Core i7-4960X "Ivy Bridge-E" Roughly 10% Faster than i7-3970X: Early Tests

[RejZoR]

Too long a walk I'm afraid - I live in New Zealand...although at those prices it does make me acutely aware of the cost of living at the last stop before Antarctica.

At $230, RejZor's argument is taking on more water than a cardboard submarine.

3770k is 309 EUR here and it's the cheapest importer for my country where i buy most of components. Still soaking up like a cardboard submarine? And if you do this every time for a mere 10% boost... do the math...

I'm placing big hopes for Skylake architecture when it arrives sometime next year (probably) but who knows.

 

[Fourstaff]

3770k is 309 EUR here and it's the cheapest importer for my country where i buy most of components. Still soaking up like a cardboard submarine? And if you do this every time for a mere 10% boost... do the math...

I'm placing big hopes for Skylake architecture when it arrives sometime next year (probably) but who knows.

How much was 920 in your place back when it was the most popular chip? Without knowing that there is no comparison. Regardless of what is costs, even if the 920 is cheaper than €309 the problem lies with your country's importers, not Intel. On top of that no one forces you to shell out every time Intel releases a new chip. New chip releases works wonders only for people either looking for more performance regardless of cost, or those with older chips (still rocking in their E8xxxx chips etc).

 

[RejZoR]

Digged the order mail from June 2009. Core i7 920 D0 that i bought was 262 EUR at that time (same store). And i know i paid tiny bit extra just to get the D0 version specifically. It'll soon be exactly 4 years ago when i bought it and i still think it's superb. Paired it with Antec H2O 920 cooler and overclocked it and it's up to every gaming task that i have and all the compression/encoding needs. 8 threads at 4+ GHz is not that weak despite the CPU age.

Granted, no one forces me to shell out such amount of money, but with such tiny boosts it makes even less sense. I was looking at Q6600 back then and then decided to shell out some more and take the newer Core i7 920. And i made a great decision. I don't think any of the current CPU's would last as long as this one did.

 

[AsRock]

Too long a walk I'm afraid - I live in New Zealand...although at those prices it does make me acutely aware of the cost of living at the last stop before Antarctica.

At $230, RejZor's argument is taking on more water than a cardboard submarine.

And even if you live in America don't mean ya be able to get to one our closest one is 315 miles away LMAO

 

[Fourstaff]

Digged the order mail from June 2009. Core i7 920 D0 that i bought was 262 EUR at that time (same store). And i know i paid tiny bit extra just to get the D0 version specifically. It'll soon be exactly 4 years ago when i bought it and i still think it's superb. Paired it with Antec H2O 920 cooler and overclocked it and it's up to every gaming task that i have and all the compression/encoding needs. 8 threads at 4+ GHz is not that weak despite the CPU age.

Granted, no one forces me to shell out such amount of money, but with such tiny boosts it makes even less sense. I was looking at Q6600 back then and then decided to shell out some more and take the newer Core i7 920. And i made a great decision. I don't think any of the current CPU's would last as long as this one did.

In trays of 1000s, i7 920 sold for USD284 in 2009, which is about €200 with June 2009 exchange rate. You paid €262, which is about 30% markup
In trays of 1000s, 3770K is selling for $332, which is about €255. Shop is selling for €309 for 20% markup.

Performance difference between both is much more than 20% in most cases(with or without overclock), so I would say that you are not getting an inferior product in any way (from 1000s tray prices, to exchange rates, to mark up). Inflation has not been taken to account yet. On top of that, its cheaper to assemble a system with 3770K rather than a 920 system, iirc it costs about $1500 for a fully functioning 920 system (with graphics card etc), while you will need to shell out about $1000 for an equivalent system (equivalent used extremely loosely here).

Have you considered 3570K instead? Much cheaper than 3770k, and performance wise not too far behind. Granted the upgrade from 920 to 3570K is not as significant but that is large due to the strength and performance of the 920 more than anything else.

http://www.anandtech.com/bench/Product/47?vs=551

 

[mlee49]

If Intel will keep on doing these idiotic 10% bumps for every series at full price every time, they can have them. Lazy money milking bastards.

Good news is the 3k series chips will come down in price some

edit: I'm still on X58, my 970 Hex-core has just been great for the last 3 years.

 

[ensabrenoir]

Here's hoping AMD has a 5 or 6 module Steamroller up the pipe for us by the end of the year, that should make short work of this... hopefully for less than half the price. Otherwise there are no upsides to these almost pathetic speed-bumps...

seriously though the entire cpu landscape is changing..... raw power is not the dominate factor anymore. 10% is more than enough for anything out there. Better temps and power consumption are king right now. Software is years behind. And it'll prob take a 20 CORE amd to match it. Most Intel users are fine with this and paying for a new board. ....There is not a lot of price haggling on Porsche lots...this is High End Desk Top. Mainstream is the value minded....and mainstream intel usually beats highend Amd

 

[midnightoil]

~2 year wait for a ~10% performance increase. That's pretty dire.

Shame that the desktop/enthusiast 'E' line only exists as a dumping ground for defective or very low bin Xeon-E parts. If this weren't the case, Intel might have actually skipped straight to Haswell-E ... but if they'd done that, nowhere to dump.

Really hope Steamroller and Excavator are on time and up to expectations. It's become patently clear that Intel won't make any push unless absolutely forced to.

 

[midnightoil]

Are these soldered? That would be just about the only interesting data point, seeing if a soldered Ivy runs cooler than a soldered Sandy, but beyond that we knew the performance of this chip before it even had a name.

They are soldered, yes. 'E' enthusiast / desktop parts are rebranded low-binned / defective Xeon-E parts.

 

[arterius2]

meh, 10% is good enough for me to shell out

 

[Octavean]

I think the reality is that Intel's priorities have changed somewhat.

AMD x86 / x64 chips are no longer much of a threat but ARM processors are selling like mad and the market for ARM based hardware is still growing. Whereas the PC industry is seeing negative growth. Intel knows they need to answer the ARM initiative. A big way of doing that is with decent performance and power efficiency not raw processing power. So this is likely where Intel is focusing their efforts.

Intel should be able to engineer monster performance chips for desktop, workstations, servers and so on while engineering power efficient chips for tablets, phones and so on,.... but that doesn't mean that they wont cut corners by focusing on power efficiency on the platforms that would better benefit from raw power,..

***edit***

Also an upcoming Intel Core i7 4960X Ivy Bridge-E processor is likely to cost just as much as its predecessor in the ~$1,000+ USD range. I'm not willing to spend that kind of money on a processor and judging from the system specs in this thread not many if any people here would either. Its all very academic to argue about the finer points of something you'll likely never buy,....

 

[midnightoil]

I think the reality is that Intel's priorities have changed somewhat.

AMD x86 / x64 chips are no longer much of a threat but ARM processors are selling like mad and the market for ARM based hardware is still growing. Whereas the PC industry is seeing negative growth. Intel knows they need to answer the ARM initiative. A big way of doing that is with decent performance and power efficiency not raw processing power. So this is likely where Intel is focusing their efforts.

Intel should be able to engineer monster performance chips for desktop, workstations, servers and so on while engineering power efficient chips for tablets, phones and so on,.... but that doesn't mean that they wont cut corners by focusing on power efficiency on the platforms that would better benefit from raw power,..

There's no possible way for Intel to match ARM though, and they know that. The only way they can even come close is using vastly more expensive, complicated and smaller processes ... but they're still a long way behind in performance per watt and massively behind in performance per $.

There's no way that Intel can make any inroads into the phone, tablet or embedded devices market, and ARM will continue to chew through vast swathes of the server (and soon workstation) market. The rate of market share loss is likely to increase exponentially for Intel once the the 64bit ARM chips start making their way into systems later this year.

It's going to get to the point soon where they either admit complete defeat or they start making more use of their existing ARM licenses (Intel are heavy licensees of ARM IP, contrary to popular belief).

The main reason Intel aren't pushing the envelope in desktop or E Xeons is because whilst they have a reasonable lead over AMD in absolute performance and performance per watt, they need the absolute maximum return possible on the minimum investment. Outside of the SSD business, their margins and marketshares are dropping like a stone everywhere else.

 

[Octavean]

Intel may make no inroads into the ARM segment of the market but that doesn't mean they wont try. The same goes for Microsoft and their efforts with Windows 8 and Windows RT with respect to the mobile market.

 

[midnightoil]

In trays of 1000s, i7 920 sold for USD284 in 2009, which is about €200 with June 2009 exchange rate. You paid €262, which is about 30% markup
In trays of 1000s, 3770K is selling for $332, which is about €255. Shop is selling for €309 for 20% markup.

Performance difference between both is much more than 20% in most cases(with or without overclock), so I would say that you are not getting an inferior product in any way (from 1000s tray prices, to exchange rates, to mark up). Inflation has not been taken to account yet. On top of that, its cheaper to assemble a system with 3770K rather than a 920 system, iirc it costs about $1500 for a fully functioning 920 system (with graphics card etc), while you will need to shell out about $1000 for an equivalent system (equivalent used extremely loosely here).

Have you considered 3570K instead? Much cheaper than 3770k, and performance wise not too far behind. Granted the upgrade from 920 to 3570K is not as significant but that is large due to the strength and performance of the 920 more than anything else.

http://www.anandtech.com/bench/Product/47?vs=551

the i7-920 was a much, much more expensive chip to produce, at the time, than the i7-3770k. there is a much, much higher margin for intel.

 

[RejZoR]

I don't care what was the price for 1000 pieces. I don't buy thousand of them, i only need 1. And the price of one is never this low...

 

[ensabrenoir]

There's no possible way for Intel to match ARM though, and they know that. The only way they can even come close is using vastly more expensive, complicated and smaller processes ... but they're still a long way behind in performance per watt and massively behind in performance per $.

There's no way that Intel can make any inroads into the phone, tablet or embedded devices market, and ARM will continue to chew through vast swathes of the server (and soon workstation) market. The rate of market share loss is likely to increase exponentially for Intel once the the 64bit ARM chips start making their way into systems later this year.

It's going to get to the point soon where they either admit complete defeat or they start making more use of their existing ARM licenses (Intel are heavy licensees of ARM IP, contrary to popular belief).

The main reason Intel aren't pushing the envelope in desktop or E Xeons is because whilst they have a reasonable lead over AMD in absolute performance and performance per watt, they need the absolute maximum return possible on the minimum investment. Outside of the SSD business, their margins and marketshares are dropping like a stone everywhere else.

True.....the landscape is chaning hence why intel's greater focus and resouces will be else where... But they do have the resources though so i wouldnt count them out just yet. Arm....was a larger pivot point than many realize...its evolve or die and im sure intel knows this
Myn new rig is an x79 in a cosmos 2(only posted on your pcatm) i got a 3820(on an asrock x79 exteme 9) to hold me over until the 6 core ivy-e comes out.

 

[EarthDog]

I don't think any of the current cpu's would last as long as this one did.

b o l o g n a.

 

[HD64G]

I cannot understand that everyone thinks there is a 10% bump in IPC! It isn't! Max gain is almost 10%. In the majority of the tests the gain is 5-7%. And that is sad since the clocks are the same as IB. Total gain is none to move somone to upgrade. I hope AMD's SR is what expected. Only then Intel is going to bring faster CPUs or lower the prices.

 

[tastegw]

Stock tests on these chips don't mean a whole lot, show us the overclocking comparison between the two on non engineering samples, that would make for a better article.

 

[Mindweaver]

10 percent over the old chip will be good if it overclocks as well. I mean a 10% increase and then you can bump it up another 10% with overclock would be a great upgrade, but if there isn't any overclocking head room then ho hum.. I'll wait to upgrade my i7 2600k.

 

[buggalugs]

Intel killed the highend for me with sandy-e and X79. Ivy-e doesn't look much better. Haswell looks a much better option.

 

[Octavean]

I cannot understand that everyone thinks there is a 10% bump in IPC! It isn't! Max gain is almost 10%. In the majority of the tests the gain is 5-7%. And that is sad since the clocks are the same as IB. Total gain is none to move somone to upgrade. I hope AMD's SR is what expected. Only then Intel is going to bring faster CPUs or lower the prices.

This is an Extreme class processor on an enthusiast level platform (currently X79). The prices are not going to go down.

 

[ensabrenoir]

I cannot understand that everyone thinks there is a 10% bump in IPC! It isn't! Max gain is almost 10%. In the majority of the tests the gain is 5-7%. And that is sad since the clocks are the same as IB. Total gain is none to move somone to upgrade. I hope AMD's SR is what expected. Only then Intel is going to bring faster CPUs or lower the prices.

...oh geeeshh lets not start that again

 

[Sasqui]

Seems like I'm the only person who's noticed that the 4960X has a higher base frequency, and supposedly a higher turbo frequency.

Which kinda negates the enthusiasm over the new CPUs since the older ones can be OC'ed.

You're not the only one who noticed.

3.5 ghz (3970X) vs 3.6 ghz(4960X) turbo and not being able to see the MB for the 4960X makes the whole article from coolaler.com WORTHLESS.

 

[radrok]

If Intel wants us X79 hexa users to upgrade they need to give us either an unlocked 8 core cpu or an unlocked qpi 6 core to play with on enthusiast 2p mobos.

Not worthy draining my loop for a 5% increase in IPC.

Might aswell bump my 3930k to 5,3Ghz and call it a day until Haswell-E comes.