What happens if Intel does end overclocking?
Posted on 29th Jul 2010 at 12:42 by Richard Swinburne with 79 comments
The feedback in our news post mirrored my own initial reaction: how dare Intel! How dare they stop us from generating £700 CPU performance from a £150 one?
Commercially speaking, there's a good argument it never made any sense to allow overclocking in the first place - I can't imagine Intel sales were happy to see etailers selling pre-overclocked bundles and whole PCs, or enthusiasts flocking to mid-range chips such as the i5-750 rather than high-end models. Flock they did - one of our most popular articles is how to overclock a Core i7-920 - despite the fact it's nearly two years old now.
While Intel's sales teams probably aren't happy about overclockers pouncing on cheap chips, they're probably not losing too much sleep over it. The enthusiast market is tiny compared to its mainstream sales and pre-OC hardware is limited to smaller companies only. If we compare the number of Scan (just one example) pre-OC kits and PCs to the amount say, Acer or Dell ship only ever at their rated speeds, it's a tiny percentage - why should Intel care? This is especially the case given the fact Intel is in its most profitable state ever and has yet to really see revenue from the explosive smartphone/tablet market - that'll further develop in 2011/2012.
In its long history, Intel has only recently acknowledged enthusiasts anyway, and as with all the other markets it competes in, it's really only interested if it can make serious money - hence its idea of making up for lack of sales volume with high prices. Spend a thousand dollars and get a CPU with a fully unlocked multiplier. Woo bargain, cheers guys. I even wrote an article about this back in 2008.
This alone makes the existing K-series CPUs something of a big deal for Intel, and I wonder then if for Intel, they're the first step to a bigger plan for defining a logical separation between the different, non-interoperable sockets it's offering to consumers.
LGA1366 proves that many enthusiasts have accepted paying more for a performance platform: £200+ motherboards and a £200 CPU, plus expensive 6GB triple channel kits still sell. I know at least one manufacturer whose biggest market is X58 boards in the UK. If LGA1366 offers more overclocking options then that's another reason for enthusiasts to buy it over LGA1155. That said, by providing the K-series CPUs, if you do want to fully overclock with Sandy Bridge, you can.
In both scenarios, you will pay more - but let's be honest here, we all love upgrading, and if the upgrade is worth it we will budget around it. AMD's protracted weakness when it comes to performance means Intel has an advantage and if Fusion isn't any good then Intel will continue to be the only option.
If Special K is your only option, will you upgrade?
That said, as I drink my ice tea, I wonder if the lack of overclocking really is a big deal. When was the last time you thought your current PC was CPU limited? Can you really tell the difference between a 3.3GHz CPU and 4GHz? Will you really miss those extra multipliers for a few extra MHz? I'm not talking about a benchmark machine where a few MHz matter for a better 'number' - I mean the one you use every day. You say you will now - you think you need it - but if it came down to working within a budget, a large proportion of the users would live without it.
I'll admit I was seething at Intel for taking away from a hobby that all enthusiasts of all ages can enjoy, not just those who have lots of cash. Nostalgia is a powerful thing until I realised my own CPU isn't even overclocked. 95 per cent of the time it sits at 1,176Mhz and I rely on TurboBoost for the rest. I get more pissed off that it would crash than if it takes a minute longer to encode something. 30 minutes longer? OK that's a big difference, but it's an extremely rare event and unusual I can't multitask around it.
Who really loses?The biggest losers aren't actually consumers, but Intel's partners: motherboard guys mostly, but also heatsink and the watercooling companies. Multiplier overclocking puts very little stress on the motherboard beyond power provision and if every CPU can achieve the top multiplier then what cards do motherboard companies left to play? The major marketing arm for almost everyone right now is overclocking - overclocking events, overclocking quality and stability, HWBot etc. It's seen as the F1 of the industry: developments here fritter down to use mere mortals through BIOS updates and the next board revision.
If that angle is taken away then we're left with less than scintillating features such as energy use, component quality and price. Do people really care about energy use? Maybe, but not in the way they do for performance. Even if they did, motherboards aren't something that factors that heavily in terms of power usage, certainly not compared to PSUs, graphics cards and CPUs.
Component quality - well, if you believe the marketing, we're already at 50,000+ hours of super-ultra-military grade OMFG capacitors that will save us from cancer, so what else do we need, and do we care as long as it works? If it breaks I want a good after sales service and warranty.
And then there's price. Well simply put, a price war is a death sentence for motherboard manufactures. For the top three companies motherboard divisions are their most profitable areas, so a price war is an absolute last resort. Remember they want to upsell you.
If I was a senior person in a motherboard division I'd be sweating through a shirt an hour right now.
Do you put more eggs into the Fusion basket? How long until it reaches the market though?
Do you push overclocking with K series CPUs and hope there isn't a huge consumer backlash? What about good publicity and reviews? Will anyone even need them if the boards all overclock the same and it's simply a case of "does it work"? Yikes! I hope not - I need to get paid!
The bottom line is that if Sandy Bridge does dramatically curtail overclocking then the enthusiast end of the PC industry will change drastically. In some respects, it already has - we already live with multiple sockets, this is just the next step. Perhaps our best hope is for a competitive future between AMD and Intel because one company dominating and dictating the market is bad news for everyone.