Leaked Intel i7-6700K Benchmarks?

I never had much faith in Skylake being a major leap forward. I always assumed it's main selling features would be better on-chip graphics, power and thermals. That's why I went ahead and jumped on a 5820K a few months ago.
 
Old_Way said:
I never had much faith in Skylake being a major leap forward. I always assumed it's main selling features would be better on-chip graphics, power and thermals. That's why I went ahead and jumped on a 5820K a few months ago.
Click to expand...

Don't tempt me bro! ;)
 
david_ said:
Everyone needs to accept that performance gains like the 2500k are the exception, not the rule.
Click to expand...

And that only happened because the i5 2500k had much improved efficiency, which opened-up clock speeds available. Previously, the i5 760 was power-limited to 2.8 Ghz (95w), so the boost to 3.3 GHz was the majority of the performance increase over Nehalem.

We also saw a large boost with the Haswell refresh Core i7 to 4.0 GHz base clock, so don't act like we haven't seen an impressive increase since Sandy Bridge.
 
Those gains look to be about what I would expect. The days of big gains are over, unfortunately. At least on the desktop.. :(
 
Performance has not increased significantly since Core i7 920 which could easily be clocked at 4.2 GHz on air.
 
For anyone who'd like to know, I did some creative guessing to calculate that the 6700K might be about twice the speed of an Intel i5-4570. I might end up going for the 6700 (non-K) to get the lower power consumption (65 watts). I don't need 100% increase but I'd be happy with 50 to 65% gain.
 
defaultluser said:
And that only happened because the i5 2500k had much improved efficiency, which opened-up clock speeds available. Previously, the i5 760 was power-limited to 2.8 Ghz (95w), so the boost to 3.3 GHz was the majority of the performance increase over Nehalem.

We also saw a large boost with the Haswell refresh Core i7 to 4.0 GHz base clock, so don't act like we haven't seen an impressive increase since Sandy Bridge.
Click to expand...
Not sure if the second part is directed at me.. I'm arguing for upgrades.. and have a 4770k.
 
chris719 said:
Performance has not increased significantly since Core i7 920 which could easily be clocked at 4.2 GHz on air.
Click to expand...

Really? Did you just say that? Maybe you meant to put Core i7-2600k.... Even the 4790k is about 15% faster than a 2600k (not necessarily in gaming keep in mind). The 2600k was a good step above a 920 as well with major gaming improvements.

There has not been much chance since the 2500k/2600k though for sure plus the overclocking headroom of the those 2 chips is still the best.
 
If these results are true, then doesn't Skylake gives around roughly a similar increase over Haswell that Sandy Bridge did over Bloomfield? ...when you look at it from the perspective of closest-spec tier offering to tier offering (close in clock speed, caches, and cores/threads); i7-975 being replaced with the i7-2600K.
 
roma said:
I just upgraded last year - to a new E8600 ($60 off Amazon). Haven't been held back in any way. I'll need to upgrade for 4K video editing though.
Click to expand...

Sorry you got ripped off like that.
A G3258 is about the same price and anywhere from considerably faster to waayyyy faster.
 
Kueller said:
Sorry you got ripped off like that.
A G3258 is about the same price and anywhere from considerably faster to waayyyy faster.
Click to expand...

Its no contest, the G3258 is better in every way, even the non-overclock pentiums are better.
 
Yeah, after waiting so long I pulled the trigger on an $80 Xeon X5660 for my comp (replacing a i7-920.) At this rate, I'll prolly change platforms in 2018 - 2020 sometime... bleh!!!
 
Trimlock said:
Its no contest, the G3258 is better in every way, even the non-overclock pentiums are better.
Click to expand...

Those Pentiums are a bargain. I went with a G2020+Motherboard for less than $100 going on two years ago. Still quite impressed with the performance even without being able to overclock.
 
Dew said:
Those Pentiums are a bargain. I went with a G2020+Motherboard for less than $100 going on two years ago. Still quite impressed with the performance even without being able to overclock.
Click to expand...

I have to admit I'm blown away with what those CPU's can do. Even the J1900 which was a stop-gap power saver is great in certain areas.

I really want to replace my G3258 in my primary rig (funny its my slowest CPU too) but I can't justify the few hundred.
 
"Skylake movement curve associated with the following processors possible , we have prepared performance graphics. Intel i7 6700 benchmark test that takes place in this comparison were prepared on the basis of performances given by Intel's processors in the previous year."

These aren't benchmarks at all. They're guesstimates.
 
extide said:
Those gains look to be about what I would expect. The days of big gains are over, unfortunately. At least on the desktop.. :(
Click to expand...

I think those days are over because of factors that indirectly impact technological improvement rather than being like the result of limited research or developmental capabilities though. I think mostly its due to...

-Little to no competition driving development in the desktop market

-Few software applications that benefit greatly from increases in personal computing processor power

-A general shift toward power/heat/space constrained mobile computing

Intel has no need to drive desktop computing and no competitive forces posing a credible threat at the moment. Instead they're focusing their efforts on bringing down power demand by improving efficiency and managing consumption along with focusing on improving integrated graphics.

The good news about that is that desktop computers don't really need processor upgrades as often since software applications don't require it and new CPUs don't cause current ones to be obsolete right away. Those aren't bad things really unless you have an itch to upgrade, but can't find a good excuse to do so because processors improvements don't justify it.
 
CreepyUncleGoogle said:
I think those days are over because of factors that indirectly impact technological improvement rather than being like the result of limited research or developmental capabilities though. I think mostly its due to...

-Little to no competition driving development in the desktop market

-Few software applications that benefit greatly from increases in personal computing processor power

-A general shift toward power/heat/space constrained mobile computing

Intel has no need to drive desktop computing and no competitive forces posing a credible threat at the moment. Instead they're focusing their efforts on bringing down power demand by improving efficiency and managing consumption along with focusing on improving integrated graphics.

The good news about that is that desktop computers don't really need processor upgrades as often since software applications don't require it and new CPUs don't cause current ones to be obsolete right away. Those aren't bad things really unless you have an itch to upgrade, but can't find a good excuse to do so because processors improvements don't justify it.
Click to expand...

I would suspect that in a community such as this one that a higher percentage of people are doing things with their computers that would benefit from faster processing speeds than would be the case with the general public. The general public, however, are the predominant buyers of technology, so your assessment is accurate. The lack of competition is not necessarily hurting the general public because the market is over-saturated with $300 Wal-Mart special craptops and they simply don't know any better, but enthusiasts are especially disappointed with said lack of competition.
 
MrCaffeineX said:
I would suspect that in a community such as this one that a higher percentage of people are doing things with their computers that would benefit from faster processing speeds than would be the case with the general public. The general public, however, are the predominant buyers of technology, so your assessment is accurate. The lack of competition is not necessarily hurting the general public because the market is over-saturated with $300 Wal-Mart special craptops and they simply don't know any better, but enthusiasts are especially disappointed with said lack of competition.
Click to expand...

Yeah, I agree with that. The general tone of the thread makes it apparent that a lot of people would prefer having a larger than incremental increase, but I'm not sure about the actual benefit. Lots of replies indicate that people are okay with keeping their Nehalem or Sandy Bridge which might indicate that there's not a lot consensus about the material benefits of more compute power.
 
CreepyUncleGoogle said:
Yeah, I agree with that. The general tone of the thread makes it apparent that a lot of people would prefer having a larger than incremental increase, but I'm not sure about the actual benefit. Lots of replies indicate that people are okay with keeping their Nehalem or Sandy Bridge which might indicate that there's not a lot consensus about the material benefits of more compute power.
Click to expand...

You might be on to something. Even if they could easily boost power by 100%, they may get a lot of immediate sales, but then people would just sit on those systems for a long time if Intel didn't continually boost speed by good amounts. Since they currently have a lock on the market, there is little incentive for them to give us more than crumbs.
 
John P. Myers said:
These aren't benchmarks at all. They're guesstimates.
Click to expand...
That's also what I thought, but didn't trust the terrible translation enough to make a definitive claim. As I mentioned on the first page, the only Skylake chips showing up in benchmark submissions are low clocked engineering samples. I don't know why people here are treating the link as real benchmarks.
 
DocSavage said:
You might be on to something. Even if they could easily boost power by 100%, they may get a lot of immediate sales, but then people would just sit on those systems for a long time if Intel didn't continually boost speed by good amounts. Since they currently have a lock on the market, there is little incentive for them to give us more than crumbs.
Click to expand...

But also the faster the processing power available, the more processing is used by many devs as advanced options which then migrate to standard features.
The lack of faster cpus does hold back software improvements which then holds back sales of faster processors.
Intel probably dont know how to make best use of the market yet.

The lack of a competing AMD has hurt us a lot though.
I bought about about 7 AMD CPUs over the years but theres no point now.
 
The problem with intels incremental updates is that they mostly are measured based on stock clocks. But for quite a while we have been able to get good enough overclocks out of processors like sandy bridge that we don't care about those increases in speed since we can just clock another 100 MHz to offset it. Intel needs to produce a new chip that either is a lot better, or gives us the overclocking headroom to see big gains.


A lot of this is also because software isn't pushing hardware. This is because everyone is trying to convince consumers that less is more. If you are making a laptop why sell a person a gaming laptop when you can convince them to give you higher profit margins with crappy ultrabook? If you are a software developer why make software that does more and craves more performance when you know a bunch of these consumers will be running shit on crap atom tablets or glorified netbooks they call ultrabooks? And on top of that you might have divided your programming talent up in order to port your software to 10 different platfroms including all the mobile ones.
 
Could be that Intel sees Moore's Law as no longer sustainable given current tech. Until a commercially viable breakthrough is achieved, they will just dole out incremental improvements to keep things moving somewhat forward.
 
Nenu said:
The lack of a competing AMD has hurt us a lot though.
I bought about about 7 AMD CPUs over the years but theres no point now.
Click to expand...

Competition from AMD has never had much impact on Intel. Its not like they rushed out anything to replace the Pentium 4 or cut prices when the Athlon 64 came out. They just leaned on OEMs to keep AMD from gaining market share. AMD just isnt big enough to matter.
 
We wouldnt have had the performance level of the 2500K so early without competition from AMD.
 
When are we expecting to see the motherboards? One of the reasons i'm looking to upgrade from my 2500k is the feature set, PCIe 3.0, m.2 and usb 3.1. What are estimates on the jump from 2500k to 6600k?
 
Nenu said:
We wouldnt have had the performance level of the 2500K so early without competition from AMD.
Click to expand...

I think we would have. Sandy Bridge was on the drawing boards a hell of a long time before it was released in 2011...they showed external demos as early as 2009.
 
DejaWiz said:
I think we would have. Sandy Bridge was on the drawing boards a hell of a long time before it was released in 2011...they showed external demos as early as 2009.
Click to expand...

And they could have sat on it for a lot longer.
 
Well sandy jump was huge, and we didnt have things called smartphones tablets etc. Nowdays the smartphone life is approximately 2 years. Faster developing and more lucrative for a big corp like intel. Hence their spending all their resources coming up to speed in the mobile sector.

I had this discussion on Atech as well, and yes the 6700K is a yawn for those who recently upgraded. However for those still chugging along with say a Core 2 Quad or even 965x maybe its enough to kick their current platform.
 
Smashing Young Man said:
Could be that Intel sees Moore's Law as no longer sustainable given current tech.
Click to expand...
1) Moore's law, as it has evolved through the years, relates to the number of transistors/density doubling every so many months (18). Intel seems to use that as a metric to drive process innovation and has kept on track for decades.

2) Intel has changed its focus. Performance improvements must be balanced against power costs. Unfortunately this is heavily weighted towards mobile on non-handheld processors.

3) While maintaining compatibility with existing x86 software, there is little low hanging fruit to tackle anymore. Even adding more execution units to the CPU core gives diminishing returns. The buggy TSX was an attempt to improve performance in other ways and the future points to more x86 extensions providing boosts than somehow making the core instruction set significantly faster each generation. ;)
 
Nenu said:
And they could have sat on it for a lot longer.
Click to expand...

Not a chance. No other follow-up product was developed to replace the CPU architecture of 1156 and 1366 platforms...Intel needed to get their iGP-equipped products on the market.
 
DejaWiz said:
Not a chance. No other follow-up product was developed to replace the CPU architecture of 1156 and 1366 platforms...Intel needed to get their iGP-equipped products on the market.
Click to expand...

There wasnt any need to release it as early and they could have started off at slower speeds.
Its widely acknowledged that the performance of the 2500K and 2600K was a surprise.

If it wasnt for competition up until then, we would have had lower performing parts trickling out.
Exactly like we are now getting because there is little competition.
 
DeathFromBelow said:
Competition from AMD has never had much impact on Intel. Its not like they rushed out anything to replace the Pentium 4 or cut prices when the Athlon 64 came out. They just leaned on OEMs to keep AMD from gaining market share. AMD just isnt big enough to matter.
Click to expand...

That's really not true and yes they most certainly did push to get Conroe (it and it's successors) out to replace the P4 and the prices that you enjoy now on the desktop are because of AMD. Also the P4 has been the shortest run of just about any architecture Intel has ever done. Even Itanium lasted longer than the P4.

Intel used to be far more expensive than they have been the last two generations. On the desktop generally Intel has been keeping the prices pretty damn low. The 2P market on the other hand has gone through the roof. When AMD was competitive Intel's prices were more aligned than they are now.
 
We definitely wouldn't of had the P4-D without AMD coming out with the Athlon 64X2 and kinda blindsiding them, but of course those proc still pretty much sucked. Heck if we didn't have AMD at all, Intel might have stuck with netburst and never even switched to Conroe! That would be a sad day. We would of had the Tejas CPU..!
 
Nenu said:
There wasnt any need to release it as early and they could have started off at slower speeds.
Its widely acknowledged that the performance of the 2500K and 2600K was a surprise.

If it wasnt for competition up until then, we would have had lower performing parts trickling out.
Exactly like we are now getting because there is little competition.
Click to expand...

I still don't see where all this "performance surprise" comes from regarding SB. Take a spec comparable prior-gen 1156/1366 product vs an 1155 product...in this case an i7-975 vs i7-2600K, and the SB gains are about what these benchmarks (if true) show for this particular Skylake over it's Haswell predecessor. I covered that in an earlier post...

Seriously, what do people expect to get that classifies as an "acceptable" minimum performance gain when a new processor is released that, oftentimes, can be put on an existing MoBo/socket? 20%? 25%? 33%?

So Intel (or AMD, at that) gives us 5-15% with every new gen, and also adds usable features (let's not focus on the shitty ones, like TIM paste under the IHS starting with IB) A gain is a gain for anyone that upgrades their home CPU every generation. If that isn't good enough to keep our games running great, then focus on what actually needs the focus lately: the GPU. If gaming is not a given person's cup of tea and they are using processor intensive programs that love a lot of cores/threads, then bitching about gains for new Mainstream consumer processors is just plain ignorant...there are 6 and 8 core processors in Intel's Performance consumer lineup and even more cores in their Enterprise lineup. Granted, some can come with heavy price tags, but they do exist...they are real...and they are even available for purchase. Crazy times we live in, huh?
 
July 1st is the launch apparently according to some sites today;) damn that came fast.

Ddr4 prices fell as well looking good
 
95 freaking watts? How the heck is this thing going to overclock past 4.3
 
You must log in or register to reply here.
Back
Top