• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

***Official Intel Haswell Thread***

ARM aren't replacing Intel because they aren't trying to; If they brought out desktop processors, companies like Apple would jump straight to them (They already are in the mobile market - Mac Air) and the lower licensing fees and such would quickly bring the rest of the market too. No matter how hard Intel try, they're never going to make the same mark that ARM has in that sector.

Silvermont is going to destroy ARM (and trust me, I take no pleasure from writing that); Intel were late to realise the importance of the ultra mobile market but they've finally started taking it seriously and the new Atoms are going to be seriously good.

http://www.anandtech.com/show/6936/...tecture-revealed-getting-serious-about-mobile

To say that x86 based architectures can't compete with RISC on power consumption is a fallacy which is going to be shattered in spectacular fashion; with Intel finally moving from a 5 year development cadence for Atom to an aggressive yearly tick-tock one, ARMs days of dominating the mobile sector are limited.
 
To say that x86 based architectures can't compete with RISC on power consumption is a fallacy which is going to be shattered in spectacular fashion; with Intel finally moving from a 5 year development cadence for Atom to an aggressive yearly tick-tock one, ARMs days of dominating the mobile sector are limited.

Thats a crying shame really.
 
i7-4770k hits 7.01Ghz overclock speed!!

http://www.xbitlabs.com/news/cpu/di...0K_Haswell_Overclocked_to_7_01GHz_Report.html

Journalists from Ocaholic web-site have discovered a record in the CPU-Z database that an unnamed overclocker had managed to boost the speed of an Intel Core i7-4770K microprocessor from 3.50GHz to whopping 7012.8MHz (91.07MHz BCLK*77). The overclocker used Asus Maximus VI Extreme Edition mainboard based on Intel Z87 core-logic. While the utility claims that the core voltage of the chip was upped to unprecedented 2.56V, it should rather be considered as a software bug, than a miracle since Core i-series 3000-family “Ivy Bridge” chips fail at 2.0V.
 
Silvermont is going to destroy ARM (and trust me, I take no pleasure from writing that); Intel were late to realise the importance of the ultra mobile market but they've finally started taking it seriously and the new Atoms are going to be seriously good.

http://www.anandtech.com/show/6936/...tecture-revealed-getting-serious-about-mobile

To say that x86 based architectures can't compete with RISC on power consumption is a fallacy which is going to be shattered in spectacular fashion; with Intel finally moving from a 5 year development cadence for Atom to an aggressive yearly tick-tock one, ARMs days of dominating the mobile sector are limited.

I heard that ARM and all its licensees don't do any R and D either.

I am a PC guy,but as you might realise I don't agree,because it is not really about the technology IMHO,ie,whether one CPU is 20% faster than the other or gives 10% extra battery life.

One of the reasons why companies have used ARM based SOCs is down to cost in both R and D, and actual manufacture costs. Companies are thinking of the costs of the whole device,not just the SOC used. You are forgetting the race to the bottom.

Intel is competing with loads of companies ranging from small companies with low margins which make SOCs for £40 phones to large companies with larger margins like Qualcomm who supplies SOCs for significantly more expensive devices. Combine that with the Chinese with their MIPs like chips which are homegrown and free from external export restrictions,and it is nothing like with the Windows PC market at all,which is 99% X86 still and effectively controlled by Intel at an IP level due to their ownership of the instruction set.

Some people feel X86 based chips need to be used in every computing device in the world and Intel by association for some weird reason.
 
Last edited:
Not to bothered about the headline clock speed, but rather the large range of multi's available.
The granular BCLK settings could be useful too.
 
Am I meant to be impressed?
Because it gets a resounding meh from me lol.

Me too, there's a video up on THG now of somebody getting a 4770K engineering sample to 8GHz by disabling half the cores and the HT, seriously that's like bragging about putting a man on the moon after he suffocates halfway there and lunar module crashes on landing >.>
 
Boomstick777 All depends on what You need from pc

For 3d gaming i need MORE ips than 5ghz Sandy can provide !!!!
Since TriDef software is ****** coded app my 7950 is often not at 99% usage. In many cases it sits around 60-70% gpu load when i am at 25-30fps. When 2 cores are at 100$ load and other 2 are at 40%

I know its software fault. But its been like that over 2 years now !!!
Like i posted here many times. If it can run 5ghz i will get one if not there is no point....

I would rather invest in some Golden 3770k that can do 5ghz :P
 
I am a PC guy,but as you might realise I don't agree,because it is not really about the technology IMHO,ie,whether one CPU is 20% faster than the other or gives 10% extra battery life.

The problem is that we're not talking about 20% faster or 10% extra battery life.

If Intel's slides are accurate (and they have a good record in this regard, a company the size of Intel can't afford to lie to shareholders) then we're talking about ~150% of the performance at half of the power consumption as compared to the ARM alternative; that's a pretty compelling argument for OEMs to switch from ARM to Atom, and in the ultra-competitive mobile market if your competitor has got a hardware advantage then you had damn well better follow suit.

Intel will price the new SoCs extremely aggresively (ie they will take a loss if necessary) in order to gain market share; all it will take is a few big design wins and suddenly they are the market leader and ARM go the way that AMD have since the debut of the CORE µArch.

edit: another good analysis here
http://www.realworldtech.com/silvermont/
 
Last edited:
Eagerly awaiting news on the the e3-1200 v3 xeons for server market.

I know there is NDA but surely some of the manufacturers would have hinted they will have haswell servers ready from June onwards?
 
The problem is that we're not talking about 20% faster or 10% extra battery life.

If Intel's slides are accurate (and they have a good record in this regard, a company the size of Intel can't afford to lie to shareholders) then we're talking about ~150% of the performance at half of the power consumption as compared to the ARM alternative; that's a pretty compelling argument for OEMs to switch from ARM to Atom, and in the ultra-competitive mobile market if your competitor has got a hardware advantage then you had damn well better follow suit.

You are also making the assumption that the ARM based SOCs have suddenly stopped evolving. FFS,we have not even seen the 64 bit ARM cores in action,and yet people are already saying ARM is dead.

You are reading why too much into the standard marketing hyperbole accompanied when companies talk about their latest and greatest. Look,at how much people were bigging up IB and Ultrabooks,etc and so on. The reality was very different. I expect the ARM based SOCs from competitors to be marketed the same way too. Meh.

People were saying ARM was dead for years. They are not dead yet.

No wonder half the industries in this country die,since we are so pessimistic about them surviving and give up the fight half way. Luckily,I don't see ARM doing that yet.


Intel will price the new SoCs extremely aggresively (ie they will take a loss if necessary) in order to gain market share; all it will take is a few big design wins and suddenly they are the market leader and ARM go the way that AMD have since the debut of the CORE µArch.

edit: another good analysis here
http://www.realworldtech.com/silvermont/

Which again is not relevant. Intel controlled the X86 market via its ownership of the instruction set. It only had very few companies to compete with,and even when AMD got a leg up on them they resorted to shady crap. They cannot do that now.

Intel stock is dependent on keeping high margins. That is their business,and that is shown by their market segmentation.
Its wishful thinking to think they will sell at a loss,especially when a number of their competitors have more free cash in hand themselves,and less hinging on high dollar value investments(which need costs to be covered).

Moreover,if Intel sells at a loss,do you think the competition won't do the same?? The ARM and MIPs based SOCs are cheaper to produce and develop so in the race to the bottom,who is going to do better?? Any and everybody can produce these SOCs now,from tiny startups to massive firms. Cost and easy licensing terms are the way ARM has done well.

The main problem is what happens,when the ARM and MIPs based SOCs hit an acceptable level of everyday performance. You can see that with the PC market where even cheap chips will do the job,and with more and more companies trying to use cloud based apps,you can see a trend in the market,where massively powerful CPUs won't be important for Joe and Jane Bloggs anymore,and only for power users.

Edit!!

http://beta.fool.com/rciura/2013/01/22/despite-shellacking-im-not-selling-intel/22093/

Intel is not doing too badly ATM,but they have $7 billion in long term debt and $6 billion in cash and equivalents.

http://www.bloomberg.com/news/2013-03-05/qualcomm-boosts-dividend-by-40-sets-5-billion-buyback.html

Qualcomm has cash and investments of over $28 billion and hardly any debt.

That is only one of many companies Intel is now competing with. They are hardly going to turn over and let Intel waltz into their backyard.

Its going to be a war,with the only winner being the consumer.

Anyway,we will have to agree to disagree about this.
 
Last edited:
A thought just occurred to me (yes it hurt :P), Ivy was faster clock for clock than Sandy due to better IPC, however because Ivy overclocked worse than Sandy due to the IHS issue it required its IPC advantage to match Sandy at high clocks (I.E a board/cooler that could do 5GHz with Sandy would only manage ~4.7 with Ivy so performance was basically the same) because of this the was no actual SB-IB performance difference for high overclockers.

So basically does this mean that (assuming Haswell isn't crippled on the IHS front) when overclocking we can expect to get the IB-Haswell % boost AND the missing SB-IB % boost too? If so it suddenly appears more interesting to me :P
 
Back
Top Bottom