Discussion Intel current and future Lakes & Rapids thread

[jpiniero]

A theoretical Xeon-D was my other thought. Would remove the need for said third party 2D graphics. Or maybe in a future server PCH.

 

[IntelUser2000]

A theoretical Xeon-D was my other thought. Would remove the need for said third party 2D graphics. Or maybe in a future server PCH.

No, it has to be more than just about saving a few bucks, which is the primary purpose of integrating the simplest configuration.

The AST chip provides more functions than just 2D graphics. Also, frequent change means reduced reliability, and no components go through more radical changes than the GPU. The chip line is probably something that has proven to work with vast different hardware combinations and multiple OSes.

 

[Hixbot]

Will the 9th gen CPU Metldown/Spectre silicon fixes suffer the same performance penalty as the OS/Microcode patches of today?
Will speculative execution be enabled with vulnerabilities resolved, or will it be silicon-level disabled?

 

[LTC8K6]

Will the 9th gen CPU Metldown/Spectre silicon fixes suffer the same performance penalty as the OS/Microcode patches of today?
Will speculative execution be enabled with vulnerabilities resolved, or will it be silicon-level disabled?

If we knew all that, we could make a lot of money doing some speculative execution of our own...

Disabled? Where'd that come from?

 

[Hixbot]

.

Disabled? Where'd that come from?

I was under the impression that current Spectre patches disables speculative execution, but after some quick reading, I see that is not the case.

 

[LTC8K6]

I did not know there were desktop Spectre patches out.

My Intel systems are Meltdown patched, but I haven't seen a Spectre patch yet.

 

[Dayman1225]

I did not know there were desktop Spectre patches out.

My Intel systems are Meltdown patched, but I haven't seen a Spectre patch yet.

I believe you have to go download a Windows Microsoft Patch your self to patch against it - as it is not pushed to auto update.

 

[LTC8K6]

I believe you have to go download a Windows Microsoft Patch your self to patch against it - as it is not pushed to auto update.

Looks like it's only for Skylake and newer chips.
I don't see anything for Haswell or Ivy Bridge.

 

[Dayman1225]

Looks like it's only for Skylake and newer chips.
I don't see anything for Haswell or Ivy Bridge.

Probably will see them in the near future.

 

[Dayman1225]

"Those "Crimson Canyon" NUCs with Bluetooth 5 certification *appear* to be Cannon Lake Core i3s (i3-8121U)." - David Schor

Link to BT5 Certification

So if OEMs won't make products with it, Intel will

 

[jpiniero]

May be premature to say that it is 8121U, but it does appear real and actually available soon. It wasn't on the NUC roadmap that was leaked earlier so this is new. I wonder what GPU it has...

 

[jpiniero]

http://winfuture.de/news,102849.html

This article claims that the Crimson Canyon indeed will ship with an 2 GB GDDR5 Radeon 500 series GPU. Not clear which one though. Also basically confirms that it is a 2+0 CPU.

 

[CatMerc]

https://twitter.com/GRbUF8qUqYB4iPy/status/987410657334255616

This is next level.
Cannon Lake might have a 50% IPC uplift. I am beyond confused.

 

[.vodka]

wat.

I don't see how is it possible to pull off another Conroe over Skylake?

That would be amazing!

I doubt it. Core i3s on the 8th Generation feature Turbo mode.

Oh, that explains it. Still, it's too huge a disparity.

 

[IntelUser2000]

https://twitter.com/GRbUF8qUqYB4iPy/status/987410657334255616

This is next level.
Cannon Lake might have a 50% IPC uplift. I am beyond confused.

I doubt it. Core i3s on the 8th Generation feature Turbo mode.

 

[CatMerc]

I doubt it. Core i3s on the 8th Generation feature Turbo mode.

There's a 3DMark 11 score, and that tends to detect Turbo. It was locked to 2.2GHz.

Trust me, I considered everything. I simply don't see any other option other than straight up false reading.

 

[LTC8K6]

https://twitter.com/GRbUF8qUqYB4iPy/status/987410657334255616

This is next level.
Cannon Lake might have a 50% IPC uplift. I am beyond confused.

Even if the numbers are exaggerated by 50% ...

But who knows if it's anything more than smoke?

 

[IntelUser2000]

There's a 3DMark 11 score, and that tends to detect Turbo. It was locked to 2.2GHz.

Trust me, I considered everything. I simply don't see any other option other than straight up false reading.

Come on.

The Core i3 8130U has a 2.2GHz base clock and 3.4GHz Turbo. Why would a i3 8121U not have one? I know its its more fun to believe in that its wrong, or that it actually has 50% IPC, but what does a more logical line of reasoning tell you?

 

[CatMerc]

Come on.

The Core i3 8130U has a 2.2GHz base clock and 3.4GHz Turbo. Why would a i3 8121U not have one? I know its its more fun to believe in that its wrong, or that it actually has 50% IPC, but what does a more logical line of reasoning tell you?

Obviously the boost is more logical. I'm just not certain about that, because there are way too many signs that this thing is locked to 2.2GHz.

We've had a leak about this very same CPU that said it was locked to 2.2GHz long before performance benchmarks, and 3DMark tends to be good at detecting clocks.

 

[IntelUser2000]

I'm telling you now, its just not showing the Turbo clocks.

Here's the result for 8130U: https://browser.geekbench.com/v4/cpu/7890958

Most results are just slightly higher than 8130U, owing to the likely scenario the 8121U having maybe 3.2GHz clock compared to 3.4GHz for the 8130U. Some are drastically better, but Cannonlake likely has some enhancements in a few areas.

You really think it makes sense that Cannonlake, which is a Tick by former Intel standards, having significantly greater advancement than anything else before it? We should remember "Conroe" only happened because some bad decisions were made with Netburst. Prior Intel chips with 5-6 years between them brought only 15-20% gains.

 

[LTC8K6]

Would AVX-512 and SHA play any role in these GB scores? They were mentioned in the Twitter posts.

 

[LTC8K6]

How is this new if it was out in January?
https://browser.geekbench.com/v4/cpu/search?q=8121u

 

[CatMerc]

How is this new if it was out in January?
https://browser.geekbench.com/v4/cpu/search?q=8121u

Nobody really paid attention to it, thinking it's 100% boosting. There was however recently a 3dMark 11 score that showed the core not boosting. That's what set off today's debate.

 

[IntelUser2000]

Here's the comparison between 8130U and 8121U: https://browser.geekbench.com/v4/cpu/compare/7890958?baseline=6246439

Few results are much higher for 8121U, but one of them is AES. The scores are not too far from each other in general though.

 

[LTC8K6]

So the justification for not comparing to the 8130u is based on the assumption that the 8121u does hot have Turbo?