4930K benchys early look. Conclusion.. Meh.

[Boomstick777]

Going to be careful how I link to the article as I don't want to suspended for hot linking Please excuse the lack of direct links. Photos have been uploaded to photobucket, that seems ok in the other threads..

hxxp://www.chinadiy.com.cn/html/53/n-10953.html

 

[alex_123_fra]

Overclocking is where it's at and the 4930k will have a much stronger IMC too. Stock v stock not surprising there is little difference.

 

[=XDC=FluphyBunny]

Its the difference you would expect going from sandyE to ivyE, we already know that the jump from sandybridge to ivybridge was incremental.


So same as before, no reason to upgrade unless you want to OC the bejesus out of it. Everyone else interested can wait for Haswell-E next year.

 

[SimonMaltby]

Also remember that there is a 100MHz increase in boosted speed between the two benchmarks so its not much increase at all from Ivy over Sandy.

As has been said earlier, its all about how it overclocks and handles memory.
Unless this is much better than the 3930K I can't see the point in upgrading a 3930K to a 4930K.

Personally I only have a 3820 which was always a stop gap while waiting for a 4930K and if the 4930K is not much better than a 3930K, well maybe I can pick up a cheaper 3930K when Ivy-E launches

 

[WingZero30]

Also remember that there is a 100MHz increase in boosted speed between the two benchmarks so its not much increase at all from Ivy over Sandy.

As has been said earlier, its all about how it overclocks and handles memory.
Unless this is much better than the 3930K I can't see the point in upgrading a 3930K to a 4930K.

Personally I only have a 3820 which was always a stop gap while waiting for a 4930K and if the 4930K is not much better than a 3930K, well maybe I can pick up a cheaper 3930K when Ivy-E launches

Exactly

 

[wazza300]

from sandy to ivy and ivy to haswell on average its been a 300mhz boost on clocks over last gen(same clocks)

so id expect the same on ivy-e

 

[Boomstick777]

So people seem content with 5% improvement on the new CPU's that replacing the now 2 year old CPU's. Ok but shouldn't the price be less, surely such little performance increase doesn't warrant the high prices that Intel charge for X79. I could accept 5% more performance if it meant that now using 22nm etc that it translated as 'ok well there isn't any more performance but we have better prices for the customer', but only 5% more performance after 2 years and the same prices bolstered seems pretty poor..

It's as much AMD's fault though tbh, if they could release a competitive part Intel wouldn't be able to get away with it lol.

Hopefully Haswell -E will bring some real performance increases and justify the prices..

 

[Jokester]

but only 5% more performance after 2 years and the same prices bolstered seems pretty poor..

I actually don't see it as much of a problem. People can now spend less cash to keep themselves at the top of the performance charts because today's hardware is going to last a lot longer than it would have 10 years ago.

 

[Boomstick777]

I actually don't see it as much of a problem. People can now spend less cash to keep themselves at the top of the performance charts because today's hardware is going to last a lot longer than it would have 10 years ago.

That just sounds like acceptance, we expect more from GPU's improvements year on year, so why should people be happy to pay the same prices we were paying two years ago for a product that only offers 5% more performance?

Can you imagine that working in the GPU space, two years and only 5% more perfomance

I hope AMD execute soon, and can make the market competitive again.. The stagnation sucks. APU's are more exciting.. That's how bad it is

 

[mmj_uk]

It's a much better chip all around for new adopters, the only people who will be disappointed is current Sandy Bridge-E owners who can't justify paying out again. Plus without knowing the price and overclockability we can't fully judge it yet.

 

[Martini1991]

This low gain was expected though.
Haswell-E isn't going to be much more exciting either, more of the same, as it was on 1150 from 1155.

 

[Boomstick777]

This low gain was expected though.
Haswell-E isn't going to be much more exciting either, more of the same, as it was on 1150 from 1155.

Agreed, the best thing about Haswell -E is the first true 8 core desktop Intel CPU, along with the massive reduction in power consumption, I think overall Haswell -E is a much more compelling upgrade than Ivybridge -E..

It's a much better chip all around for new adopters, the only people who will be disappointed is current Sandy Bridge-E owners who can't justify paying out again. Plus without knowing the price and overclockability we can't fully judge it yet.

I do agree with this ^^, it's a def a better chip, lower power consumption etc, it's also much cheaper for Intel to make, and it's a shame with such little performance increase that Intel won't pass that saving onto the customer, but rather bolster the same high price point. If only AMD could make a decent CPU, they could really disrupt the market and force huge price drops.. Come on AMD get Steamroller out of the gate..

 

[Martini1991]

Intel do have a problem though, if Steamroller's half decent, Intel need to bring more cores in, short of producing a 5 core chip (lol) I'm not sure how they're going to do it without messing up socket 2011 and annoying all their 39XX and Ivy-E hex users.

 

[Jokester]

That just sounds like acceptance, we expect more from GPU's improvements year on year, so why should people be happy to pay the same prices we were paying two years ago for a product that only offers 5% more performance?

Can you imagine that working in the GPU space, two years and only 5% more perfomance

But that's the thing, you don't have to pay for that 5% more. You're still effectively as fast as you can get, yet will now last effectively 2 (or even more) CPU generation iterations, whilst when there's massive gains every year you're more likely to have to pay out every time a new architecture was released to keep up.

 

[mmj_uk]

Intel would probably just lower the cost of IB-E hex cores if Steamroller is something groundbreaking, there will still be a market for Haswell with OEM's who want low tdp and onboard graphics.

 

[Boomstick777]

Intel do have a problem though, if Steamroller's half decent, Intel need to bring more cores in, short of producing a 5 core chip (lol) I'm not sure how they're going to do it without messing up socket 2011 and annoying all their 39XX and Ivy-E hex users.

Agreed, everything is setup for AMD to bring a decent part out and disrupt the current market and be competitive, it all depends on how quickly when they can launch new parts. I'm always rooting for AMD because they are the underdog, it's like they have these awesome ideas but bring them to market to late when they are starting to become irrelevant in comparison to what Intel is doing.. Hopefully AMD can get Steamroller parts out before Haswell -E and the Haswell refresh. Otherwise they might only be considered 'value' alternatives again.

AMD's APU's are awesome though, that is one area where they are doing really well..

But that's the thing, you don't have to pay for that 5% more. You're still effectively as fast as you can get, yet will now last effectively 2 (or even more) CPU generation iterations, whilst when there's massive gains every year you're more likely to have to pay out every time a new architecture was released to keep up.

That's one way of looking at it. I just see the high end CPU market as stagnated and overpriced atm, Intel do offer the best value with the mainstream Haswell stuff right now, but would be nice to see some 6 - 8 core parts that aren't priced so high..

AMD are also to blame by not providing any competition. Hopefully that will change with Steamroller..

 

[ubersonic]

so why should people be happy to pay the same prices we were paying two years ago for a product that only offers 5% more performance?

The IB-E chips will reportedly be priced slightly lower than their SB-E counterparts, plus the 5% difference is at stock and nobody buys SB/IB-E to run it at stock.

Intel do have a problem though, if Steamroller's half decent, Intel need to bring more cores in, short of producing a 5 core chip (lol) I'm not sure how they're going to do it without messing up socket 2011 and annoying all their 39XX and Ivy-E hex users.

Is that even possible? because of Intel's higher performance per core than AMD (a fact not a dig) I was under the impression they couldn't do mainstream 6/8 core CPU's as dual channel memory would bottleneck under load (it was true in tests done on Gulftown so I don't see how it couldn't be true for the more powerful Haswell architecture).

 

[Martini1991]

They can get a lot more down the dual channel now though, I mean RAM's moved on quite a bit from then, IMC's have vastly improved from Intel.

There might be a bottleneck, but what can they do?

With the Haswell refresh, it's not like they're going to bump up the 1150 base spec, unless they make 1150 another 1156 and sack it straight off.

Also, this memory bottleneck, doesn't that potentially limit just how good SR can be?

 

[Boomstick777]

The IB-E chips will reportedly be priced slightly lower than their SB-E counterparts, plus the 5% difference is at stock and nobody buys SB/IB-E to run it at stock.

Didn't know they would be priced lower, that's what I was saying, the cost reduction for Intel should be passed onto the consumer, especially with such little overall performance increase. If they do come in cheaper than that would be good. The power consumption on Ivy -E will be a lot better than Sandy -E as well. So if they are cheaper than I guess it's not such a bad upgrade. I just hope AMD bring it with Steamroller and the high end CPU market becomes more exciting / compelling to upgrade. Right now APU's are the most interesting thing in computing, that's how boring the high end has become .

 

[NirK]

Considering that the clock speed are 3-6% faster.... There is actually no improvement at all / worse performance.