Skylake-X (Core i9) - Lineup, Specifications and Reviews!

Good is a relative term I guess. The 1800x lagged behind but may have been pre Agesa 1.0.0.6
 
When will i9 10core processors start showing up in dell/Hp pre built? I am having problems with my Ryzen build and if the new memory does not fix it I am going to RMA the whole system back to n ew egg and take a loss on the Ebay processor.
 
This is pretty surprising. Look at the difference lowering mesh voltage by .033v (still at 3200MHz) and turning down LLC one notch made in temps. All other setting are the same. As you can see, the core voltage is unchanged, and comparing the average and peak watts, they are almost identical ~1% lower. But look at those temps! Max and average both dropped by 7c. I really did not think the mesh voltage affected temps this much.

Before lowering LLC and Mesh...
I2ec6aQ.jpg


After:
0rlUsBo.jpg
 
Tweaktown 7820x review with latest bios:
http://www.tweaktown.com/reviews/82...x-series-skylake-8c-16t-cpu-review/index.html

The focus of the bios was to reduce power and did so by 10 watts or so.
It sewmed to help gaming min fps also.

As I keep saying reviews that used launch BIOS didn't draw an accurate picture of the new hardware, because the BIOS affected key factors as performance, power consumption, and overclocking headroom.

Moreover, the reviewer mentions this is not still the last word:

I wasn't surprised to find that the new BIOS ushered in slightly lower power consumption, but rumors have it that a new beta microcode will reduce those numbers even further.
 
Last edited:
Some interesting graphs in there for the people that say x299 isn't good for gaming.

The social media would have you to believe otherwise lol

Indeed, they pretend me to believe that 7820X being ~8% behind the 7700k on games is a "disaster", but the 1800X lagging a ~30% behind the same 7700k is "good enough for gaming".
 
Indeed, they pretend me to believe that 7820X being ~8% behind the 7700k on games is a "disaster", but the 1800X lagging a ~30% behind the same 7700k is "good enough for gaming".
That lag is largely due to the change in cache behavior, which in time compiler optimizations will reduce, or even flip around (4 times as much L2 cache per core on Skylake-X is huge, waiting to be taken advantage of).

Also, as game engines shift to taking advantage of more cores, the distance between the 7820x and 7700k will grow, in the other direction. Don't think game engines will take advantage of more cores? Take a look at how many cores the latest gen game consoles have.
 
Indeed, they pretend me to believe that 7820X being ~8% behind the 7700k on games is a "disaster", but the 1800X lagging a ~30% behind the same 7700k is "good enough for gaming".

rekt
 
That lag is largely due to the change in cache behavior, which in time compiler optimizations will reduce, or even flip around (4 times as much L2 cache per core on Skylake-X is huge, waiting to be taken advantage of).

Also, as game engines shift to taking advantage of more cores, the distance between the 7820x and 7700k will grow, in the other direction. Don't think game engines will take advantage of more cores? Take a look at how many cores the latest gen game consoles have.

The gap with the 7700k can be also reduced overclocking the mesh on the 7820X, which reduces latency and improves performance on games.

As I stated in some other part the current consoles have 8-cores because game developers rejected anything above 8-cores, due to difficulties with programing. Therefore I expect CPUs with 6 or 8 cores clocked high to be enough for gaming for another round.
 
The gap with the 7700k can be also reduced overclocking the mesh on the 7820X, which reduces latency and improves performance on games.

As I stated in some other part the current consoles have 8-cores because game developers rejected anything above 8-cores, due to difficulties with programing. Therefore I expect CPUs with 6 or 8 cores clocked high to be enough for gaming for another round.

A blanket statement that is only true for a few games. It really doesn't do all that much in most games.
 
Jumping in on Skylake-X

Ordered a 7820X and a Gigabyte Aorus Gaming 9 motherboard.

I was going to go Threadripper 1950X but the gaming performance was a little disappointing to me, and that's what I will primarily be using this build for. The whole gaming mode vs creation mode switching on Threadripper is a huge annoyance factor to me.

I have 2 Samsung 960 Pro 1TB NVMe drives ready to go along with a Thermaltake P3 Case (Red and Tempered Glass Upgrade) along with a Thermaltake Riing 360mm AIO for the CPU. Also have G.SKILL 32GB (4 x 8GB) Ripjaws V (DDR4-3200 @ 14-14-14-34) memory for the build (F4-3200C14Q-32GVK). Will be using the MSI Gaming X 1080 Ti, Samsung 950 Pro 512GB, and EVGA SuperNova T2, 80+ Titanium 1000W PSU from my current build.
 
Jumping in on Skylake-X

Ordered a 7820X and a Gigabyte Aorus Gaming 9 motherboard.

I was going to go Threadripper 1950X but the gaming performance was a little disappointing to me, and that's what I will primarily be using this build for. The whole gaming mode vs creation mode switching on Threadripper is a huge annoyance factor to me.

I have 2 Samsung 960 Pro 1TB NVMe drives ready to go along with a Thermaltake P3 Case (Red and Tempered Glass Upgrade) along with a Thermaltake Riing 360mm AIO for the CPU. Also have G.SKILL 32GB (4 x 8GB) Ripjaws V (DDR4-3200 @ 14-14-14-34) memory for the build (F4-3200C14Q-32GVK). Will be using the MSI Gaming X 1080 Ti, Samsung 950 Pro 512GB, and EVGA SuperNova T2, 80+ Titanium 1000W PSU from my current build.

Nice! Sounds similar to what I am running right now.
 
A blanket statement that is only true for a few games. It really doesn't do all that much in most games.

Yes, he keeps pushing the mesh issue even though it adds even more heat and does very little.

As for the 30% lower Ryzen performance, these guys are perfectly fine using early Ryzen bios to compare.
 
As for the 30% lower Ryzen performance, these guys are perfectly fine using early Ryzen bios to compare.

The claim it is "good enough for gaming" was done then, with those "early Ryzen bios". I will phrase the point in an alternative way: The same people that then labeled as irrelevant the ~30% gap between the 7770k and the 1800X is now pretending that the ~8% gap between the 7770k and the 7820X is something terrible for gaming.

Sure that BIOS updates have improved on RyZen side. BIOS updates have also improved things on SKL side.
 
The claim it is "good enough for gaming" was done then, with those "early Ryzen bios". I will phrase the point in an alternative way: The same people that diminished the ~30% gap between the 7770k and the 1800X then is now pretending that the ~8% gap between the 7770k and the 7820X is something terrible for gaming.

Sure that BIOS updates have improved on RyZen side. BIOS updates have also improved things on SKL side.

After CFL it will be "more performance than I need so didn't buy it", wanted another platform and how slower is smoother excuses. ;)
 
After CFL it will be "more performance than I need so didn't buy it", wanted another platform and how slower is smoother excuses. ;)

At least we can buy a AMD chip. I think you will find it hard to plug that Coffelake processor in when it's made out of paper.
 
Even Dell got CFL ready.
http://ranker.sisoftware.net/show_r...d5e2d2e2d6e2d6f082bf8fa9cca994a482f1ccf4&l=en

Dont worry, you can get Ryzen for half price after its launch if you can find one that isn't bugged with crashes and corrupt data. ;)

If your so sure on that why dont we make a bet. I will run the linux setup that is out there that people are having trouble with. If my 1700 system has the error I will stay out of the Intel forum for a year, if it runs without issue then you stay out of the AMD forum for a year. Let me know if your [H]ard enough for the bet.
 
Last edited:
If your so sure on that why dont we make a bet. I will run the linux setup that is out there that people are having trouble with. If my 1700 system has the error I will stay out of the Intel forum for a year, if it runs without issue then you stay out of the AMD forum for a year. Let me know if your [H]ard enough for the bet.

AMD has accepted there is a problem and taking in RMAs while still trying to figure out how to solve it besides disabling SMT+Opcache+ASLR. So you can only want to prove something that isn't right, not exactly trustworthy. I also remember your rage about lies when the bug was mentioned and how it was only B350 boards and other fictional ideas, yet where are we today?

Funny enough the same people that ignored the data corruption bug is mentioning ECC support.

Yes, must have memory protection for single bit errors while your CPU garbage your data. Not to mention how OC isn't a problem while you throw any form of data integrity out the door ;)
 
Last edited:
AMD has accepted there is a problem and taking in RMAs while still trying to figure out how to solve it besides disabling SMT+Opcache+ASLR. So you can only want to prove something that isn't right, not exactly trustworthy. I also remember your rage about lies when the bug was mentioned and how it was only B350 boards and other fictional ideas, yet where are we today?



Yes, must have memory protection for single bit errors while your CPU garbage your data. Not to mention how OC isn't a problem while you throw any form of data integrity out the door ;)

Yep what I figured you are unwilling to take the bet and yes I still think most of it is people causing the error, why they get a RMA chip and it does it again. You can talk all you want but the reality is plenty can run linux without issue only a tiny amount have a issue. I have not ran it since I could care less about Linux so I have no clue if mine will pass or not but I am far more confident then perhaps you are.
 
Yep what I figured you are unwilling to take the bet and yes I still think most of it is people causing the error, why they get a RMA chip and it does it again. You can talk all you want but the reality is plenty can run linux without issue only a tiny amount have a issue. I have not ran it since I could care less about Linux so I have no clue if mine will pass or not but I am far more confident then perhaps you are.

Fail to see your logic for the bet as the external party (AMD) to this pissing match, already confirmed the issue exists. If your CPU doesn't throw up errors, will all the affected chips get fixed or what? Don't get so attached to components, their just sand, lol.
 
Fail to see your logic for the bet as the external party (AMD) to this pissing match, already confirmed the issue exists. If your CPU doesn't throw up errors, will all the affected chips get fixed or what? Don't get so attached to components, their just sand, lol.

There is no logic in his post. It is unadulterated denialism, once AMD already confirmed the issue exists.

denialism1final.jpg
 
Last edited:
Early hours of overclocking. Getting some good results. 4.8GHz across all 8 cores at 1.27volts on average. I say on average because I'm using offset voltages (I like some power savings). Lowest core is 1.266, core 4 likes a little more voltage apparently at 1.285 volts. Have the MESH overclocked to 3.0GHz from 2.4GHz. Hitting 83C max.

It would seem that VRM's overheating on X299 is seriously blown out of proportion. Highest temp I am getting on my motherboard is 76C. Though I do think the Gigabyte Gaming 9 is the best motherboard in terms of VRM heatsinks.

So far I am impressed. Max overclock I could get on my 6700K was 4.6GHz. This thing has double the core count and is achieving a higher overclock. From 3.6GHz base clock to 4.8GHz is impressive imo, a 1.2GHz overclock on a 8 core CPU.
 
Anyone seen or heard anything about the new 12core 7920x? Wasn't it supposed to be released in august?
 
I know Proshop very well. It says they dont have it but might have it in a week. They don't have it. They just refer to their suppliers catalog of delivery dates. Which again always are in one to two weeks. I build such b2b2c systems, so I know how they work. :)

The big suppliers here, like Amazon.de don't have it at all. Caseking.de says 25th of september. That is the guess I trust most atm.

But we'll see.
 
I know Proshop very well. It says they dont have it but might have it in a week. They don't have it. They just refer to their suppliers catalog of delivery dates. Which again always are in one to two weeks. I build such b2b2c systems, so I know how they work. :)

The big suppliers here, like Amazon.de don't have it at all. Caseking.de says 25th of september. That is the guess I trust most atm.

But we'll see.

They have it 100% confirmed. Its just at a remote warehouse(Read distributor, quite commonly used here). At sept. 7 they expect to have it at their own warehouse as well.
 
Last edited:
Back
Top