Intel shifting to two year cycles for desktop with no Broadwell being released?

[bicpug]

Yearly releases don't make sense for Intel, i like to have a fast pc but i only update the cpu every 3-4 years; only the most rabid gamer with more money than sense feels the need to update to every new cpu/gpu that's released.

 

[WingZero30]

If I remember correctly Q6600 was the first mainstream quad released in beginning of January 2007 by Intel. So for almost 7 years (going into 8th year with rumored Haswell refresh in 2014), we have been on Intel quad cores in the mainstream segment. That's pretty impressive!!

However what I would like to know is when we can expect 6 cores mainstream cpus from Intel in the future? I would be somewhat disappointed if there was no 6 cores mainstream Skylake processor.

 

[=XDC=FluphyBunny]

....
However what I would like to know is when we can expect 6 cores mainstream cpus from Intel in the future? I would be somewhat disappointed if there was no 6 cores mainstream Skylake processor.

Intels problem in introducing a 6 core mainstream cpu is their own enthusiast 2011 chips (or whatever their next platform is). It would be very difficult to fit them in without taking sales off themselves. Imagine a hex core Haswell now against a hex core ivy-E. It would also highlight how the E chips are a generation behind the mainstream.

This may be a daft question but why are the Enthusiast chips released a generation behind the the mainstream? (ivy bridge - haswell ect)

 

[drunkenmaster]

So still a haswell 'refresh'

And



Yes but that's due to a slow down in per cycle performance. Give a performance increase that is worth changing for and people will upgrade.

edit: What we really need is AMD to bring out a Steamroller chip that is equal in performance and undercuts on price. Hopes are not high.

Yup, Intel seem to have missed that it goes hardware then software then the need for upgrade is generated and people move on.

What they've done is destroy performance improvements, focus on all the wrong areas(for desktop), no new fancy software that can use 50% more performance or double the core number and as such stuff that ran on a CPU bought in their Dell 5 years ago is still running fine.

What I find unbelievable is Intel both don't know this, are hurting their own business with their fabs being no where near fully utilised(which is inefficient and costs a lot more). The ridiculous thing is these two things are related, they used to push faster chips every year with bigger die sizes, the bigger die sizes meant less chips coming off each wafer, and creating significantly faster chips which drives up demand. By going small chip they've both drastically reduced foundry utilisation and stunted performance growth which has stunted demand.

Yes, they want to jump into mobile more and push performance/power usage/battery life in that segment. But there is a reason most companies have different products for drastically different segments. By trying to make a chip that can work in (very fake) 5W up to 100W.... it's daft. They are making the best high power chips but they haven't moved forwards with them in ages, but they aren't even close to the best chips in mobile and still need Atom for very low power usage.

AMD has stagnated for years but Jaguar looks capable of anything from 3-25W and giving good performance and power/battery life/usability in any of its form factors. Intel both because of their money and ability really should have split phone(atom), added a low power mobile chip for tablets through pretty decent laptops then have a 30-100W high end chip for more powerful laptops and desktops. High end Intel performance has stagnated mostly because they are trying to jam this chip into tablet power bracket for absolutely no reason. No other company in the world thinks a 100W capable chip should also work well in a 5W tablet.... for a reason.

Realistically if Atom had always been focused at sub 3-5W then that would likely be a better chip as well for mobile than it is now.

Anyway, if Intel just made a stonking octo core in mainstream it would drive software, drive demand, sell well at £300 with failed chips being hexcore a little cheaper and it would also drive up usage at the fabs.

 

[John24]

Question: Presuming AMD don't produce anything suitable, when is the next big performance jump for video encoding from Intel and will that jump when it comes be it for the following 2 years?

 

[accod]

Anyway, if Intel just made a stonking octo core in mainstream it would drive software, drive demand, sell well at £300 with failed chips being hexcore a little cheaper and it would also drive up usage at the fabs.

Completely agree. I think Intel are still looking at what happened with the P4 and are in the mindset that they must reduce power usage like they did with the core series, forgetting that a decent performance increase is required too.

 

[Jumper118]

amd need to invent a 128bit, eight core, 100w tdp monster cpu from nowhere and sell it for £250-then intel won't be doing their 2 years cycles. just a shame that isn't going to happen. a 128bit cpu is needed now that we can get a 64bit phone.

 

[Rooster212]

amd need to invent a 128bit, eight core, 100w tdp monster cpu from nowhere and sell it for £250-then intel won't be doing their 2 years cycles. just a shame that isn't going to happen. a 128bit cpu is needed now that we can get a 64bit phone.

128 bit OS is completely unnecessary. 64 bit can already access billions of gigabytes of RAM, and everything is being designed around 64 bit.

I reckon at least 20 years before we see 128 bit OSs at all anywhere, maybe datacentres.

Some interesting points though if you want to read up on them.

Stack Overflow

 

[Seanoog]

I don't like this new development from intel.

I have a i5 2500K @ 4.4GHz, 2x ATI 6970 2GB, 16GB Ram, 2x256GB SSD, Dell U3014.
A powerful an expensive system, an yet I still have that itch to upgrade (Thankfully not yearly.)
Since buying the i5, Intel an AMD have released new chips, but I haven't been compelled to upgrade due to the tiny performance increases with each release + the staggering costs with the new chips (looks at Intel specifically.)

With this 2 year release cycle, they are affectively killing the desktop, oem markets (Servers will struggle on.)
R&D investments won't increase, few products will reach a dwindling market an costs of production + at market will increase...

My 2 pence..

 

[Sampsy]

Really it doesn't matter much whether they are releasing new CPUs every year or every two years if they manage to improve them enough. Right now there is next to no point in me buying a new desktop CPU - I have a 2500k @5Ghz - is the tiny gaming performance gain worth £200+? Not even close. If they actually release a sensible mid-high end dedicated CPU without wasting transistors (and my money) on a useless integrated, low-end GPU that I am never, ever, going to use THEN I might upgrade.