Intel Skylake / Kaby Lake

[ShintaiDK]

Argh, only 1 M.2 slot on the Stinger >.<

 

[phillyman36]

So far I see 2 z170 mobo at my local Microcenter. Asus z170AR And A Gigabyte board. No cpus up yet. With Asus would post their info so i can see what is what.

 

[Borealis7]

looks like TPU got a much better chip than AnandTech, their chip does the MSI Z170 auto-OC to 5.00GHz and boots into windows

http://www.techpowerup.com/reviews/MSI/Z170A_GAMING_M7/7.html

 

[AtenRa]

Well, Skylake GT2 even with DDR4 is still slower than Kaveri 65W A8-7600.

If you want DDR-4 you better go for the 2011-V3 and a real 6-core.

 

[witeken]

Very nice power as well, 77W.

They also hit 5Ghz.

That's quite a difference compared to the Broadwell review a few days ago.

 

[coercitiv]

Looks like Broadwell's L4 cache is the real star in today's reviews.

 

[BFG10K]

http://www.tomshardware.com/reviews/skylake-intel-core-i7-6700k-core-i5-6600k,4252-9.html

So where's the "performance destroying IGP" we were supposed to be getting? The slides were quoting 50% faster. It's slower than Broadwell; just realized it doesn't even have the 128MB eDRAM!

I guess they were comparing it to the Haswell IGP, which is the same perf % inflating stunt they pulled with Broadwell Iris instead of comparing it the previous Iris.

 

[ShintaiDK]

So far I see 2 z170 mobo at my local Microcenter. Asus z170AR And A Gigabyte board. No cpus up yet. With Asus would post their info so i can see what is what.

New boards keep popping up here. Gigabyte ITX etc. Plenty of CPUs in stock.

Will see what I will pick, EVGA completely dumped it.

 

[Abwx]

Some really serious reviewing :

http://www.computerbase.de/2015-08/intel-core-i5-6600k-i7-6700k-test-benchmark-skylake/

 

[ShintaiDK]

From the favourite hardware.fr.

 

[Kallogan]

just looking at anand review, it seems like the most useless upgrade ever.

Makes broadwell look good imo

overclocking seems very decent tho

 

[cytg111]

What the actual ..hell? Thats a pretty huge *meh* for me.
ASUS PB287Q, here i come.

 

[AtenRa]

http://www.tomshardware.com/reviews/skylake-intel-core-i7-6700k-core-i5-6600k,4252-9.html

So where's the "performance destroying IGP" we were supposed to be getting? The slides were quoting 50% faster. It's slower than Broadwell; just realized it doesn't even have the 128MB eDRAM!

I guess they were comparing it to the Haswell IGP, which is the same perf % inflating stunt they pulled with Broadwell Iris instead of comparing it the previous Iris.

Even against Haswell GT2 is not more than 15-20% faster on average.
Still slower than 1.5 year old 65W Kaveri.

 

[cytg111]

The real disaster is to come IMO.
kabylake will be more of the same.
10nm cannonlake delayed AND possible more of the same.
zen will be a disaster.
we all screwed.
Looking forward to the next 10 years of anantech forums being the same 10 ppl engaged in the same trolldom flameshite arguing the same damn numbers.
Whatever.

 

[ShintaiDK]

Even against Haswell GT2 is not more than 15-20% faster on average.
Still slower than 1.5 year old 65W Kaveri.

Thats not what his link shows.

 

[coercitiv]

overclocking seems very decent tho

It is decent, but it also uses more power than Haswell at stock settings. Granted efficiency kept up the pace, but where is the 14nm advantage?

The real disaster is to come IMO.
kabylake will be more of the same.
10nm cannonlake delayed AND possible more of the same.
zen will be a disaster.
we all screwed.

We're not screwed: that 4k monitor looks sweet.

 

[Sweepr]

From the favourite hardware.fr.

~23.5% better performance per clock than Sandy Bridge, they will add Haswell results soon.

Looking at their Core i5 6600K and Core i5 4690K averages (same base and turbo clocks) Skylake is 8-10% faster per clock (just like inf64 predicted). Core i5 5675C is still slightly faster than Core i5 6600K in CPU limited games but Core i7 6700K beats Core i5 5765C. eDRAM will provide a nicke boost to LGA Skylake next year.

 

[witeken]

Spoiler alert: there will never be huge leaps in performance again for CPUs unless there's some fundamental transistor and interconnect breakthrough allowing for much higher clock speeds with same power consumption.

So too bad there's not yet a full deep dive, although that's expected.

 

[Sweepr]

It is decent, but it also uses more power than Haswell at stock settings. Granted efficiency kept up the pace, but where is the 14nm advantage?

According to Computerbase it uses +25W less power than Core i7 4790K @ CPU load. Not sure about other reviews.

http://www.computerbase.de/2015-08/intel-core-i5-6600k-i7-6700k-test-benchmark-skylake/8/

 

[Tovarisc]

Anadtech forgot to add graphs into their review

http://www.anandtech.com/show/9483/intel-skylake-review-6700k-6600k-ddr4-ddr3-ipc-6th-generation/21

Edit: apparently I'm blind and missed crucial bit of info, nvm!

 

[AtenRa]

Thats not what his link shows.

yea Half Life 2 DX-9, you can play that with Sandy/IvyBridge HD

 

[AtenRa]

Anadtech forgot to add graphs into their review

http://www.anandtech.com/show/9483/intel-skylake-review-6700k-6600k-ddr4-ddr3-ipc-6th-generation/21

nope,

Ryan Smith - Wednesday, August 05, 2015 - link

We still have graphs going up (it's a LOT of data), so if you see any missing graphs, we're working on it. In the mean time try to hold tight.=)

 

[BigDaveX]

I'm guessing the IGP's drivers still need some work, as results are all over the place at the moment; in some games it's even slower than a Haswell GT2, and in others it's noticeably faster.

Getting away from the performance situation for the moment, UK retailers are all pricing the 6700K in line with the 5775C, rather than the 4790K. I only hope this is a case of them scalping early upgraders rather than a permanent increase in the price of the higher-end i7s, because if it's the latter, frankly it's even more of a reason to go with a Haswell-E setup.

 

[Sweepr]

HD Graphics 530 clearly needs some driver work but just like Ivy Bridge back at launch there's potential.

 

[coercitiv]

According to Computerbase it uses +25W less power than Core i7 4790K @ CPU load. Not sure about other reviews.

http://www.computerbase.de/2015-08/intel-core-i5-6600k-i7-6700k-test-benchmark-skylake/8/

Techreport and Tomshardware say different. They might all be right, with good and bad, since one reports Cinebench power usage, one encodes x264 and the last measures power usage while gaming on discrete cards. Nevertheless, I was expecting better results power wise (at least stay within 4790K envelope in each case).