Any information about overclocking? :) Or is this NDA at the moment? I think you have the first review online so well done with the time management!
Stuen4y
Any information about overclocking? :) Or is this NDA at the moment? I think you have the first review online so well done with the time management!
Says on the test config page that their engineering sample chip (not the same as one you'd buy in store) is an unexpectedly bad overclocker and they can't push it further. We're still waiting on retail devil's canyon for overclocks and proper temperatures. From this it seems delidding is still likely the only way though…
CampGareth
Says on the test config page that their engineering sample chip (not the same as one you'd buy in store) is an unexpectedly bad overclocker and they can't push it further. We're still waiting on retail devil's canyon for overclocks and proper temperatures. From this it seems delidding is still likely the only way though…
I do not know with what mind does Intel ship ES samples of overclocker tailored chips that will not perform in the most important aspect of the whole product as the retail samples… Thanks for the info though.
At least it shows they don't give out their best binned chips though..
The Pentium G3258 is £59.99 on pre-order:
http://www.overclockers.co.uk/showproduct.php?prodid=CP-531-INOTH,once you add the cost of a cooler and a Z series motherboard,it might be better just to get a Core i3 or a FX6300 with a £50 motherboard. In some games,like Thief,Watch Dogs and BF4 MP,the Pentium dual cores do badly indeed(even with Mantle) according to the articles I have seen.
However,it might be good for more lightly threaded games,and negate the need for more expensive CPUs.
I think I'll be upgrading my 4670 to this, so long as ASUS come out with a BIOS update to support it.
Newer games have started to take advantage of more cores. I noticed with my i3 that performance wasn't as good as benchmarks showed with my GPU at the time.
MrRockliffe
Newer games have started to take advantage of more cores. I noticed with my i3 that performance wasn't as good as benchmarks showed with my GPU at the time.
Same here. I had a socket 1155 Core i3,which did very well at the time,but in some newer games it started to struggle,so changed it over to a Xeon E3 instead.
I bought a second hand 3470 which has been working flawlessly since I bought it.
I need to see what's under that IHS
This is an underwhelming release, at least until overclocking results prove otherwise.
If the underlying silicon is the same as any other Haswell refresh CPU, why should the engineering sample be any different from a retail chip? Doesn't the engineering sample have the upgraded thermal interface material? Sadly the review is a bit pointless until they can test retail CPUs.
Jimbobgod1969
If the underlying silicon is the same as any other Haswell refresh CPU, why should the engineering sample be any different from a retail chip? Doesn't the engineering sample have the upgraded thermal interface material? Sadly the review is a bit pointless until they can test retail CPUs.
Well at least they got the review out weeks before anyone else,as Hexus is the only site to have reviewed the Core i7 4790K.
The thing is though,Intel sends ES samples to reviewers anyway,not retail samples AFAIK.
Tarinder - have offered the use of my FX9590 , board and ram before for and AMD comparison ;)
CAT-THE-FIFTH
Well at least they got the review out weeks before anyone else!!
True! They even beat Anandtech to the punch :)
Jimbobgod1969
If the underlying silicon is the same as any other Haswell refresh CPU, why should the engineering sample be any different from a retail chip?
Because an engineering sample isn't a reviewer's sample, it's a sample that's used by the engineers to check the chip before it goes to retail production. We've seen several steppings of the same ES chip in the past when erratia has been found.
The retail production run might have very minor alterations that make a huge difference. In some cases it can be small changes in the architecture, in others a simple microcode update (which alone can make a huge difference).
Why set the New K-processors vs the AMD's APUs? Why not put them up against the FX-series? Sure, they lack IGP, but they perform better that the AMD-K's
Hardly beats my i7 920. wake me up when Intel finally makes a CPU thats 70-90% faster clock for clock in every way.
So in conclusion My next PC build includes Devil's Canyon and Z97 motherboard because it feels right and my current build power hungrier than I thought.
abychristy
So in conclusion My next PC build includes Devil's Canyon and Z97 motherboard because it feels right and my current build power hungrier than I thought.
The power consumption difference is not going to make any realworld difference,unless you spend your entire life gaming. Think more of the performance upgrade.
virtuo
I think I'll be upgrading my 4670 to this, so long as ASUS come out with a BIOS update to support it.
With you on this one.
abychristy
So in conclusion My next PC build includes Devil's Canyon and Z97 motherboard because it feels right and my current build power hungrier than I thought.
If there's a bios update I'll be taking one of these and my new maximus formula vi
Agent
The retail production run might have very minor alterations that make a huge difference. In some cases it can be small changes in the architecture, in others a simple microcode update (which alone can make a huge difference).
I have a feeling that with those binned chips, people will be back to batch hunting. :D
Does the Intel Core i7-4790K have the Intel visualization tech that wasn't in the 4770k but was in the 4770?
j.o.s.h.1408;3288677
Hardly beats my i7 920. wake me up when Intel finally makes a CPU thats 70-90% faster clock for clock in every way.
Seconded!
CampGareth
Stuen4y
Any information about overclocking? :) Or is this NDA at the moment? I think you have the first review online so well done with the time management!
Says on the test config page that their engineering sample chip (not the same as one you'd buy in store) is an unexpectedly bad overclocker and they can't push it further. We're still waiting on retail devil's canyon for overclocks and proper temperatures. From this it seems delidding is still likely the only way though…
Usually its the opposite. Intel purposely sends out samples that overclock really well to get better reviews…
Ross1
Usually its the opposite. Intel purposely sends out samples that overclock really well to get better reviews…
People have claimed they overclock better for years. People have claimed they overclock worse for years. People have claimed they're about the same as retail. The slightly smarter people have claimed they are the luck of the draw, like retail.
The truth is no one knows or has any evidence for any of them.
Agent
People have claimed they overclock better for years. People have claimed they overclock worse for years. People have claimed they're about the same as retail. The slightly smarter people have claimed they are the luck of the draw, like retail.
The truth is no one knows or has any evidence for any of them.
Well we do have one bit of evidence at least regarding Ivy Bridge. Engineering samples used different dies to retail samples:
http://www.chip-architect.com/news/2012_04_19_Ivy_Bridges_GPU_2-25_times_Sandys.html
CAT-THE-FIFTH
Well we do have one bit of evidence at least regarding Ivy Bridge. Engineering samples used different dies to retail samples:
That's evidence they used different dies, not that they tried to manipulate benchmarks to show better overclocking.
This kind of thing is
expected. They are
samples for checking the
engineering of the chip. I don't see how it has anything to do with benchmark manipulation :)
Agent
That's evidence they used different dies, not that they tried to manipulate benchmarks to show better overclocking.
This kind of thing is expected. They are samples for checking the engineering of the chip. I don't see how it has anything to do with benchmark manipulation :)
The problem is that the bigger die was a production sample of an IB Core i5,obtained by the company which took the pictures. The earlier picture is of an engineering sample(many reviews use these like Hexus for example),meaning there is a distinct
possibility that reviews were testing a chip not representative of what you could actually buy. My major concern is not overclocking but things like power consumption for example.
j.o.s.h.1408;3288677
Hardly beats my i7 920. wake me up when Intel finally makes a CPU thats 70-90% faster clock for clock in every way.
I think those days are gone. Application specific accelerators will get added in, like we already have video encoding, encryption, visualization & positional audio. Other than that, we are seeing deeper queues, more cache and tweaks to the pipelines. Process shrinks don't give automatic performance improvements any more. Give it another 5 years and I expect x86 to be 20% faster per clock and maybe touching 5GHz, hardly enough to notice.
I'm surprised Intel haven't tweaked the encryption instructions for BitCoin. That might cause some ripples :)
Was going to sell my i7 2700k and asus formula v mobo to upgrade with this new chip. Now wondering whether it is worth it.
Kovoet
Was going to sell my i7 2700k and asus formula v mobo to upgrade with this new chip. Now wondering whether it is worth it.
Honestly,with things like Mantle and DX12 in the next year,together with the main multi-platform engines threading better,your Core i7 2700K has some life left in it still,especially if overclocked IMHO.
Agreed. Even a 920 will do
CAT-THE-FIFTH
The problem is that the bigger die was a production sample of an IB Core i5,obtained by the company which took the pictures. The earlier picture is of an engineering sample(many reviews use these like Hexus for example),meaning there is a distinct possibility that reviews were testing a chip not representative of what you could actually buy.
It's also a
possibility that the ES chip was simply part of their process to work towards a final retail chip. The change could have been last second for all we know.
I'm not trying to defend Intel here, I really don't care about them any more than AMD, but there are so many
possibilities and general speculation that it's impossible to state that they were intentionally misleading anyone.
DanceswithUnix
I'm surprised Intel haven't tweaked the encryption instructions for BitCoin. That might cause some ripples :)
I really don't think it would make a difference given the insane speed and mass production of ASIC based systems now. By the time they got into peoples hands, and people were generating coins, the value would be of little consequence unless a huge part of the die was dedicated to it.
Agent
I really don't think it would make a difference given the insane speed and mass production of ASIC based systems now. By the time they got into peoples hands, and people were generating coins, the value would be of little consequence unless a huge part of the die was dedicated to it.
Really should have googled that before I posted :D
https://software.intel.com/en-us/articles/intel-sha-extensionsSo SHA support goes into Skylake. Now that isn't going to make you rich, but if it is energy efficient enough to have it ticking away while you are surfing etc and it makes rather than loses you money then potentially every PC can be adding to the global hash rate.
Got to be at least more useful than Physix? ;)
How hot would you expect a 4770K to run under the stock cooler and being fed 1.152v? The review mentions the 4790K ran at 88 degrees compared to 78 degrees for the 4770K, but that was with the 4770K at 1.08v
CAT-THE-FIFTH
Kovoet
Was going to sell my i7 2700k and asus formula v mobo to upgrade with this new chip. Now wondering whether it is worth it.
Honestly,with things like Mantle and DX12 in the next year,together with the main multi-platform engines threading better,your Core i7 2700K has some life left in it still,especially if overclocked IMHO.
I have the 2600k running arount 4.5Ghz and was also wondering if it was time to upgrade, I guess I shall hang onto my cash for a bit longer.
Why not use the latest stable or beta Catalyst drivers though… Catalyst 14.4 (the latest stable version) has been out for at least a month and there's also 14.6 beta availlable. Unless there's only one motherboard available for APUs at the lab, shouldn't be that hard to update, nor is it particularly hard to add the non-present devices system variable and remove old display adapters from device manager.
APUs aside, shouldn't there be a comparison with AMD's FX offerings, binned or not?
eatonm62
I have the 2600k running arount 4.5Ghz and was also wondering if it was time to upgrade, I guess I shall hang onto my cash for a bit longer.
I've done it for one reason. The formula vi is a gorgeous board and happy with it. Got it the 4770k running at 4.2ghz easily but the best thing is the new bios