Intel Skylake / Kaby Lake

Page 504 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

vissarix

Senior member
Jun 12, 2015
297
96
101
Are we reading the same charts / reviews?!
What charts are you reading? Where is the review for the overclocked i7 7800X?

Lets use logic for a moment....Delidded i7 7800X with a proper cooler can do 5ghz...Make it 4.8ghz without delid but with a good cooler..

1eamfs.png


As you can see the stock i7 7800x is already very close to ryzen flagship...ryzen has little oc headroom while the i7 7800x have a lot...So even when all cores will be utilized the i7 7800x will end up faster...






You have been warned that unsubstantiated claims only cause more problems and is considered flamebait. You are trolling this forum. and it is not allowed.


esquared
Anandtech Forum Director
 
Last edited by a moderator:

DDH

Member
May 30, 2015
168
168
111
No thermal results and no overclocking on the anandtech review? Otherwise very good

How can the 7800x use almost the same power as the 7900x with 40% less cores though?
 

Markfw

Moderator Emeritus, Elite Member
May 16, 2002
25,478
14,434
136
What charts are you reading? Where is the review for the overclocked i7 7800X?

Lets use logic for a moment....Delidded i7 7800X with a proper cooler can do 5ghz...Make it 4.8ghz without delid but with a good cooler..

As you can see the i7 7800x is very close to ryzen flagship...ryzen has little oc headroom while the i7 7800x have a lot...So even when all cores will be utilized the i7 7800x will end up faster...

You are just guessing, and I have warned you before on this.

KNOCK IT OFF !! STOP the trolling !
 

DrMrLordX

Lifer
Apr 27, 2000
21,582
10,785
136
*redacted*
Waiting for Covfefelake

Amusing memetics aside, your stance seems to be the logical one. That looks like the Blue Team's next best product.

I will say that it'll be interesting to see what kind of efficiency endusers can achieve by tuning for lowest possible volts rather than trying to hit highest possible clocks. It looks to me like Intel pushed themselves out of 14nm+'s ideal voltage curve range.
 

plopke

Senior member
Jan 26, 2010
238
74
101
Aah nice to see we have so many trades off to make < 1000$ bracket. These power numbers are a little bit meeeh , they will have to reduce that quiet a bit if they want to squeeze out 8 more cores on the same platform. Starting to wonder why they even bothered with offering 18 on the X299 platform if these 10 are on the limit. 2-4 more sure ,6-8 really ?
 

IndyColtsFan

Lifer
Sep 22, 2007
33,656
687
126
The reviews so far reinforce my plan to wait for TR and Coffee Lake to see what happens. I was really hoping for the 7900, but those temps really concern me - my office is warm as-is, and that thing would heat my whole second floor. Maybe once things have settled and we have some reports from delidded CPUs, my thinking will change. There is no way I am building a custom loop for one of these guys though, so if high-end air or AIO isn't sufficient, I'll have to look in another direction.

There was a rumor floating around about Intel releasing "soldered" HEDT chips later this year - my thinking is that they're going to have to solder the 12C and above CPUs to get better heat transfer, and that might be where this rumor started. Either way, it doesn't matter too much - I can wait to upgrade to see how it all pans out. :)
 

ManyThreads

Member
Mar 6, 2017
99
29
51
Can someone explain to me please why the vertical M2 drive on the Asus boards is an issue? I ordered a TUF MK1 but now I am worried...I assume a board like that has no trouble maxing out a 960 PRO SSD in any configuration - is that wrong?
 

DDH

Member
May 30, 2015
168
168
111
Aah nice to see we have so many trades off to make < 1000$ bracket. These power numbers are a little bit meeeh , they will have to reduce that quiet a bit if they want to squeeze out 8 more cores on the same platform. Starting to wonder why they even bothered with offering 18 on the X299 platform if these 10 are on the limit. 2-4 more sure ,6-8 really ?


I was thinking along the same lines. Taking the thermal and overclock results from TH's review, How on earth they plan on releasing 14, 16, and 18 core variants and maintain decent cooling and power figures if they dont resolve their thermal paste to IHS issue is beyond me.
 

lolfail9001

Golden Member
Sep 9, 2016
1,056
353
96
I will say that it'll be interesting to see what kind of efficiency endusers can achieve by tuning for lowest possible volts rather than trying to hit highest possible clocks. It looks to me like Intel pushed themselves out of 14nm+'s ideal voltage curve range.
I dunno, Kev, 1.1V for 4.3Ghz is pretty impressive and that looks to be entirely feasible. If anything, SKL-X looks to be severely thermally limited even on LN2 + delid rather than voltage curve limited.

And of course it does not help that stress tests turn it into a furnace cause of how vector units are laid out, limiting stability testing you can do.
 

Malogeek

Golden Member
Mar 5, 2017
1,390
778
136
yaktribe.org
Can someone explain to me please why the vertical M2 drive on the Asus boards is an issue? I ordered a TUF MK1 but now I am worried...I assume a board like that has no trouble maxing out a 960 PRO SSD in any configuration - is that wrong?
In theory, sounds like a great idea for allowing air to flow past the drive. M.2 drives definitely have heat problems.
 

TahoeDust

Senior member
Nov 29, 2011
557
404
136
Can someone explain to me please why the vertical M2 drive on the Asus boards is an issue? I ordered a TUF MK1 but now I am worried...I assume a board like that has no trouble maxing out a 960 PRO SSD in any configuration - is that wrong?

For me, it is 100% aesthetics. It just looks like an afterthought to me. I am usually a form follows function guy, but it just irks me.
 
  • Like
Reactions: ManyThreads

moonbogg

Lifer
Jan 8, 2011
10,635
3,095
136
Aah nice to see we have so many trades off to make < 1000$ bracket. These power numbers are a little bit meeeh , they will have to reduce that quiet a bit if they want to squeeze out 8 more cores on the same platform. Starting to wonder why they even bothered with offering 18 on the X299 platform if these 10 are on the limit. 2-4 more sure ,6-8 really ?

There was a cheesy looking rumor that Intel would release a revised socket to go along with the higher core count chips. It sounded almost unbelievable at first, but seeing these power figures, that rumor doesn't sound so far fetched at all now. Relying on the power headroom in these nice boards to pick up some unexpected load would have been fine, but the power required here is kind of ridiculous. I am doubtful that maintaining decent clocks on a 12+ core chip will be possible without a more robust specification for power delivery.
Also, adding solder to the higher core count chips will clearly help with temps, but I am doubting it will help all that much with power requirements. Yeah chips eat more power when hot, but the power being pulled by that 10 core leads me to think that no amount of solder or cooling will keep power requirements in check. I wouldn't be surprised in the slightest if Intel kicked their enthusiasts right in the face by releasing soldered chips and more robust boards right after they spent tons of money on one of these thermal pasted nightmares.
Being completely honest here, my suspicions couldn't possibly have any merit, right? Especially considering how severe the backlash would be. It would be completely hostile and crazy. It can't be true, but at the same time I see no way around it. I really think Intel is in a very tight spot here.
 
  • Like
Reactions: DarthKyrie

raghu78

Diamond Member
Aug 23, 2012
4,093
1,475
136
There was a cheesy looking rumor that Intel would release a revised socket to go along with the higher core count chips. It sounded almost unbelievable at first, but seeing these power figures, that rumor doesn't sound so far fetched at all now. Relying on the power headroom in these nice boards to pick up some unexpected load would have been fine, but the power required here is kind of ridiculous. I am doubtful that maintaining decent clocks on a 12+ core chip will be possible without a more robust specification for power delivery.
Also, adding solder to the higher core count chips will clearly help with temps, but I am doubting it will help all that much with power requirements. Yeah chips eat more power when hot, but the power being pulled by that 10 core leads me to think that no amount of solder or cooling will keep power requirements in check. I wouldn't be surprised in the slightest if Intel kicked their enthusiasts right in the face by releasing soldered chips and more robust boards right after they spent tons of money on one of these thermal pasted nightmares.
Being completely honest here, my suspicions couldn't possibly have any merit, right? Especially considering how severe the backlash would be. It would be completely hostile and crazy. It can't be true, but at the same time I see no way around it. I really think Intel is in a very tight spot here.

I think IHS and a socket 2066v2 with higher power specification could surely be on the cards. bitsandchips have been making some statements about rumours they keep hearing from their sources.

http://semiaccurate.com/forums/showpost.php?p=291375&postcount=80

http://www.bitsandchips.it/52-engli...e-x-skus-with-soldered-ihs-in-the-near-future

http://www.bitsandchips.it/52-engli...commercialize-a-v2-edition-of-the-socket-2066
 
Mar 10, 2006
11,715
2,012
126
I think IHS and a socket 2066v2 with higher power specification could surely be on the cards. bitsandchips have been making some statements about rumours they keep hearing from their sources.

http://semiaccurate.com/forums/showpost.php?p=291375&postcount=80

http://www.bitsandchips.it/52-engli...e-x-skus-with-soldered-ihs-in-the-near-future

http://www.bitsandchips.it/52-engli...commercialize-a-v2-edition-of-the-socket-2066

This is the same source that said that 14-18 core SKX wouldn't work in LGA 2066 when Intel clearly said it did at the launch.

I don't think bits & chips is a reliable source, TBH.
 

raghu78

Diamond Member
Aug 23, 2012
4,093
1,475
136
This is the same source that said that 14-18 core SKX wouldn't work in LGA 2066 when Intel clearly said it did at the launch.

I don't think bits & chips is a reliable source, TBH.

bitsandchips have been spot on with Zen related information. Could you quote what they exactly said about Skylake-X 14-18 core SKUs and 2066.
 
  • Like
Reactions: DarthKyrie

DrMrLordX

Lifer
Apr 27, 2000
21,582
10,785
136
I dunno, Kev, 1.1V for 4.3Ghz is pretty impressive and that looks to be entirely feasible. If anything, SKL-X looks to be severely thermally limited even on LN2 + delid rather than voltage curve limited.

And of course it does not help that stress tests turn it into a furnace cause of how vector units are laid out, limiting stability testing you can do.

Volts mean nothing if we don't also examine current draw. Take a look at the 9590 compared to the 8370e for example. Different process and uarch, but same basic principle. I do expect some users to explore undervolting in the 3.5-4.0 GHz range. We'll see what can be done there.

If Intel does need a new socket for higher-core parts (and really, how is the 12c part going to fare on the current boards?) then . . . ouch.
 
  • Like
Reactions: moonbogg

moonbogg

Lifer
Jan 8, 2011
10,635
3,095
136
This is the same source that said that 14-18 core SKX wouldn't work in LGA 2066 when Intel clearly said it did at the launch.

I don't think bits & chips is a reliable source, TBH.

I'm curious about the following. If Intel doesn't revise things, such as power delivery and soldering, I'm wondering what effect that could have on the clocks for the higher core count chips. I wonder if a 16 or 18 core chip at 4+ghz could actually exceed the power capabilities of most boards out there. Are they robust enough already? I'm wondering what the effect of thermal paste being used in a 16 or 18 core chip could be regarding clocks and temps with so many cores. Lastly, I wonder how current or soon to be Skylake-X owners may feel if Intel released future chips with solder and much better thermal performance. These are probably some tough questions that I'd like Intel to answer, but its not like I can ask them, so I'm wondering what the thoughts are of the users here regarding my questions.
 
  • Like
Reactions: DarthKyrie

IEC

Elite Member
Super Moderator
Jun 10, 2004
14,323
4,904
136
I pre-ordered a second Skylake-X + Mobo combo. This time, 7820X + Gigabyte AORUS Gaming 7.

This is going to be a lot of fun!

Are you going to run BOINC on them? Curious to see what the practical maximum OC is for 100% 24/7 load.
 
  • Like
Reactions: Drazick

formulav8

Diamond Member
Sep 18, 2000
7,004
522
126
What a pathetic release Intel. I hope it was only panic mode and things get better. (I would Not void my warranty on a 1k cpu for higher clocks and lower temps by delidding.)

There is NO reason for Intel to NOT solder the lid except for getting their extra 25 cent.

Even the lowly, so-called cash strapped AMD spent the extra money for a high quality conductor.

What is weird is there are still persons that thinks AMD's TR didn't get noticed by Intel.

Intel can't even have a higher than 20 thread cpu hedt before AMD drops the 32 thread release.