Intel Skylake Core i7-6700K IPC & Overclocking Review @ [H]

Well looks like my 2500K can keep up trucking for another year it seems -- I have it at 4.3Ghz and it works just fine for me gaming at 1440p with Xfire 290s

If I do come into any upgrade money, seems to me like I'll retrofit two M2 sata drives via pcie adapters into my system for a nice 512GB worth of Raid0 action across the pcie bus. (my single 240GB ssd is getting long in the tooth)
 
Hmm. My current 3770k Ivy Bridge Z77 motherboard is screwed. I'm barely getting by with my setup. I don't know if I should upgrade or not. I could sell the Ivy Bridge and ram and recoup my losses.

I just bought 2 240GB SSDs for $120 for a Raid-0 setup so I'm all set for that. Buy a nice Skylake mobo with 9-10 SATA ports, 3x PCI-e x16 lanes. etc... Decisions, Decisions...
 
Hmm. My current 3770k Ivy Bridge Z77 motherboard is screwed. I'm barely getting by with my setup. I don't know if I should upgrade or not. I could sell the Ivy Bridge and ram and recoup my losses.

I just bought 2 240GB SSDs for $120 for a Raid-0 setup so I'm all set for that. Buy a nice Skylake mobo with 9-10 SATA ports, 3x PCI-e x16 lanes. etc... Decisions, Decisions...

??? I'm surprised that you find your 3770k lacking. Are you overclocking it? I'm not seeing much reason to upgrade as far as CPU performance goes. The new mobo features are nice and all, but not so compelling to me that I feel the need to switch out.
 
??? I'm surprised that you find your 3770k lacking. Are you overclocking it? I'm not seeing much reason to upgrade as far as CPU performance goes. The new mobo features are nice and all, but not so compelling to me that I feel the need to switch out.

I am o/c'ing to 4.4GHz. It's just that when I installed my 2nd 980Ti a few weeks ago my motherboard pretty much died. (I changed PSUs), etc. Only the 2nd and 3rd slots (980Ti ACX 2.0 cards right next to each other) work now so the top card runs @ like 90F the bottom @ 70F. I'm afraid I'm going to fry my setup.:eek:

I could buy a refurb Z77 mobo for ~$80 but since I need to replace motherboards I figure I might as well just do the whole setup.
 
I'd like to know this too. I've looked everywhere I can think of to look.

Most major etailers are showing preorders only and Aug. 14 is the only date I've seen. And believe me I've been checking everyone in North America.
 
I am o/c'ing to 4.4GHz. It's just that when I installed my 2nd 980Ti a few weeks ago my motherboard pretty much died. (I changed PSUs), etc. Only the 2nd and 3rd slots (980Ti ACX 2.0 cards right next to each other) work now so the top card runs @ like 90F the bottom @ 70F. I'm afraid I'm going to fry my setup.:eek:

I would love to have an SLI setup run at 90F/70F. ;)
 
In A.D. 2101... I'm still the last i7 920 holdout!

When faster DDR4 comes down in price I'd definitely get a 6700K and feel confident it would last for multiple years.
 
I would love to have an SLI setup run at 90F/70F. ;)

I suppose I'm not hurting it too much by running the card @ 90F because I run MSI Afterburner with a max temp @ 84F, it just downclocks both cards. If I had a functioning motherboard it would space out the cards so they wouldn't run so hot and be downclocked all the time.

Plus I've really only played Witcher 3 since I installed the 980Tis, and that game crashes a lot anyways, so it's really hard to tell exactly what is going on.:D

But I know my motherboard is bad- if I try to run 2 980Tis with a card in PCI-e slot 1 it shuts down my PC.:( I know it's not my PSU because I replaced it.

PCI-e slot 4 doesn't work at all.:( (ironically it all worked fine with 3 OG 6GB Titans).
 
Hmm. My current 3770k Ivy Bridge Z77 motherboard is screwed. I'm barely getting by with my setup. I don't know if I should upgrade or not. I could sell the Ivy Bridge and ram and recoup my losses.

I just bought 2 240GB SSDs for $120 for a Raid-0 setup so I'm all set for that. Buy a nice Skylake mobo with 9-10 SATA ports, 3x PCI-e x16 lanes. etc... Decisions, Decisions...

Skylake won't give you three x16 PCIE slots with full x16 lanes. Most are going to be 8x8x4 at best. You csn opt for a motherboard with a PLX chip but your shoving all of that data down the same x16 lanes to the CPU rendering that largely pointless.
 
In A.D. 2101... I'm still the last i7 920 holdout!

When faster DDR4 comes down in price I'd definitely get a 6700K and feel confident it would last for multiple years.

I only recently upgraded a couple of systems from i7 920s to X5675s and pretty happy with the performance. Dirt cheap upgrades and works perfectly fine as both systems have USB 3.0 and I may add some PCI-E SSDs in the future.
 
I am o/c'ing to 4.4GHz. It's just that when I installed my 2nd 980Ti a few weeks ago my motherboard pretty much died. (I changed PSUs), etc. Only the 2nd and 3rd slots (980Ti ACX 2.0 cards right next to each other) work now so the top card runs @ like 90F the bottom @ 70F. I'm afraid I'm going to fry my setup.:eek:

I could buy a refurb Z77 mobo for ~$80 but since I need to replace motherboards I figure I might as well just do the whole setup.

I wouldn't be too concerned if your cards are at 90F/70F. It's hotter than that outside.
 
I have to wonder if, for a non-gamer, that currently does a lot of encoding using QSV, if sky lake is right for me? I do some CAD and photo editing so going 4k will happen this year...but would a 5960 actually be a better option (cheaper) versus 6700+video card.

I don't usually see reviews where they over clock the cpu AND run integrated graphics. Any thoughts?

Currently using an oc-2500k.
 
Good job on the review. You make it the way it should be done, last 4 gens all at the same clock speed, then the new flagship at it's average good OC speed. Very nice.
 
I wouldn't be too concerned if your cards are at 90F/70F. It's hotter than that outside.

Yeah. I just can't overclock my cards which is kind of a bummer. Also since all my other PCI-e slots are dead I can't upgrade. I game @ 4k. I was thinking I might add a 3rd card eventually. Oh well. I guess I'll stick with it for a while.

I needed to upgrade from my 90GB SSD (I think it's broken too, could be the motherboard) to 480GB SSD anyways, now I can do that. I guess I'll find out if my SATA controller is shot.
 
Hmm. My current 3770k Ivy Bridge Z77 motherboard is screwed. I'm barely getting by with my setup. I don't know if I should upgrade or not. I could sell the Ivy Bridge and ram and recoup my losses.

I just bought 2 240GB SSDs for $120 for a Raid-0 setup so I'm all set for that. Buy a nice Skylake mobo with 9-10 SATA ports, 3x PCI-e x16 lanes. etc... Decisions, Decisions...

3770k is a beast over locked.

Watch raid 0 with SSDs. You don't really gain any performance with raid 0 for real world use (except for moving TB files or video encoding). More importantly it's been reported it doesn't always play nice and you get SMART errors for some reason. I am not well versed in raid but I've seen logical explanations to stay away with SSDs. Commuting so I can't google right now.
 
Nah, I'm just keeping my i5-2500K, let's see what AMD comes up with with their new core design. I expect it will either be excellent or atrocious, with no middle ground. Can't wait to see. But if I had to upgrade this second it would probably be Haswell-E.

Same here. My i7 2600 does everything I need it to and more, especially considering I have a "lowly" 2GB GTX 960 paired with it. Zen will either be a winner and a return to greatness (or at least competitiveness), or a flop. The last nine years of history doesn't give me much hope, considering the closest AMD has ever been to competing was Deneb and Thuban versus Yorkfield and Clarkdale - six years ago. If it's a flop, Skylake/Kaby Lake it is. Until then, I'll bide my time.

Great article as always guys.
 
Last edited:
3770k is a beast over locked.

Watch raid 0 with SSDs. You don't really gain any performance with raid 0 for real world use (except for moving TB files or video encoding). More importantly it's been reported it doesn't always play nice and you get SMART errors for some reason. I am not well versed in raid but I've seen logical explanations to stay away with SSDs. Commuting so I can't google right now.

I've done a little research on SSD Raid-0 setups. It seems mostly hit or miss. I don't plan to keep any sensitive data on there anyways, just the OS and some games. So most of it will be recoverable or backed up to the cloud (like Steam, etc). I'm not too worried. Besides, I already bought the drives. $120 for 2 240GB SSDs I couldn't resist.:D

What's a good OS migration tool?
 
I'll keep my 2600k @4.6 for now as I mainly use it for office work, browsing and occasional gaming at 1600p with a single card.

By the time DX12 games are out, we'll see if there are any actual gains with skylake and realistically, we'll likely be on a refresh with better performance and perhaps less heat.

Skylake looks to be a good setup for water cooling but running on air would make cautious as it's voltage and temps.
 
Damn, what a times when I moved from P4 to E6600 and then to 2500K. When the excitement is there, that's all it matters. As of Skylake, there is zero interest to be honest.. My Sandybridge is still going strong 24/7 since the release day..

I agree those were the days. I also went from Conroe to Sandy straight (with some laptops in between). Back in the day, you would only upgrade when everyone was losing their minds over the (gaming) performance. The only landmark before that was the coppermine pentium-III (1Ghz).

These personal desktop computers are only good for gaming. People who do actual compute work on their desktops are using workstation grade components (xeon, ECC). And most people who do actual compute work don't use desktops, there are amazon EC2 instances for that. Therefore the only real value in a consumer desktop today is gaming, which is also in a rut. 120fps 4k gaming is still by and large out of reach. Few people playing games go above 1080p 60hz and for good reason. As long as pc gaming simply tracks the console industry, pc gaming will be disappointing for enthusiasts.

So when do I think it's worth upgrading from Sandy Bridge? When we have a generation of 4k consoles.
 
Last edited:
Skylake won't give you three x16 PCIE slots with full x16 lanes. Most are going to be 8x8x4 at best. You csn opt for a motherboard with a PLX chip but your shoving all of that data down the same x16 lanes to the CPU rendering that largely pointless.

I thought I read today it was 16x and 16x and a third card would be 16x, 8x, 8x
 
Bought one myself yesterday :) RAM is a pain in the ass to find though. Every performance DDR module seems to be 50 ft tall.

This is how I feel too, I may just wait a while longer and see what happens.

I feel like the I7 5820k is the new I7 920. If it lasts me just as long I will have no issues.

I am waiting to see if these new Skylake motherboards have anything to offer that makes them more worth while..
 
The only landmark before that was the coppermine pentium-III (1Ghz).
There were alot of really good AMD procs in there, some that were also very good overclockers.


So when do I think it's worth upgrading from Sandy Bridge? When we have a generation of 4k consoles.

you might be right, but even then those 4k consoles might be heavy on pixels, light on geometry.

Nobody seems to give a shit about single socket max cpu performance but us.
 
Sigh. This thread just goes to show how much you just can't please people no matter what you do. Skylake is a better processor, period. Better IPC, support faster RAM, works on a new platform (I can always use more USB 3.0), etc. Am I going to upgrade? No, I am perfectly happy with my [email protected] with a GTX970 for work and gaming @ 1900x1200 X 2. At the same time, you are not going to hear me call Skylake a disappointment. Admittedly Intel needs to tweak the DDR memory controller if this previously mention link is correct. http://www.hardware.fr/articles/940-5/cpu-ddr4-vs-ddr3-pratique.html
But its still a better processor. I was hoping to see more in depth analysis of the processor and I don't mean 1080p/4K gaming results that some morns still want to see.
 
Now we know why Intel is planning layoffs. Still, they have balls charging $350 while admitting that there is weakening demand in the consumer market. Perhaps they are hoping people will buy these for the platform improvements (which, when you get down to brass tacks, aren't going to be very noticeable) and the anticipation that you can use the iGPU in DX12 (which won't happen btw).

It seems like the transitions to both finfet and 14nm didn't really yield any significant improvements in either performance or power-efficiency.
 
2600k @ 4.6

Not compelling enough for me. Skylake-e perhaps.

Except, Skylake-E isn't on any official roadmap that Intel has released. There is some conjecture and supposedly leaked slides showing a Skylake-E on the roadmap, but the next chip we currently know about for X99 / HEDT segment is Broadwell-E.

I thought I read today it was 16x and 16x and a third card would be 16x, 8x, 8x

I am not sure how you got that impression. The CPU has 16 lanes period. These can be divided up however a motherboard manufacturer likes. Typically what you'll see in an SLI solution is x8x8 with two GPUs. The rest of the lanes are on the PCH. Those lanes are divided up among other features such as M.2 and the like and whatever if anything that's left over can be allocated to the expansion slots. Depending on the switch chip installed on the motherboard, allocation of these can be more dynamic. You can share bandwidth in the expansion slot area with other devices and effectively give the expansion slot priority. Typically what you see or will see is 8x8x4.

If you see more than that the motherboard features a PLX chip. GIGABYTE is heavily embracing the PLX chip this generation but be warned. You will pay for it. The cost of the PLX is not insignificant. A PLX chip "adds" lanes but it does so at the expense of the lanes the CPU has. In other words you can expand your PCIe lanes to 32, but your still shoving all the data from your expansion slots and thus GPUs, down the same x16 lanes the CPU offers. Ultimately you gain latency and the PLX can act as a PCIe lane switch, but again your still limited by the PCIe lanes of the original platform and any benefits come at a slight latency penalty.

Again, there is not an insignificant cost involved. The Anandtech article I linked is old, but most of the information is still relevant. I happen to know the PLX 8747 is cheaper than the article details now, but not by much.

And looking at the manual for a PLX'less Z170 motherboard I'm still seeing 8x8x4 in the manual.
 
This review is ALL OVER THE PLACE.

1) Why use Windows 7 for tests? Why not the more stable, efficient and better optimized Windows 8.1? Better yet, why not Windows 10? What doesn't work properly yet?

2) Why use a Titan? It's a generation old and 3 flagships old GPU. The 780 Ti, 980 or 980 Ti would've been better choices. Also, why use 1 year old drivers? 320.xx drivers? Seriously ?

3) Why are using OLD software to benchmark? Cinebench 11.5? Winrar 4? What in the.... ?

4) In gaming benchmarks, DON'T TEST for LOW settings at 480p. WTH. 1080/1440p tests with max in-game settings. If you want to test the CPU, just use FXAA for AA and voila !

5) You're seriously testing a game from 2007. Why again ?


It's all nice and dandy that 6700K is 30% faster overall than a 2600K at 4.5 GHz, but you DID NOT use realistic settings, tests or ANYTHING. This whole review is a myth in my eyes.
 
Except, Skylake-E isn't on any official roadmap that Intel has released. There is some conjecture and supposedly leaked slides showing a Skylake-E on the roadmap, but the next chip we currently know about for X99 / HEDT segment is Broadwell-E.

Last year the Haswell-E dropped a few months after consumer launch (devils canyon).

Any chance we see that this year?
 
This is pathetic. 25% improvement over a 4+ YEARS OLD CPU! And for over $700 (CPU, mobo, RAM)!

Either Moore's law is truly dead or Intel is not even trying anymore.
 
This is pathetic. 25% improvement over a 4+ YEARS OLD CPU! And for over $700 (CPU, mobo, RAM)!

Either Moore's law is truly dead or Intel is not even trying anymore.

Intel is not trying.

They are focused on performance per watt.

This is what happens when AMD sucks. We NEED AMD.
 
Except, Skylake-E isn't on any official roadmap that Intel has released. There is some conjecture and supposedly leaked slides showing a Skylake-E on the roadmap, but the next chip we currently know about for X99 / HEDT segment is Broadwell-E.



I am not sure how you got that impression. The CPU has 16 lanes period. These can be divided up however a motherboard manufacturer likes. Typically what you'll see in an SLI solution is x8x8 with two GPUs. The rest of the lanes are on the PCH. Those lanes are divided up among other features such as M.2 and the like and whatever if anything that's left over can be allocated to the expansion slots. Depending on the switch chip installed on the motherboard, allocation of these can be more dynamic. You can share bandwidth in the expansion slot area with other devices and effectively give the expansion slot priority. Typically what you see or will see is 8x8x4.

If you see more than that the motherboard features a PLX chip. GIGABYTE is heavily embracing the PLX chip this generation but be warned. You will pay for it. The cost of the PLX is not insignificant. A PLX chip "adds" lanes but it does so at the expense of the lanes the CPU has. In other words you can expand your PCIe lanes to 32, but your still shoving all the data from your expansion slots and thus GPUs, down the same x16 lanes the CPU offers. Ultimately you gain latency and the PLX can act as a PCIe lane switch, but again your still limited by the PCIe lanes of the original platform and any benefits come at a slight latency penalty.

Again, there is not an insignificant cost involved. The Anandtech article I linked is old, but most of the information is still relevant. I happen to know the PLX 8747 is cheaper than the article details now, but not by much.

And looking at the manual for a PLX'less Z170 motherboard I'm still seeing 8x8x4 in the manual.

Sorry Dan. Thanks for clarifying.
 
Intel is not trying.

They are focused on performance per watt.

This is what happens when AMD sucks. We NEED AMD.

I think your off base here. Intel is focused on performance per watt because energy savings is what the industry demands in both the server and mobile markets. The consumer / enthusiast desktop market is literally the only one that doesn't give a shit about that.
If Intel could push out a 6GHz CPU tomorrow that would get all your 2500/2600K holdouts to finally buy something from them again don't you think they would? Competition from AMD and others is good. No one would argue against that but it isn't the only reason for Intel to try. Intel needs customers' repeat business. To do that they have to offer products you want so that you spend money with them again. I think they just haven't been able to get more clock speed out of the silicon, but have had some success at improving efficiency and power with the existing clocks they've had. Intel seems to consider the few clock speed concessions they've made in recent years worth while for the mobile and server markets.

For the most part, they are probably right.

Last year the Haswell-E dropped a few months after consumer launch (devils canyon).

Any chance we see that this year?

I haven't heard anything of the sort. Honestly I won't know until very close to a product launch and I'd be under NDA anyway if I did know.

Speculation: I don't think we will see that at all. Broadwell-E was pushed back to 2016. There are some leaked slides out there showing a possible cancellation of Broadwell-E in favor of Skylake-E. While that would be nice it would likely push a release even further out. As one website pointed out, HEDT processors are based on Xeon E5 series cores. Intel would likely have to cancel Broadwell-EP and develop a Skylake-EP in it's place. Intel doesn't normally alter Xeon roadmaps too much. Server customers are a different breed and require long validation times and QVL testing to ensure compatibility in their platforms.

I don't see Skylake-E coming to replace Broadwell-E/EP unless Intel seems to think they can do it and keep their server customers happy. If anything I think we'll see an unusually long cycle for an HEDT CPU refresh.

Sorry Dan. Thanks for clarifying.

No biggie. :)
 
It's shame we can't buy a i7-5775 (c or r), I think I'd buy that for my upcoming HTPC if I could. Looks like the vapor broadwell chips would be better.
 
I think your off base here. Intel is focused on performance per watt because energy savings is what the industry demands in both the server and mobile markets. The consumer / enthusiast desktop market is literally the only one that doesn't give a shit about that.

If Intel could push out a 6GHz CPU tomorrow that would get all your 2500/2600K holdouts to finally buy something from them again don't you think they would?

I think if Intel re-prioritized the way they design chips we could get our old generational jumps back.

But to your first point, 99% of the people buying CPUs do care enormously about performance per watt. I mean there is no consumer desktop market, it's dead, and it's never coming back. It's really only enthusiast and prosumer workstation stuff now. I mean look how long Apple lets their desktops languish.

To the point about AMD, I'd guess that they could find a niche with us instead of trying to compete for that much bigger market. But it's likely a small slice of the whole pie would be better than a big slice of the enthusiast market.

So in the end I do agree, if Intel could ship a 6ghz chip that also made for great laptops and high-end server chips they would. But since we are the one market were performance per watt doesn't matter we get left in the cold.
 
If Intel could push out a 6GHz CPU tomorrow that would get all your 2500/2600K holdouts to finally buy something from them again don't you think they would?

No, I don't, simply because Intel appears to have made a quick back of the envelope style calculation that the money required to create said 6GHz CPU would not be profitable enough to warrant the effort. And I think they're wildly wrong about that, especially since the market in question is low volume/high margin.
 
No, I don't, simply because Intel appears to have made a quick back of the envelope style calculation that the money required to create said 6GHz CPU would not be profitable enough to warrant the effort. And I think they're wildly wrong about that, especially since the market in question is low volume/high margin.

I don't. I think if they could create 6GHz chips with reasonable thermals they would be upping the turbo clocks and overclocking headroom at least gradually each generation. They haven't been doing that.

I've spoken to people from Intel. They are indeed aware of the fact that tons of 2500K and 2600K users have had zero reason to upgrade and flat out haven't. They know there is a current 3-5 year life cycle on PC gaming CPUs and motherboards. Intel wants to change that and get more of your money. Oviously hasn't made a compelling enough product for many of you to do that. I think if they could have given us 5GHz overclocks on air they would have. Intel has promised us this before and always fallen short.
 
No, I don't, simply because Intel appears to have made a quick back of the envelope style calculation that the money required to create said 6GHz CPU would not be profitable enough to warrant the effort. And I think they're wildly wrong about that, especially since the market in question is low volume/high margin.

I doubt they are wrong about that. The money to design the processor is astronomical and enthusiasts are notoriously cheap. We're the guys that buy the $110 celeron 300A instead of the $500 PII 450mhz.
 
Back
Top