Will Haswell be the clincher CPU of this decade?

[SickBeast]

Originally Posted by Idontcare

[Xpage]

A drop from 17w to 10w, will also call for less cooling, thus fans could be off or lowered rpms on idle, thus less energy for cooling too. LCD screens will still take up a majority fo the battery power.


Broadwell I think will have major limitations due to the amount of heat generated per mm2, thus will not be a good overclocker. Unless some form of exotic cooling is used eg. IBMs water cooling http://www.sciencebuzz.org/blog/ibm-...crochips-water

[IntelUser2000]

Originally Posted by Xpage

A drop from 17w to 10w, will also call for less cooling, thus fans could be off or lowered rpms on idle, thus less energy for cooling too. LCD screens will still take up a majority fo the battery power.

It already can be turned off. TDP is for worst case scenario, like when you are doing something on load. And I don't mean "load" by movie watching and internet browsing, or even flash games. But running games, doing photo/video editing, and benchmarks like LinX. Since its likely you won't be doing that on battery, it ends up being relevant solely for system design purposes.

[jpiniero]

Originally Posted by Fjodor2001

What would be the purpose of that? Are they aiming at enabling small compact desktop PCs such as the Mac Mini? Possibly fanless?

To save costs as the share of desktops in terms of Intel's sales continue to decline.

Otherwise, what reasons would the consumers owning a SB/IB/Haswell Desktop PC have to upgrade if the performance will be the same or less compared to their old computer due to the low power => low performance constraints?

That's Intel's problem. The need for faster processors for the vast majority is shrinking. For those who do need the performance, that's what the Extreme line is for.

[pablo87]

I'm not very familiar with Haswell and we don't have reviews yet but rather than looking at technological achievements or how it kills AMD, perhaps we should be looking at what user problem it solves:

improved IPC - well for most segments that is always a good thing to have more MIPS/watt.

gaming level graphics - it doesn't sound like it to me. The only real beneficiary I see is vastly improved graphics on lower power devices, otherwise its a betterc speed and feed that is of limited value to most of those buyers.

improved battery life - of value for sure but it appears that if it's not W8 or panel with dram, it's impact on the overall system is limited.

cheaper - doesn't appear so...

Tablet - if its 10W, it's basically not in the game yet though it might get there eventually and maybe that's the ultimate goal.

Thin / ultra - if it makes them thinner with improved battery life and better graphics then yes it's a winner which leads me to ask this: was Haswell designed primarily for Ultrabook? Seems that way...or perhaps its design was heavily influenced by a major OEM who were considering an alternative architecture...

[Ferzerp]

There is an old study, that, at the time, split the power usage in to 3rds.

1/3 was CPU, 1/3 was screen, 1/3 was everything else. However, I think this info is relatively old.

[Intel17]

Originally Posted by pablo87

I'm not very familiar with Haswell and we don't have reviews yet but rather than looking at technological achievements or how it kills AMD, perhaps we should be looking at what user problem it solves:

improved IPC - well for most segments that is always a good thing to have more MIPS/watt.

gaming level graphics - it doesn't sound like it to me. The only real beneficiary I see is vastly improved graphics on lower power devices, otherwise its a betterc speed and feed that is of limited value to most of those buyers.

improved battery life - of value for sure but it appears that if it's not W8 or panel with dram, it's impact on the overall system is limited.

cheaper - doesn't appear so...

Tablet - if its 10W, it's basically not in the game yet though it might get there eventually and maybe that's the ultimate goal.

Thin / ultra - if it makes them thinner with improved battery life and better graphics then yes it's a winner which leads me to ask this: was Haswell designed primarily for Ultrabook? Seems that way...or perhaps its design was heavily influenced by a major OEM who were considering an alternative architecture...

Haswell is billed as being the first product designed specifically for ultrabooks.

[SickBeast]

If Haswell's quad core laptop CPUs are cheap with powerful integrated graphics I will definitely be getting one.

[hokies83]

Originally Posted by SickBeast

Haswell will be AMD's death blow.

Sb&IB is already outrageously Dominating Amd...

Now think if Broadwell is out before Steamroller... If there is even going to be a steamroller that is...

[Zodiark1593]

Originally Posted by hokies83

Sb&IB is already outrageously Dominating Amd...

Now think if Broadwell is out before Steamroller... If there is even going to be a steamroller that is...

Heck, I think Nehalem is still topping AMD's best chips.

[Ajay]

Originally Posted by SickBeast

Haswell will be AMD's death blow.

Bulldozer was AMD's death blow. Now we are just tracking the blood trail to we get to the kill.

[dguy6789]

Originally Posted by Zodiark1593

Heck, I think Nehalem is still topping AMD's best chips.

As far as I know Penryn is still

[Exophase]

Originally Posted by AtenRa

100% of Laptops are APUs
~45% of Desktop is APUs

http://www.3dcenter.org/news/genauer...narchitekturen

There's a difference between has an APU and is an APU. Discrete GPUs are still very common in laptops. I explicitly looked for an IB laptop that did not have discrete and couldn't find one that also had the other features I wanted.

Why would laptops have so many more APUs than desktops anyway? Are you saying people upgrade their desktops much more frequently? Because Intel CPUs have all had IGPs for a couple generations, and while you can't get AMD's current CPU-only chips in laptops they aren't exactly covering a lot of the desktop market either.

[AtenRa]

Originally Posted by Idontcare

OK, so we'll say 75% of AMD's market share is based on APU products.

According to Mercury Research, as reported by xbitlabs, AMD's unit market share has now dropped below 17%.

75% x 17% = 12.75%

So we are looking at AMD's APUs addressing at most 13% of the total market in terms of units.

What I don't know is how much of the budget segment is represented by AMD's 17% market share. Does AMD own 90% of the budget segment? 80%? 70%?

Numerically speaking, provided AMD owns more than 0.5/0.75 = 66.7% of the budget market, and 75% of their units are APUs then they can claim to be responsible for bringing 3D gaming to the budget buying masses.

The market share numbers are for the entire x86 market including Servers. Because we only care about Desktop and Mobile(Laptops) here, AMDs market share is larger because their server market share is small compared to Intel's. I dont have the Q3 2012 numbers but a year ago it was close to 33% in desktop.

But i was only talking about the product itself when i have said that we already have a product that brought 3D gaming for the budget buyer.

Any way, I dont want to continue this over here

Edit: 17% market share is for Desktop + Mobile only

[Nemesis 1]

I believe Haswell will be the big core that ties IE7 socket and Itanic into the same form factor . But The Big core of these times is PHi. 61 small cores on 1 die . THIS IS THE FUTURE here now. Intel will be releasing the 22nm merrifield phone SoC soon . Imagine 100s of these on a die core. PHi is the REAL CLINCHER but not in its present form.

[Ajay]

Originally Posted by Nemesis 1

I believe Haswell will be the big core that ties IE7 socket and Itanic into the same form factor . But The Big core of these times is PHi. 61 small cores on 1 die . THIS IS THE FUTURE here now. Intel will be releasing the 22nm merrifield phone SoC soon . Imagine 100s of these on a die core. PHi is the REAL CLINCHER but not in its present form.

I don't think so. Xeon Phi is aimed at the HPC market for a reason. It's performance is excellent, and it's programming model looks to have a lower barrier to entry than CUDA. But its excellent performance is only for extremely parallel code. It's designed to offer this performance only in applications that support very high levels of concurrency. It's ST performance is nothing to write home about.

[Ajay]

Originally Posted by Haserath

I think he means clincher because Haswell will be the last significant increase we see for awhile, on the CPU side at least.

Nehalem was something like 15% better clock for clock over the latest Core 2, but it brought hyperthreading for some extra oomph, it wasn't really all that big of a jump otherwise.

Sandy Bridge was 10-15% better over Nehalem, brought some new instructions, and brought CPU power levels down a lot.

Haswell will be 10%? better than Sandy Bridge, will have new instructions(what else could they bring for big benefits now?), and will bring power levels down again on the CPU side even if TDP for the whole chip isn't lower.

There seems to be a trend of lower CPU performance gains, unless Haswell is actually faster than rumors have said it is.

If that was the posters intention, then my answer would is, in all probability, yes. I find it hard to imagine that Intel will achieve >10% increase in IPC on any future chip this decade on previously compiled binaries. Of course, if Intel can pull out a few shockers from it's bag of materials science tricks, then they may keep pulling 10% per tock.

iGPU is different, with increasing possibilities to put larger eDRAM configurations on an interposer or even on chip and with a very wide memory interface to the iGPU, we could see respectable performance by the 14nm node (Skylake gets a new GPU core). This does depend on Intel improving it's drivers. It is possible we may actually see something impressive by the end of the decade, since GPU designers may not have access to process technology that allows them to compete as effectively with Intel as they do now.

[Ferzerp]

Wow, someone tried to steer the conversation to igpus... Such a surprise.

[Ajay]

Originally Posted by Ferzerp

Wow, someone tried to steer the conversation to igpus... Such a surprise.

I wasn't trying to steer anything. Just pointing out that Intel has more room to improve the performance of it's iGPUs than its CPUs. No agenda here.

[Ferzerp]

Originally Posted by Ajay

I wasn't trying to steer anything. Just pointing out that Intel has more room to improve the performance of it's iGPUs than its CPUs. No agenda here.

Let me specify. Someone tried to steer the conversation to how "good" amd igpus are. When all you have is a tack hammer....

[frozentundra123456]

Originally Posted by AtenRa

The market share numbers are for the entire x86 market including Servers. Because we only care about Desktop and Mobile(Laptops) here, AMDs market share is larger because their server market share is small compared to Intel's. I dont have the Q3 2012 numbers but a year ago it was close to 33% in desktop.

But i was only talking about the product itself when i have said that we already have a product that brought 3D gaming for the budget buyer.

Any way, I dont want to continue this over here

Edit: 17% market share is for Desktop + Mobile only

Depends on how you define "3D gaming". On the desktop any igpu is a very unsatisfactory solution for gaming, when it can be beaten by a low end card like the HD7750.

[Ajay]

Originally Posted by Ferzerp

Let me specify. Someone tried to steer the conversation to how "good" amd igpus are. When all you have is a tack hammer....

Ah, yes.