This is it, boys and girls. We're nearing the end of days for the PC processor as we know it. There are storms of change on the horizon and it's anyone's guess what the PC will be like in hardware terms when it blows over.
Right now, things are much as they've always been. You pays your money, you takes your choice. In other words, you get to choose any CPU you like and match it with a motherboard and a graphics card. You've got both AMD and Intel options. And in many cases, you've still got full control over the chip you buy. You can overclock it, underclock it, swap it out and generally mess about with it.
Trust us on this - much of that could begin to disappear within the next 18 months, so enjoy it while it lasts.
If you're wondering why, there are a number of trends at work. Firstly, AMD's position is pretty precarious. We'll come to that in more detail later, but AMD is truly teetering on the edge of oblivion. Then there's the market's obsession with all things ultra-mobile and the technological trend towards greater feature integration that entails.
Very likely, it won't be long before you can't buy a drop-in CPU. They'll come soldered onto motherboards. So while we've a few complaints about the current state of play in CPUs, there's a chance we'll soon be looking back on this as a golden age in terms of choice and flexibility. So get out there and revel in it, we say.
There are some great CPUs from both AMD and Intel that can still be enjoyed in true enthusiast fashion. They're fully drop-in-able. They're tweakable. They're fun. And the way the CPU market is going, they'll probably keep getting the job done for at least a couple of years.
During the making of this CPU roundup, it felt like we were living on borrowed time. The PC is in a transitional period and five years from now much of what you take for granted when you spec up a rig will either be gone or very different.
There are two major drivers here: the trend towards ultra-mobile and AMD's failure to really stick it to Intel at the performance end of the market, even if it produces good chips for tighter budgets.
But let's start with that ultra-mobile mania. It explains why all of Intel's mainstream PC processors now contain on-die graphics. With any generation of computer chip, you have a given quantity of transistors available. That transistor 'budget' increases over time as manufacturing technology shrinks individual transistors.
In the past, it was pretty much all spent on improving CPU performance. More complex execution units, more cache, more cores, added features to help the cores like an on-die memory controller.
Already, however, that process has slowed. Intel's latest Ivy Bridge processors are a great example. At around 1.4 billion transistors for the quad-core version, such as a Core i5-3570K, Ivy Bridge is fully 240 million transistors bigger than the Sandy Bridge quad-core chip it replaced, but it doesn't have any additional cores or extra cache. Okay, the execution units are slightly tweaked, but we're talking typically low single-digit improvements in per-clock performance. That's not a lot to show for a 20 per cent increase in complexity.
The logical explanation, of course, is that Intel chucked almost all those 200-odd million transistors at Ivy Bridge's graphics core. The same thing will apply next year when the new Intel Haswell chips arrive. They will still be four-core beasts at best, and most of the increase in transistor count will be blown on improving the integrated graphics.
The problem is that, to date, Intel's on-die processor graphics has not been gaming worthy. In a mobile PC context, the power efficiency of integrated is great, but on the desktop and if you're into games, it's dead silicon. Worse than that, it means Intel is compromising processor performance - performance you'll actually use - in favour of improving integrated graphics performance that you won't use from crap to merely mediocre.
Eventually Intel's processor graphics will come good for gaming, but we're still several years away from that happening. Anyway, all this is because mobile computing is driving CPU design. Actually, that's not entirely the case - it's also because AMD hasn't stepped up to the plate.
AMD can't compete with the sheer raw performance of Intel's fastest current four-core chips in the LGA1155 socket. And that means it's nowhere near Intel's high-end chips in the LGA2011 socket. LGA2011 chips, of course, don't have processor graphics and are entirely focussed on CPU performance. But without AMD keeping Intel honest, LGA2011 chips are intentionally hobbled and very expensive.
Put it this way: if AMD had a competitive CPU, Intel's six-core LGA2011 CPUs would probably be half the price they are today, and there would also be eight cores on top. Put it all together and the unavoidable, undeniable conclusion is that Intel's desktop CPUs are already nothing like what they would be if Intel was simply focusing on performance.
But what of AMD? Well, that's an entirely different problem. And it's all to do with execution. Put simply, everything AMD has launched in the past five years has been too late and too slow. That's a great pity because AMD is more likely to sell straightforward CPUs in the configurations that desktop PC enthusiasts want. Plus, if those CPUs were more competitive, Intel would surely be forced to do things differently, too.
On the bright side
At this stage, we've painted a pretty bleak picture of the state of PC processors. But actually, things are still pretty good. You can still buy CPUs separately and mix and match them with motherboards and GPUs, allowing you to get the performance balance just so.
And AMD's chips are still competitive at certain price points, which has a knock-on effect across the market. More to the point, while it's likely CPU performance would be even higher if AMD had played a better game in recent years, of course today's processors are still extremely effective bits of kit. Intel may not have actually added cores to its mainstream chips, but it has done a very good job of improving per-core performance.
Sandy Bridge was a huge step forward in that regard and the latest Ivy Bridge processors raised the game a little further. All of which means that these are still the good times, right now. Five years from now, it's hard to say, but it's extremely likely you'll have a lot less choice, and year-on-year CPU performance increases may have slowed to a trickle - AMD may be a goner, for instance, and it's likely you won't be able to buy a stand alone CPU and drop it into your motherboard of choice. A few years after that, you may have to swallow motherboard, CPU and graphics in one big pill.
Back in the here and now though, let's enjoy what's on offer. If you're gaming mad, like us, the good news is that you don't need to go right to the top of Intel's current catalogue to get great performance. Intel's mainstream quads are still outrageously good. For those on tighter budgets, there are some very compelling options, some of which come from AMD.
If you've got a ton of cash, of course, there are even more options. In fact, we've thrown an Intel Xeon chip into the mix to show both how things might have been at the high end and also how you can get round Intel's increasing tendency to sandbag.
It's also worth noting that from a PC performance and gaming enthusiast perspective, now is a really great time to buy. Next year's Haswell chips from Intel are highly unlikely to bring dramatic increases in CPU performance. On the AMD side, we had hoped to see the company really raise its game next year, but now that's looking unlikely before 2014. If ever.
So it's fair to say that a decent CPU bought today will still be competitive for several years to come. As Arnie says, then, do it. Do it now.