• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Are you upgrading to Skylake?

Page 6 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

Upgrading to Skylake-S?

  • Yes, absolutely!

  • No way!

  • Not sure.

Results are only viewable after voting.
jdubs03 said:
That's an interesting find, what is surprising is that they had Skylake for only 2 quarters, seems quite a bit short to me. I would expect Skylake to be on the market for at least a year before Cannonlake would succeed it.
Click to expand...

It doesn't make sense, so it won't happen.

They are in it to make money, and OEMs are too. Actually the latter wishes are far more important. If they realize Skylake is a 6 months product, they'll skip it and use Cannonlake. I've seen these kind of rumors far too many times. Do you think these companies that have margins <10% are going to go for half the expected product lifespan? That's crazy!

Unless there's another caveat. Maybe Skylake is a higher end SKU and Cannonlake a low end one, due to low yielding 10nm process in the beginning. Or maybe limited Skylake availability with mass availability on Cannonlake. If that's true, neither is good. Either 14nm sucks or 10nm sucks.
 
Last edited:
IntelUser2000 said:
It doesn't make sense, so it won't happen.

They are in it to make money, and OEMs are too. Actually the latter wishes are far more important. If they realize Skylake is a 6 months product, they'll skip it and use Cannonlake. I've seen these kind of rumors far too many times. Do you think these companies that have margins <10% are going to go for half the expected product lifespan? That's crazy!

Unless there's another caveat. Maybe Skylake is a higher end SKU and Cannonlake a low end one, due to low yielding 10nm process in the beginning. Or maybe limited Skylake availability with mass availability on Cannonlake. If that's true, neither is good. Either 14nm sucks or 10nm sucks.
Click to expand...

Cannonlake is platform compatible. So there isnt any change as such.
 
If you apply the usual 5-10% performance boost and some performance/watt benefits. Then its not that hard. The platform, core count etc is all known.
 
Last edited:
I'm lazy and skipped a few pages, so completely missed the benchmarks on Skylake. Are they awesome?

Without seeing some test results, tough to make a an upgrade decision. Strange Skylake isn't shown in the Anand comparo database - and they call themselves a tech journalism site. Pffft...with no tech psychics on staff how can we possibly make informed future speculation purchases?

So...can I answer this question in late 2017, early 2018(ish), when Skylake-E might actually be shipping. For sure. Possibly. Maybe.

Sucks about the AVX512 exclusion - market segmentation or ? Of course, since AVX2 has been so widely embraced in the desktop world, we could squeak by with it for a few more tickety tocks.

;-)
 
coolpurplefan said:
Is this a big deal? http://www.kitguru.net/components/c...or-pcs-will-not-support-avx-512-instructions/

It says Skylake won't be supporting AVX-512 instructions. Looking at the table below, it looks like it's the same as Haswell. That could be good for compatibility but I don't know what's the real difference.
Click to expand...


We know it since a year or so. And no it isn't a big deal.

The hardware to support AVX-512 is in the processors
Click to expand...

A typical Shilov posting. It isn't certain that AVX-512 is in hardware because even Xeon E3 won't have it. Only high-end Xeon some years later.
 
My 1155 mb burnt up last week, so after spending 150 getting it back up and running, it will probably be after all of the newness wears off before I do another build.
 
Dave2150 said:
What makes you believe that's a real skylake result?

I assume anyone could fake a result like this if they really wanted to?
Click to expand...


The ID belongs to Skylake platform. No idea if this can be faked but wouldn't make sense.


Deders said:
Bear in mind if that result is skylake, it's only clocked at 2.6GHz. I think we can expect nearly double the clockspeed upon release.
Click to expand...


Over 5 Ghz? Surely not. We should be happy if SKL could match Devil's Canyon clock speed.
 
When this is due, Q2/Q3?

Not many engineering samples spotted out in the wild?

Redentor said:
A little news about skylake: http://www.bitsandchips.it/52-engli...lar-to-the-transition-from-prescott-to-conroe

Not bad at all, if true.
Click to expand...
Can it walk the talk? If it's Prescott-to-Conroe like, sure as hell, Intel would have been a bit more vocal. No AVX 3.0 on mainstream chips? Hmm. Is that "TSX" thing has been fixed yet?

ShintaiDK said:
If you apply the usual 5-10% performance boost
Click to expand...
I hope it is more, than that.

3) With the FIVR* removed, it would be interesting to see what kind of power consumption / performance it has in idle/sleep states.

* - If that is removed, to only come back later, looks like just another compromise to me. If Skylake v1 doesn't overclock well, I will definitely skip. Afraid to ask, is the mainstream "K" model going to use indium solder? :hmm:
 
Last edited:
Deders said:
Bear in mind if that result is skylake, it's only clocked at 2.6GHz. I think we can expect nearly double the clockspeed upon release.
Click to expand...
More than 5Ghz? This is even more unlikely than a brand new micro-architecture.

My point is that even at 2.6 GHz this would be similar to Haswell scores at similar frequency, so the masked claim of the article referenced above that IPC will skyrocket is unlikely.

Of course I don't claim anything, I'm just guessing 🙂
 
Magic Carpet said:
When this is due, Q2/Q3?

Not many engineering samples spotted out in the wild?

If it's Prescott-to-Conroe like, sure as hell, Intel would have been a bit more vocal. No AVX 3.0 on mainstream chips? Hmm.


I hope it is more, than that.

3) With the FIVR removed, it would be interesting to see what kind of power consumption / performance it has in idle/sleep states.
Click to expand...

Yea, I agree. I seriously doubt we will see a conroe like jump, but I am trying to remain hopeful of a decent improvement. Skylake will be a 2 generation jump though, so anything less than 15 percent would be less than from SB to Haswell.

I am concerned about the lack of leaks, seems to me that either the performance is not up to what they hoped for or there will be more delays, or both.
 
frozentundra123456 said:
Yea, I agree. I seriously doubt we will see a conroe like jump, but I am trying to remain hopeful of a decent improvement. Skylake will be a 2 generation jump though, so anything less than 15 percent would be less than from SB to Haswell.

I am concerned about the lack of leaks, seems to me that either the performance is not up to what they hoped for or there will be more delays, or both.
Click to expand...
Well for me, it's clear, my current 4770K is a poor overclocker. So, if the Skylake unlocked SKU overclocks reasonably well, has good temps and can offer at least 10% IPC boost in every CPU metric, out of the box, I will most likely upgrade to it. If Intel wants my money, of course.

P.S. Sad, that they decided to part with VGA, though. I still have a few VGA-only panels around for misc. use.
 
Last edited:
Magic Carpet said:
Well for me, it's clear, my current 4770K is a poor overclocker. So, if the Skylake unlocked SKU overclocks reasonably well, has good temps and can offer at least 10% IPC boost in every CPU metric, out of the box, I will most likely upgrade to it. If Intel wants my money, of course.

P.S. Sad, that they decided to part with VGA, though. I still have a few VGA-only panels around for misc. use.
Click to expand...

I think it's nice to be out with the old, in with the new. Good riddance, VGA!
 
frozentundra123456 said:
So if you have a dgpu with a VGA connector would that work?
Click to expand...
Yeah, it works fine with a dgpu, no problem at all. I like to use 2-4 monitors depending on task at hand, 2 connected to onboard (VGA+DVI) and another two connected thru the dgpu (GTX 670, again VGA+DVI). No issues with drivers or excessive power consumption (HD 4600 works really well w/ dgpu). With Skylake, I might have to get another DVI (hard to find a quality 16:10 panel these days, not a fan of 16:9) monitor instead.
 
Last edited:
NTMBK said:
Maybe, depending on personal finances and how well my rig holds up. Phenom II 965 and HD7770, 6GB DDR2. It keeps on playing all the games I want it to, so it's hard to justify a rebuild!
Click to expand...

I was in the same boat as you with a PII 955 and R9 270 but then 3 weeks ago my CPU died after less than 4 years of use. Now that i have a Devil Canyon 4790K i will definitely skip Skylake.
 
Magic Carpet said:
When this is due, Q2/Q3?

Not many engineering samples spotted out in the wild?
Click to expand...

Intel is real tight with the ES chips nowadays, it seems. That or they have NDAs with big nasty teeth (compared to previous NDAs). Intel assassin squads? Who knows.

Afraid to ask, is the mainstream "K" model going to use indium solder? :hmm:
Click to expand...

Don't count on it. Both Intel and AMD have switched to TIM under the IHS for everything that isn't a server CPU or a derivitive thereof (hence the solder under the IHS of FX chips). I don't even know if the "stuff" under the IHS of HEDT Haswell-E chips counts as solder, it's more of a metalized epoxy or something, but it seems to work okay.

Anyway plan on delid/relid operations or suffering the TIM madness. Learn to embrace the CLU.
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top