, Duron, Sempron, and Opteron lines. Intel had the loyal Celeron, Pentium, and Itanium ranges, and these were simple and easy to understand as well.  They were all differentiated by clockspeeds, and things were easy enough to figure out. And then, something happened.

It's confusing enough already...

The Core series happened. Intel released the first Core chip to the market in the form of a bunch of mid-range chips under the Intel Core nomenclature in 2006. These were later joined by the Core 2 ranges (including Core 2 Solo and Core 2 Quad) as time passed. These no longer followed a standard that the public could follow and understand, because Intel used the same naming system on their mobile chips, confusing consumers even more when trying to make comparisons between desktops and their mobile counterparts. Meanwhile, AMD chucked out the Duron and Sempron ranges to make way for Phenom, keeping their naming system simple and easy to follow.

And now we’ve come to the new bully on the block, the Core i7. The name confused many because Intel didn’t immediately say why they chose this name for the chip, but rather they left things open to speculation. I still can’t see why they chose it, but they’ve now decided to apply it to their whole entire range of CPUs. There is now officially going to be a Core i5 and a Core i3 range, and a star rating system for performance comparisons.

The new i5 range (from the rumor mill) covers not only the i5 chips being released in October, but it could also encompass some Core 2 Duo chips both in the desktop and mobile sectors. This should also cover the P4x, P5x, and X38/X48 chipsets.  The Core i3 range should encompass lower-end Core 2 Duos and Core 2 Quads, as well as some mobile chips (namely the T3400 and T4200) and the current Pentium Processors. The G3x and P9xx chipsets follow suit, seeing as they are low-end versions.

See what I mean? The star rating makes sense

But here’s where it gets confusing. This now means that current Core 2 Duo owners now fall under the Core i3 or Core i5 brand, but the chips are still going to be sold under the older naming system. Users shopping for a Core 2 Duo system in a few months, however, will instead find a “Core i3” rig in its place. It’s the same computer, but it has a different name, which will confuse a lot of people in the process of the name changes. Worse still, the revisions of current processors will probably adopt the new naming system, so the same architecture found in the CPUs will have two names. I can see it now:

“I have a Core i3 E7200 running at 4Ghz! Beat that!”

“I have a Core 2 Duo E7200 at 4Ghz as well.”

“So what?  Mine’s still better.”

“…”

“…”

“… Idiot”

At least the rating system is informative.