• 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.

Intel Skylake / Kaby Lake

Page 657 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
osgorth said:
Did you read their Conclusion paragraph? They agree with you, as do I.

As for voltages on TR: I'm assuming TR has the same "issue" that Ryzen has sometimes when it just refuses to hit 4GHz unless you really abuse it. Others do 4.0 and even 4.1 with comparative ease. It seems both GamersNexus and Tom Logan over at oc3d.net have had major problems with that 4GHz wall lately. Silicon lottery at play, I'm sure.
Click to expand...
Well, I agree except on one point. Unless I missed it, all of our users in the TR builders threads have NOT said they need excessive voltage (over 1.3) to get 4.0. Its the over 4.0 that gets iffy and required vcore. I think I even read at many places the 4.0 is the "wall" not much vcore to get there, but a lot to get over it. 4.1 seems doable on water, but 4.2 or more is LN2 territory.
 
Markfw said:
Well, I agree except on one point. Unless I missed it, all of our users in the TR builders threads have NOT said they need excessive voltage (over 1.3) to get 4.0. Its the over 4.0 that gets iffy and required vcore. I think I even read at many places the 4.0 is the "wall" not much vcore to get there, but a lot to get over it. 4.1 seems doable on water, but 4.2 or more is LN2 territory.
Click to expand...

Even some single die Ryzen CPUs don't reach 4.0GHz. This is right around the area where the silicon lottery kicks in on the Zeppelin die.
 
PeterScott said:
Even some single die Ryzen CPUs don't reach 4.0GHz. This is right around the area where the silicon lottery kicks in on the Zeppelin die.
Click to expand...
You forget I have 2 of those, and they both get there. But only one stable at 4.0 ghz.

And threadripper is more binned. If you don't know what you are talking about, stay out of the thread.

If you guys would stop trying to confused the issue of the TR 4 ghz OC at less than 1.4 vcore, I would LOVE to get back to this thread !!!!
 
Looks like the 7980XE is king of HEDT when overclocked, but at a rather steep power cost (and 2x the price of the TR 1950X). It's pretty natural it'd get compared to the TR lineup given that these are the two shiniest platforms right now, even if the market segment itself is small compared to mainstream. Which itself should be interesting once Coffee Lake hits the shelves.

P.S. All my Ryzen chips can hit 4.0GHz, it's just the question of the voltage required to be stable. TR is supposedly binned as the top 2% or 5% of Zeppelin dies (depending on slide/source) so a TR 1950X requiring 1.44V to hit 4GHz (worse than my worst R7 1700s) means either 1) the reviewer got a dud or 2) the reviewer never bothered minimizing the voltage required for the OC, and simply used the same values as the R7 chips they OC'd. My 1950X only requires 1.25V to be stable at 4.0GHz per AIDA64 stress test. I give it a little more to offset the vdroop at load so I don't have to run aggressive LLC.
 
Markfw said:
You forget I have 2 of those, and they both get there. But only one stable at 4.0 ghz.

And threadripper is more binned. If you don't know what you are talking about, stay out of the thread.

If you guys would stop trying to confused the issue of the TR 4 ghz OC at less than 1.4 vcore, I would LOVE to get back to this thread !!!!
Click to expand...



This thread should be about the new Coffee Lake info from today, but some people like to talk about what they own and derail it.
 
Okay what the hell, can the moderators stay on topic on this thread. You guys have a TR thread going on. I get it you guys are like vegans and need to tell EVERYONE about your cpu. I get it but this is a skylake,kabylake and right now a coffeelake thread. Why cant you mods control your own actions. I mean what's the point of even having you as mods in the first place if you are going to act like the wccftech cancer discussion on their site.




You cannot call out mod by his title when he posts as a regular member, which is his right.


esquared
Anandtech Forum Director
 
Last edited by a moderator:
Burpo said:
I give up.. Maybe I can get some Intel info from the other sites I visit. Can't wait to get my hands on an 8700K. Finally gonna upgrade from X58.. Any boards available yet?
Click to expand...

At this point wccftech is more mature than this

You can check out videocardz. Whycry is the one that leaked all the coffeelake info and made Intel lift the embargo for coffeelake.
 
Burpo said:
I give up.. Maybe I can get some Intel info from the other sites I visit. Can't wait to get my hands on an 8700K. Finally gonna upgrade from X58.. Any boards available yet?
Click to expand...

Should be available Oct. 5 along with the CPUs. Lots of unboxing vids of the Z370 boards on YouTube.
 
tamz_msc said:
The 7800X wins in Sandra because of AVX512, most of those tests have got nothing to do with memory bandwidth.
Click to expand...

Ok, thanks, that didn't immediately occur to me.

Memory BW was the only the first thing that came to mind.

So outside of AVX512 Coffee Lake kicks ass?
 
Hayateazekura said:
Okay what the hell, can the moderators stay on topic on this thread. You guys have a TR thread going on. I get it you guys are like vegans and need to tell EVERYONE about your cpu. I get it but this is a skylake,kabylake and right now a coffeelake thread. Why cant you mods control your own actions. I mean what's the point of even having you as mods in the first place if you are going to act like the wccftech cancer discussion on their site.
Click to expand...
Well, maybe if you guys could stop derailing your own thread with misinformation, we could get back to the subject of the thread.
 
osgorth said:
Did you read their Conclusion paragraph? They agree with you, as do I.

As for voltages on TR: I'm assuming TR has the same "issue" that Ryzen has sometimes when it just refuses to hit 4GHz unless you really abuse it. Others do 4.0 and even 4.1 with comparative ease. It seems both GamersNexus and Tom Logan over at oc3d.net have had major problems with that 4GHz wall lately. Silicon lottery at play, I'm sure.
Click to expand...
This.

Here's a quote from the Threadripper Builder's thread.
Fir said:
The board suits my purposes fine.
I only have a corsair 115i AIO on hand to use and I believe that's holding the higher oc back.
4GHz is the end of the line for these chips. It may go higher but stability is questionable and it may seem stable but with 16 cores/32 threads it can be challenging to keep the processor busy enough to stress it to show instability. Some are pushing 1.45+ volts. YIKES! If that's your thing and you have a TR4 block designed to cool these chips and can handle TDP values over half a kilowatt then the Zenith is THE board to use as its VRM can handle the power.

It's rare for a CPU to go. I had an Asus Rampage X99 that killed a 5960X due to the BIOS bug that fed the chip over 1.8V on POST! That board still works but has all sorts of issues.

Both of my Zenith boards have cold boot problems when the PSU is turned off they go through a few ON OFF cycles before posting.

Aside from that they are fine.
Click to expand...
On the issue of stability, besides the memory related crashes some users in that thread experienced from trying to run memory at 3600Mhz CAS 16, most of the 4Ghz overclocks in that thread won't pass a Y-Cruncher stress test because they're running borderline overclocks and yet are unwilling to go over a certain voltage limit (1.4v or so). The reviews in question have no such problems pushing the voltages. They're not worried about long term overclocks, just what these threadripper chips would need to run at that satisfying, albeit borderline 4Ghz mark for their testing purposes, I suppose. See quoted post above. True stability for threadripper at 4Ghz comes at a steep power price.
Intel's AVX-512 gobbling chips get a bad rap for power consumption when none of that is the case in real, everyday use.

Edit: Sorry guys, this is going to be my last post about threadripper in this thread.
 
Last edited:
According to the article of PC Watch (Japanese), a supply of CFL-S at launch date is not sufficient, especially outside US. OEM suggests that it won't be available until late November in Japan.
 
Bouowmx said:
Intel Coffee Lake in ARK database: https://ark.intel.com/compare/126684,126686,126685,126687,126689,126688

Core i5-8400 and i3 have only 23 of 24 execution units enabled. I know Intel has done disabling 1 EU before, but I found this a bit amusing.
Click to expand...


Finally a small Mhz bump for Core i7 at least.

PeterScott said:
Does anyone remember if this leaked review came up:
https://videocardz.com/72915/first-review-of-intel-core-i7-8700k-leaks-out

I like it because they do a clock for clock comparison with the 7700K 8700K and 7800X.

The 8700K beats them both in most cases, but a few go to the 7800X likely from it's greater memory BW.
Click to expand...


I have seen this. Even this review wasn't fully working because Single Turbo didn't seem to work for them, only 191 points in Cinebench R15. Looks like board OEMs are struggling with the Bios. I think Asus does better. But it's nice to see that the core scaling over 7700k is pretty good at 4.5 Ghz with DDR4-2400 only.

In the pconline review all 8600k results are clearly nonsense, way too slow. With 6 physical cores you would expect much better MT results. In Cinebench R15 there is a 75% gap to i7-8700k, no chance that SMT makes such a big difference.
 
Last edited:
naonao said:
According to the article of PC Watch (Japanese), a supply of CFL-S at launch date is not sufficient, especially outside US. OEM suggests that it won't be available until late November in Japan.
Click to expand...

No idea if that is true or not, but part of me has thought that supply of the new chips isn't going to be great (at least at first) and that is a big part of the new motherboard requirement. Intel gets their good list prices (and they look pretty good to me!) but they still get more money because you're forced to buy a high end board (are there any plans to release lower end boards? I haven't heard any) This lets you brag about the cpu cost but the platform cost makes up the difference to some extent and manages supply in a way.

I can't find much on the boards yet, here's a leak of pre-order pricing for asus where the cheapest is $135.
https://www.reddit.com/r/intel/comments/71bncn/asus_z370_motherboards_preorder_pricing_us/
 
mikk said:
In the pconline review all 8600k results are clearly nonsense, way too slow. With 6 physical cores you would expect much better MT results. In Cinebench R15 there is a 75% gap to i7-8700k, no chance that SMT makes such a big difference.
Click to expand...

Multicore enhancement might have been enabled on the 8700K on that test.
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top