News Another Intel Broadwell delay rumoured

With the improvements getting smaller and smaller these days, it's best not to think tooo much about it really.

When the time comes for a new PC, simply look what's best bang for your buck from the currently available hardware and don't look what get's released in the future.

Since the i-Series started, we've basically seen an improvement of ~5-10% for the CPUs for every generation, and the iGPU is getting way more attention actually than the CPU these days. The differences between each generation simply isn't big enough to wait for anything new, especially with prices staying the same.

 

On the whole I agree jrs77, it's just that it can get a little confusing as if you read articles, reviews and especially comment sections things can be a little misleading. If you judged things by what others had to say the general feeling of Haswell is one of negativity but only from those with IB, SB or a select few processors from a gen or two before that. Upgrading from a Q6600 or less would see a massive improvement.
Although it is also good to keep one eye on the future as I mentioned earlier regarding Broadwell and Skylake. The move to DDR4 memory could be significant while requiring a platform change rather than just a CPU upgrade, investing in a new build within 12 months of that release would be seen as folly by many.

 

Personally don't think CPU performance and price has gone anywhere because they have not had to.

There top end high margin stuff is unchallenged in the desktop area. There mobile stuff performs well but consumes a lot of power doing so. Servers they are unchallenged.

Only really mobile where they are been pushed.

Dou't we would of even seen a better Gpu if it was not for apple complaining about it and threatening to go elsewhere.

If they were other brands of CPUs offering better performance for less then we would of saw a more faster and cheaper top end. At the high end your choices are super limited. Hence the prices are crazy. I expect the top end broad well desktop the 4770k equive today to be £300+ When it does launch and it will still sell.

The 980 everyone said was crazy money 4 years back today it's still faster than anything the competition has and is likely to stay that way for the considerable future.( £150 ownership per year if you brought it then)

Still hope someone buys out AMD and actually goes back to pushing the high end of the market.

 

Competition is an argument that held true for a very, very long time but I don't think it does any more, at least not the way it used to. Just look at the GPU market to see that. My first proper GPU was a GeForce4 4600Ti for around £240 I think, retail. For many years £250-£300 was the mark for high-end GPUs but look at their prices now. AMD and nVidia are on par with each other each gen give or take slight ups and downs and they are both charging big for their high-end cards these days. Whether you need it or not but £1000 for a single consumer sold GPU is a massive change from how things used to be.

 

If you Upgrade to Haswell coming from a Q6600 then you have 4 generations, i.e. ~40% improvement, and sure that feels huge. But the difference between Haswell and SandyBridge is only some ~15-20, and that's not that big anymore, especially since you don't have the benefit of coming from a 45nm part reducing TDP/SDP.

DDR4 won't be any better than DDR3 for the first one or two years. Was the same when we switched from DDR2 to DDR3. So it's not a really good reason to wait for the next generation if you're in the market for a new PC today.

Stop looking what might be, or what could've been a better choice. The moment you decide that it's time for a new PC is just the right moment to buy.
If you start looking ahead and try to anticipate how much better it would be to wait for the next gen you'll end up with not buying a new PC at all until you really have top because your PC died.

Nowadays hardware simply doesn't improve that much anymore between single generations to hold back.

 

I don't think you have too much to worry about on the AMD side of the fence. Their main aim now is to make a CPU/GPU combo that's actually good for gaming, so I can see the APU ending up a great tool for PC gaming. They also have Mantle on the go, as well as the slightly lesser TressFX.

I think their main mission now is to continue making their APUs more and more powerful so that they can actually game with the best GPUs. So I can't see them letting off the gas because of a tablet, for example. They have too much resting with things like the consoles and Mantle ETC to go all lame on us. I must say that although it would be rather boring it would be pretty epic if an APU came out that could actually pee with say a R280x.

Intel though? I'm far more suspicious of them tbh. They have pretty much made it clear now that they have no interest in the performance PC sector and are doing absolutely everything to get away from it. I can't say I blame them, many have predicted the death of the desktop PC. The thing is there still needs to be powerful PCs for designers/workstations and so on so they're either happy to sit back on their 2011 range for a while or are just not bothered. I have to say, if you were wanting raw power the CPUs they've already made are more than good enough really. It's kind of like things have caught up with themselves and it would be a tail chasing exercise.

Not much fun though. I guess we're going to have to just get used to the fact that the CPU market is about to freeze for the foreseeable future.

It sucks, but tbh the way things are for Intel right now seems way more dangerous than it does for AMD. AMD can stick to what they have now and can sit on it for a bit but Intel do not have the console/being enthusiastic about new API thing going on. They need to get into something quick or it could all go a bit pear shaped pretty fast.

 

@ jrs77

'DDR4 won't be any better than DDR3 for the first one or two years. Was the same when we switched from DDR2 to DDR3. So it's not a really good reason to wait for the next generation if you're in the market for a new PC today.'

I think you are judging the move to DDR4 the same as DDR3 but with APUs now using system memory and showing significant performance increases from faster and faster RAM there is much more of a reason for the speed increases than there used to be. DDR4 is going to be massive for IGPs.

'Stop looking what might be, or what could've been a better choice. The moment you decide that it's time for a new PC is just the right moment to buy.'

Sorry but I disagree, I think your attitude is completely callus. I take my time, think about what I want and what is available right now and in the near future and then I make my choices. For this reason I am waiting for the release of Windows Phone 8.1 rather than getting a 'soon to be out of date' WP8 device.

 

The choice Intel seems to have made over SATA Express, both are ways to open 2 PCIe lanes to be used by SSD's and both should mean around 16GB/s transfer speeds. I'm guessing Intel have decided to push M2 over SATA Express due to it being better suited for smaller devices.

SATA Express has backward compatibility.
M.2 is better for smaller devices.

Note: Not sure if i have my speeds correct but AFAIK both formats come out the same in terms of faster transfer speeds.

I respectfully disagree that you would only see around %40 improvement, my guess is it would be closer to %100, either way it would be a big improvement going from a Q6600 to Haswell so i wont split hairs

I am also going to disagree that you won't see a big improvement going from DDR3 to DDR4, sorry.
Even right of the bat DDR4 is 2133–4266 MT/s (million transfers per second), versus DDR3's 800 to 2133 MT/s. Crucial demonstrated some DDR4 at this years CES that was running at 2133MHz the lowest speed that DDR4 will be introduced at, eventually its expected to reach 4.2GHz

Bit-Tech wrote a nice article back in 2010 that goes into more details.

Where i do agree with you wholeheartedly is when you say to stop looking at what might be, as there is always something just around the corner that leaves you undecided if you should upgrade or not. It's not worth thinking about what maybe, otherwise you just end up wishing you life away, or maybe that should be hardware

 

I'm also waiting for next gen with DDR4 and m.2 to upgrade my main PC... for other things like storage server I am looking at Avoton based instead. They are simply a much better value at around $300-350 for an 8 core CPU and motherboard at once. Likewise for HTPC and small workstations the new Bay Trail quads are going to be excellent in the $60-1?0 tier.
Intel would be a lot smarter not to have such a rapid release schedule when there is no need for it. They are competing with themselves and it's simply foolish.

 

I just want to clarify my stance on 'waiting'. I believe there is a vast difference in waiting 3-6 months before building for an imminent release and waiting 12-18 months for a generation that hasn't even been finalised yet. So as examples, I would consider waiting for the Haswell refresh but I would not wait for Broadwell, just as a month or two ago I would have waited for the new AMD GPUs rather than jump in with the 7xxx series cards or wait for nVidia's full fat Maxwell GPUs.

 

Recently Asus demonstrated a SATA Express Mobo at CES and kitguru managed to get a hand on preview of it so that maybe worth waiting for. And rumor says Haswell may see a refresh in April.

 

Rapid release schedule does not really hurt your company if people are willing to buy your products, the $33bil revenue Intel made from the Pc Client group suggests people are willing to buy.

They could release a product once every few years but there revenues would tank. Some people just like in other technology groups have to have the latest products be it CPU, GPU, Smartphone or Tablets.

Use Samsung as a example they release a new product for 2 quaters that product brings increased revenue for example the SG4. The 2 quaters after people are waiting for the next release to launch building hype and expectation of what it will do.

Enthusaists would prefer a longer release schedule simply for money reasons 90% of the time but we are not the ones giving Intel $33 bil a year. That money comes from the likes of Dell, Lenova, HP, Apple been the big 4 who will contribute to that revenue figure.

The total of Retail sales of there cpus will be alot lower than most assume or think. Even Assuming theres 6million enthusaists ( this is a overestimate, with recent estimates putting the figure between 3-4million) buying new cpus on every release thats only around $2.4bil revenue ( If everyone brought a 4770k which did not happen).

Which is around 7% of Intels total revenue stream for the PC Client group. Apple alone last quater give Intel $1.7billion for chips it brought ( This is shown in Apples latest Figures)

Apple, HP, Dell, Lenova all have yearly release schedules to co inside with this new launch of technology from the developers of it.

 

Yes, DDR4 is told too go up as high as some 4+ GHz, but the first year or so we'll probably see only DDR4 2133-2400, and we allready have these speeds with DDR3 these days. It was the same with DDR2 and DDR3a few years back.

And yes, the IGP heavily profits from the high-speed DIMMs, but the CPU doesn't as much. Those who don't use the IGP will be totally fine with DDR3-1600 even when DDR4-2400 is allready available, and the difference won't be very noticable for most tasks.

And for the difference between CPU-generations. You basically don't see much difference between a SandyBridge and a Haswell-refresh chip when playing games. It's all about the GPU here.

I basically look what's the currently best available CPU for €200 and be done with it. No second thoughts on what might come in a few month. Pair it up with a €200 GPU and a €100 motherboard and that's it. A new gaming-rig for €500 every two years aslong as the DIMMs and the HDD/SSD are still working fine.

 

maybe Samsung can buy intel as they earn 3x times as much profit and have over 100% more revenue....


on a serious note , DDR4 will be a major benefit to AMD APU`s , as its sadly obvious they are quite bandwidth starved.


although

http://www.3dmark.com/3dm11/7922575

`performance` pre set for 3dmark 11 - using PC2133 CL10 ram on an A10-5800K

 

We are told Crucial will be offering 3Ghz DDR4 sometime in the third quarter, around the same time it is rumored Haswell-E will hit.

Looking into my crystal ball i would say we will see a Haswell refresh along with M.2 / SATA Express around April, and then not much until the end of this year when i suspect we will see Haswell-E and DDR4. I can't see Broadwell arriving this side of Christmas.

 

We're soldly into the DDR3 2133-2400 territory now and we can hit 4GHz DDR3 in special conditions, so 3GHz DDR4 isn't that unfeasible this year. Depends on what the IMCs and memory ICs are capable of tbh; people won't hold back if they have to.

 

I’m in the market for pretty much a complete overhaul; basically every component except my hard drives badly needs replacing.  I’m coming from an E8400 paired with a Radeon HD6770, so from my point of view pretty much *anything* I buy is going to be a quantum leap in performance, be it SB, IB, Haswell or even an AMD APU.  With such massive performance gains on the cards I really can’t be bothered to wait for the Haswell refresh or whatever else is around the corner; the cost will probably be more than I’m going to pay now, and the gains over a Haswell i7 4770k are likely to be marginal.  If the Haswell refresh really is all about the iGPU then it’s not worth it for me: I’m going to be pairing this new build with at least a GTX770 so the iGPU will never see the light of day. 

 

Still running Lynfield i750 with no problems does everthing I want it to but Broadwell could be time to upgrade to a new system

 

Z97 and Haswell will still be a good buy this year as multi-year investment (unless you specifically want X99). Everyone kept waiting for the 'next gen' back when Sandy/Ivy/Haswell launched, so Broadwell should mirror the Sandy-Ivy difference as it's the same mArch. In April Intel may detail its 14nm process at IDF, or maybe you'll have to wait until Sept - from that you'll learn its advantages and estimations for Broadwell.