• Guest, The rules for the P & N subforum have been updated to prohibit "ad hominem" or personal attacks against other posters. See the full details in the post "Politics and News Rules & Guidelines."
  • Community Question: What makes a good motherboard?

Intel Skylake / Kaby Lake

Page 299 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

Abwx

Diamond Member
Apr 2, 2011
9,117
902
126
That's a good increase for just a new stepping, in fact the shift from Broadwell to Skylake wasn't higher afaik. Of course the improvements comes also from the improved 14nm process.
Sysmark, WebXPRT in the slide posted by ShintaiDK...

As for the improvement in clock it come mainly from power management if we are to follow what was exposed at HChips.
 
Mar 10, 2006
11,719
2,003
126
Sysmark, WebXPRT in the slide posted by ShintaiDK...

As for the improvement in clock i come mainly from power management if we are to follow what was exposed at HChips.
It's an improved process (14nm+) and a better implementation of the same basic Skylake core.
 
  • Like
Reactions: Sweepr

VirtualLarry

Lifer
Aug 25, 2001
49,529
5,594
126
LOL. SYSmark and WEBxpert. Two Intel-controlled benchmarks.

Edit: Yeah, I realized that's basically still OK, since they're comparing like with like, after I posted.
 
Last edited:
Mar 10, 2006
11,719
2,003
126
LOL. SYSmark and WEBxpert. Two Intel-controlled benchmarks.
They're comparing Intel processor to Intel processor, though...and they're the same exact architecture.

What these benchmarks show is that these chips can hit higher frequencies and sustain them for longer.
 
Aug 11, 2008
10,457
641
126
LOL. SYSmark and WEBxpert. Two Intel-controlled benchmarks.
As usual, I am sure they picked a best case benchmark. However, it is comparing intel to intel, so vendor optimization might not play such a big part.
More the question to me is can they maintain the faster clocks under sustained load.

Edit: looks like Arach beat me to it. I agree though.
 
Mar 10, 2006
11,719
2,003
126
As usual, I am sure they picked a best case benchmark. However, it is comparing intel to intel, so vendor optimization might not play such a big part.
More the question to me is can they maintain the faster clocks under sustained load.

Edit: looks like Arach beat me to it. I agree though.
Yeah, I would say vendor provided benchmarks are pretty much best case. Anything else would be downright foolish on the part of marketing. Reviews will show the average case.
 

JoeRambo

Senior member
Jun 13, 2013
913
647
136
I am following this thread, but I guess Intel still does not want to sell me that unlocked Skylake* quad with 128MB of EDRAM. ( 4770K user here). Oh well, that's lost business for them and i guess there are plenty of others who would love it.

*any Lake would do.
 
  • Like
Reactions: beginner99
Mar 10, 2006
11,719
2,003
126
Two words: "Disney Vault". Hopefully, Intel isn't following in that path.
I doubt it. Anyway, fab cycle time is around 3-4 months for these leading edge processes, so my guess is that the chips recently PRQ'd and are now running through the fabs. Makes sense for them to launch in January as that's when those wafers will roll off the line.
 

Sweepr

Diamond Member
May 12, 2006
5,151
1,127
131
Digital Foundry: Let's build a 1080p60 gaming PC

DF using an i5-6600K in their 1080p60 build. Some intesresting bits about using fast DDR4 and i5 vs i7 here:

In terms of memory, we'd recommend spending a little extra on fast DDR4. In CPU-bound scenarios, the faster your memory, the less constricted your processor is. This goes against the conventional wisdom that 'any RAM will do' but you can see the evidence in the Core i5 2500K video embedded further down in this piece. The 3000MHz-rated Corsair Vengeance LPX DDR4 [?] is fast enough but overclocks to 3200MHz with laughable ease. Beyond that, additional DDR4 bandwidth offers limited returns based on our testing.
The Core i5 6600K is a great chip, but we did find some scenarios where an i7 would improve performance. Crysis 3 on very high settings (shadows dropped to high) could still stutter below 60fps in the jungle areas, or when heavy alpha and intense physics were deployed. Even with the i5 at 4.0GHz, the vintage 2013 title still causes issues for modern hardware. We've played this game a lot, and even older generation i7s hand in smoother performance on the very high preset. We also noted that the village stage in Rise of the Tomb Raider could max out the i5, causing stutter (a touch reduced in DX12), while galloping through Novigrad City could also tap out the i5 to its limits (though the 60fps lock remained intact).

Helping to prolong the life of PCs past, present and future is the fact that current - and indeed, future - consoles are based on relatively weak AMD CPU cores designed primarily for mobile applications. PlayStation Neo features the exact same CPU set-up, albeit with a 31 per cent clock-speed increase. We feel fairly confident that an i5 6600K can keep up for a good few years yet. On top of that, the Z170 board should be compatible with the future Intel Kaby Lake CPUs, plus you can upgrade from an i5 to an i7 for further gaming performance.
www.eurogamer.net/articles/digitalfoundry-2016-lets-build-a-1080p60-gaming-pc



Deus EX: Mankind Divided CPU Performance





Those of you running a Core i5 or Core i7 processor have little to worry about -- even the old Sandy Bridge 2500K kept up with newer models, not to mention the Core i3-6100 that managed to match the Ivy Bridge Core i5-3470 and beat the FX-9590 by a slim margin.

The only processors to really suffer were Intel's dual cores without Hyper Threading and AMD's Athlon X4 860

Not a lot to say here, CPU utilization wasn't even at 50% on the 6700K when clocked at 2.5GHz. Overclocking your Intel processor for better frame rates in Deus Ex: Mankind Divided is pointless.
http://www.techspot.com/review/1235-deus-ex-mankind-divided-benchmarks/page5.html



http://gamegpu.com/action-/-fps-/-tps/deus-ex-mankind-divided-test-gpu
 
Last edited:

witeken

Diamond Member
Dec 25, 2013
3,876
154
106
Core i7-7Y75 has a boost to 3.6GHz. I presume it's still 4.5W TDP, so that's crazy. And I thought the Snapdragon series boosts to 2.5GHz or so were already a lot. The burst speed of this SoC will be enormous.
 

nerp

Diamond Member
Dec 31, 2005
9,829
76
91
Mobile parts don't seem like a huge leap, really. People who just bought skylake shouldn't really care about these new parts at all. I love how they claim 12 percent performance increase yet the clocks in the comparison are about 12 percent higher for KBL. Duh.
 

witeken

Diamond Member
Dec 25, 2013
3,876
154
106
Mobile parts don't seem like a huge leap, really. People who just bought skylake shouldn't really care about these new parts at all. I love how they claim 12 percent performance increase yet the clocks in the comparison are about 12 percent higher for KBL. Duh.
Whut? How else would you want to improve performance besides higher clock speed? I don't understand this comment at all.
 

VirtualLarry

Lifer
Aug 25, 2001
49,529
5,594
126
Whut? How else would you want to improve performance besides higher clock speed? I don't understand this comment at all.
I think that what he's trying to point out, is that if you equalize clocks between SKL and KBL, there's virtually no IPC increase. Which shouldn't be too surprising.
 
Mar 10, 2006
11,719
2,003
126
Mobile parts don't seem like a huge leap, really. People who just bought skylake shouldn't really care about these new parts at all. I love how they claim 12 percent performance increase yet the clocks in the comparison are about 12 percent higher for KBL. Duh.
The performance increase comes from an enhanced 14nm process that allows for higher clocks at a given level of power consumption. Don't downplay this, this is a good achievement from Intel. The process and circuit implementation folks really did a good job here.

We'll see the micro-architects strut their stuff in Cannonlake.
 
  • Like
Reactions: Phynaz

ASK THE COMMUNITY