Intel Skylake / Kaby Lake

[Sweepr]

Also I don't know how credible that site is, (im not familiar with it) but this roadmap was floating around here in May reported by numerous sites with no broadwell-e on it but it was deemed fake. I wouldn't be surprised if they got a hold of the same one..

http://www.kitguru.net/wp-content/uploads/2015/05/intel_roadmap_skylake_cannonlake_10nm1.jpg

I doubt it.
No Braswell or Apollo Lake reference, Cannonlake @ Q3-2016? Also by then I'm pretty sure Broadwell-E and Broadwell-EP/EX were still in the cards since the leaked slides detailing Purley and other server products belong to an Intel presentation @ late May.

According to the server roadmap Broadwell-EP and Broadwell-E would be launched in early 2016 with no successors in sight till sometime in 2017 (Purley platform), doesn't look like Q1 to me. That would be a very long wait, while I don't expect Skylake server/enthusiast parts as soon as Broadwell-EP/E they could very well arrive sometime in H2-2016, most likely sooner than before.

 

[moonbogg]

Skylake-E coming early would be sick. By then my funds will be recuperated and I'll be all ready to FINALLY get off Sandy Bridge. I love my Sandy Bridge though so this is also kind of saddening.

 

[mikk]

Also I don't know how credible that site is, (im not familiar with it) but this roadmap was floating around here in May reported by numerous sites with no broadwell-e on it but it was deemed fake. I wouldn't be surprised if they got a hold of the same one..

http://www2.warwick.ac.uk/services/...pspcs/supplierpresentations/viglen_launch.pdf


This is no fake. But not relevant anymore: 3rd April 2014

 

[tenks]

http://www2.warwick.ac.uk/services/...pspcs/supplierpresentations/viglen_launch.pdf


This is no fake. But not relevant anymore: 3rd April 2014

Ah nice find, thank you! That roadmap caused me a lot of grief a few weeks ago on various forums. A bunch of crappy sites got a hold of that and were using that as the newest roadmap. There were various inconsistencies that led me to beleive it was fake but in the end, its just old and outdated.

The funny thing is they all reported Broadwell-E was gone, but it was based off an old roadmap, and the irony is now they're right, but by accident. lol

 

[dark zero]

So broadwell ended to be an aborted generation?

 

[moonbogg]

So broadwell ended to be an aborted generation?

God I hope so. I want it dead because that means Skylake-E and MY WALLET will align with each other in a nice, timely fashion

 

[Sweepr]

Skylake-E coming early would be sick. By then my funds will be recuperated and I'll be all ready to FINALLY get off Sandy Bridge. I love my Sandy Bridge though so this is also kind of saddening.

I really like the fact that mobile, desktop and server will be using the same x86 core next year (assuming Kabylake packs Skylake cores instead of a new x86 core). Earlier rumours led us to believe there would be Broadwell-E, Skylake Refresh and mobile Cannonlake around the same time.

So broadwell ended to be an aborted generation?

In servers, most likely yes. Haswell-EP ended up being better than most people predicted, while rumours were talking about 14-core Intel delivered a 18-core monster. That's 50% more than Ivy Bridge-EP (12-core) which was 50% more than Sandy Bridge-EP (8-core). Broadwell-EP was rumoured to pack 22 cores, only 22.2% more than current Haswell-EP at unknown clocks. Maybe it didn't meet Intel's expectations.
With up to 28 cores, 6-channel DDR4, new memory technology and a bunch of important platform advancements server Skylake is a much better successor.

 

[Fjodor2001]

So now Skylake-E is interesting all of a sudden because it has moar coares? Wasn't that what Intel used to mock AMD for... The irony...

 

[Arachnotronic]

So now Skylake-E is interesting all of a sudden because it has moar coares? Wasn't that what Intel used to mock AMD for... The irony...

It has "moar" very fast cores.

 

[Fjodor2001]

Slower in ST performance than 4790K.

Also a lot of the complaints with moar cores was that the extra cores would not be used anyway, due too poorly multithreaded SW. If you believe that to be the case, then that argument would hold true regardless of how fast each core is.

 

[ShintaiDK]

You try too hard.

More cores on server = good.

 

[Sweepr]

So now Skylake-E is interesting all of a sudden because it has moar coares? Wasn't that what Intel used to mock AMD for... The irony...

We're talking about server Skylake, not Skylake-E. There's no info on Skylake-E core count yet.

Slower in ST performance than 4790K.

You must be basing this on your hopes, because there's no way to determine how Skylake stacks up compared to Haswell based on the leaks so far (we've had 0 to more than 15% faster per clock depending on the leak) and nobody knows Skylake-E final clocks.

By the way, it wouldn't surprise me if Skylake-E made it to the market earlier than Summit Ridge and with a 28-core server version on its way who knows how Intel will shape the enthusiast desktop chips to kill AMD's hope in this segment.

 

[LTC8K6]

Slower in ST performance than 4790K.

Also a lot of the complaints with moar cores was that the extra cores would not be used anyway, due too poorly multithreaded SW. If you believe that to be the case, then that argument would hold true regardless of how fast each core is.

Is there any reason to bet the farm on the leaks?

 

[Sweepr]

EVGA shows off Z170-series motherboards ahead of Intel Skylake release

Intel’s Skylake processors are on the verge of release; while we don’t know exactly when they’ll arrive, it may be as early as August 5th, and will be no later than September. Motherboard manufacturers are showing off their latest-and-greatest in preparation. EVGA has provided details and high-resolution photos of three motherboards; the Z170 Stinger, the Z170 FTW, and the Z170 Classified.

Z170 Classified

Z170 FTW

Z170 Stinger

www.digitaltrends.com/computing/evg...therboards-ahead-of-intel-skylake-release/#/3

 

[Excessi0n]

Slower in ST performance than 4790K.

Probably, but it's kind of silly to compare a massively multicore server chip to what is effectively a factory-overclocked quad-core.

Also a lot of the complaints with moar cores was that the extra cores would not be used anyway, due too poorly multithreaded SW. If you believe that to be the case, then that argument would hold true regardless of how fast each core is.

These aren't consumer chips running consumer software. These are professional chips running professional software. Pay attention.

 

[Hugo Drax]

95% of user desktops run 99% idle most of the time. Intel is gonna have a hard time keeping sales up. There is not much justification to upgrade workstations on a 4 year cycle anymore.

 

[ClockHound]

EVGA shows off Z170-series motherboards ahead of Intel Skylake release
....

Intel's 'Shroud of Secrecy' around skylake seems to have extended to their partners design choices... ;-)

 

[iSkylaker]

Haven't seen a single mATX Z170 board just yet, its scaring me.

 

[Burpo]

ECS Z170-Blade

Gigabyte GA-B150M-D3H

Gigabyte GA-Z170M-D3H

Gigabyte_H170M-D3H
http://www.tomshardware.com/gallery/gigabyte_H170M-D3H,0101-500876-15231-0-0-0-jpg-.html

"Asustek will also offer two – Z170M-E and Z170i Pro Gaming – motherboards for compact systems in micro-ATX and mini-ITX form-factors respectively."

 

[dark zero]

We're talking about server Skylake, not Skylake-E. There's no info on Skylake-E core count yet.



You must be basing this on your hopes, because there's no way to determine how Skylake stacks up compared to Haswell based on the leaks so far (we've had 0 to more than 15% faster per clock depending on the leak) and nobody knows Skylake-E final clocks.

By the way, it wouldn't surprise me if Skylake-E made it to the market earlier than Summit Ridge and with a 28-core server version on its way who knows how Intel will shape the enthusiast desktop chips to kill AMD's hope in this segment.

We don't know even the real power of AMD... however we might predict the real power of Intel in this generation and it depends heavily on cores.... and in ST is not as great as expecting... maybe Intel is preparing the Xeon Phi to be a obligatory upgrade with their big cores.

ECS Z170-Blade

Gigabyte GA-B150M-D3H

Gigabyte GA-Z170M-D3H

"Asustek will also offer two – Z170M-E and Z170i Pro Gaming – motherboards for compact systems in micro-ATX and mini-ITX form-factors respectively."

VGA and PS2 are still alive. Old Tech is still usable until now.

 

[iSkylaker]

Thanks for the pics Burpo I saw some of those before, but didn't think/notice they were Z-chipset.

 

[Fjodor2001]

You try too hard.

More cores on server = good.

Probably, but it's kind of silly to compare a massively multicore server chip to what is effectively a factory-overclocked quad-core.

These aren't consumer chips running consumer software. These are professional chips running professional software. Pay attention.

Seems like you guys missed a few post. We were discussing the option of moving to Skylake-E, since mainstream Intel desktop performance improvements have been so poor for several years, and the future looks bleak too.

So we were not talking about Skylake-E for server use, but for desktop use. You do know the HEDT CPUs are not only for server use, right?

And then the question arises what good "moar cores" will be on desktop now all of a sudden, when its an Intel CPU that is providing those cores.

 

[steve wilson]

Seems like you guys missed a few post. We were discussing the option of moving to Skylake-E, since mainstream Intel desktop performance improvements have been so poor for several years, and the future looks bleak too.

So we were not talking about Skylake-E for server use, but for desktop use. You do know the HEDT CPUs are not only for server use, right?

And then the question arises what good "moar cores" will be on desktop now all of a sudden, when its an Intel CPU that is providing those cores.

Software is getting better all the time with multi threading. I think it's a bit of chicken and egg. Why produce good multi threading software, when there isn't the hardware to do it. Why produce good multi core chips when there isn't the software to run it. I'll be glad when Intel releases chips with more cores so the software has to catch up, which is a little ironic as it has traditionally been the other way around.

 

[ShintaiDK]

So we were not talking about Skylake-E for server use, but for desktop use. You do know the HEDT CPUs are not only for server use, right?

Right, they are also for workstation usage

 

[ShintaiDK]

EVGA shows off Z170-series motherboards ahead of Intel Skylake release

Z170 Stinger

www.digitaltrends.com/computing/evga-shows-off-z170-series-motherboards-ahead-of-intel-skylake-release/#/3

I posted a few of the EVGA series too.

Its very nice to see the confirmation of the M.2 slot on the back of the Stinger. The slot next to the battery is a mPCIe riser if anyone is interested. So put the WiFi there