Discussion Intel current and future Lakes & Rapids thread

[DrMrLordX]

Not at 28 W. H35 is 3 Ghz at 28 W, albeit I believe that's using AVX-512.

I was merely listing TDP values and official boost figures. All AVX/AVX2/AVX-512 workloads will lower the all-core boost. As we also know, Intel CPUs frequently exceed listed TDP values. The point of the mental exercise was to evaluate (based on Tiger Lake-U on 10SF) whether or not 8c Golden Cove/Alder Lake-S could reach clocks of ~4.7 GHz (in non-AVX workloads) in a TDP of 105W/actual power envelope of 142W. I'm pretty sure it could. The added Gracemont cores will of course add a few watts more.

 

[jpiniero]

The added Gracemont cores will of course add a few watts more.

Depends on how high they can clock/draw. There was even some rumors that Intel was looking at even going to 140-145 W for the base TDP on K, presumably to keep the base clock high while having the small cores active. Either way I think they are sticking with 250 W for the default PL2.

 

[repoman27]

Yes, that's Cooper Lake and 14nm.

But in full disclosure the 1068NG7 was just released with a turbo of 4.1GHz.
But still, as I wrote Ice Lake is Ice Lake. Good for Intel if they tweaked a new stepping and got another 200MHz from the 1065G7.

1068NG7 has been available since May 2020 in the 13-inch MacBook Pro. The N ("nano" package) chips were primarily (exclusively?) for Apple. Regardless, 10SF seems like it can clock a lot higher than 10+.

Ice Lake / 10+:

Core i7-1068NG7: 4C, 28 W TDP, 2.3 base, 4.1 max turbo

Xeon Silver 4309Y: 8C, 105 W TDP, 2.8 base, 3.6 max turbo
Xeon Gold 6334: 8C, 165 W TDP, 3.6 base, 3.7 max turbo
Xeon Platinum 8368Q: 38C, 270 W TDP, 2.6 base, 3.7 max turbo, liquid cooled
Xeon Platinum 8380: 40C, 270 W TDP, 2.3 base, 3.4 max turbo

Tiger Lake / 10SF:

Core i7-11375H: 4C, 28 W TDP-down, 3.0 base, 5.0 max turbo
Core i7-11375H: 4C, 35 W TDP-up, 3.3 base, 5.0 max turbo

 

[LightningZ71]

Yes, you are absolutely right.
RDNA2 IGP discussion is interesting I don't disagree, but this is simply not the right place for It. Rembrandt has Its own thread, discuss It there!

I'm sorry for getting sidetracked here. I participate in both threads and forgot which one I was on.

My main point, though, is that either Intel or AMD could produce a CPU with enough of an iGPU to easily satisfy most of the 1080p gaming crowd in the next generation of products without the product being massive. Looking at Tiger Lake U, as it exists right now, is a decent place to start. Just doubling the 96 EUs to 192 and using some decently specced DDR5, it should give reasonable performance there. My son's Lenovo Flex 5 with the G7 i5 with only 80eu does a decent job on everything that he plays. That's in the tight confines and low TDP of a 2 in 1. The resulting chip wouldn't be massive, it would barely be 50% larger than it currently is and would easily pull an extra $50+ ASP. Granted, we don't know how poor that process yields (10sf), so it could wind up cost prohibitive. However, it is going to be substantially less cost prohibitive that Kady Lake G was and likely get very close to it in gaming performance. Heck, they might even choose to provide drivers for it for longer than 24 months!

 

[TESKATLIPOKA]

I'm sorry for getting sidetracked here. I participate in both threads and forgot which one I was on.

I am not angry at you or anything like that, It was just a reminder to discuss It in the right thread.

 

[RTX]

Xeon Silver 4309Y: 8C, 105 W TDP, 2.8 base, 3.6 max turbo
Xeon Gold 6334: 8C, 165 W TDP, 3.6 base, 3.7 max turbo
Xeon Platinum 8368Q: 38C, 270 W TDP, 2.6 base, 3.7 max turbo, liquid cooled
Xeon Platinum 8380: 40C, 270 W TDP, 2.3 base, 3.4 max turbo

It appears the extra 3 UPI Links add another 45 watts assuming they used the same voltage for all 3 processors. 14nm+++ has 16% increased performance at equal/low voltage vs 14nm+
8180M 2.5 Ghz 205W 3UPI Links 14nm+
8280M 2.7 Ghz 205W 3UPI Links 14nm++
8380H 2.9 Ghz 250W 6UPI Links 14nm+++

1068NG7 has been available since May 2020 in the 13-inch MacBook Pro. The N ("nano" package) chips were primarily (exclusively?) for Apple. Regardless, 10SF seems like it can clock a lot higher than 10+.

Ice Lake / 10+:

Core i7-1068NG7: 4C, 28 W TDP, 2.3 base, 4.1 max turbo

Tiger Lake / 10SF:

Core i7-11375H: 4C, 28 W TDP-down, 3.0 base, 5.0 max turbo
Core i7-11375H: 4C, 35 W TDP-up, 3.3 base, 5.0 max turbo

i3-8121U 2C 15W 2.2 Ghz with GPU fused off 10nm
i7-1068NG7 4C 28W 2.3 Ghz 64 EU 1100 Mhz 10nm+
i7-1185G7 4c 28W 3.0 Ghz 96 EU 1350 Mhz 10SF

Intel confirms Apple MacBook Pro exclusivity for its 28 W Ice Lake processors; kills off the Core i7-1068G7 in the process

The faint glimmer of Core i7-1068G7 laptops arriving this quarter is seemingly no more. Removed from Intel's ARK database, the Core i7-1068G7 has been replaced by the Core i7-1068NG7, the 28 W processor found in the new MacBook Pro 13. The Core i5-1038NG7 has been added too, Intel's other 28 W...

www.notebookcheck.net

It's the same chip under a different name.

To add to what everyone else said, top TGL-H has no issue boosting to 5Ghz.

If the NUCs are anything to go by, the only thing holding laptop SKUs are the machine's VRM limitations and the cooling system.

 

[IntelUser2000]

i3-8121U 2C 15W 2.2 Ghz with GPU fused off 10nm
i7-1068NG7 4C 28W 2.3 Ghz 64 EU 1100 Mhz 10nm+
i7-1185G7 4c 28W 3.0 Ghz 96 EU 1350 Mhz 10SFView attachment 42965

Can't use base frequencies because according to that Icelake is at half the clocks of Tigerlake and we know that's not true.

The resulting chip wouldn't be massive, it would barely be 50% larger than it currently is and would easily pull an extra $50+ ASP.

I suggested why not a doubled up Tigerlake with 8 cores and 192EUs? But they chose one with a 32EU.

Now if you take Tigerlake-U and double up the iGPU, you'd end up at 190mm2. But that's not what they are doing. No signs of higher EU counts even with Alderlake!

And clearly if you look at profitability between the two companies AMD switched their strategies because a faster CPU has much better earnings potential, unless you are leading on CPU and you can use the GPU as a selling point like Intel did with Iris parts and only for few. Manufacturers charged $150-200 on TOP of the normal i7 variants, and not all models offered it.

If you think the two companies will put a radically faster iGPU and not charge you a lot for it I think you need to adjust expectations a bit.

 

[mikk]

AMD even downgraded Vega from 10/11 CUs in Picasso to 8 CUs in Renoir, because they could reach much higher GPU clock speeds at the same time. Now Intel reached GPU clock speeds that are comparable to Picasso which is rather low clocked compared to Renoir/Lucienne/Cezanne.

 

[repoman27]

It's the same chip under a different name.

There's only one Ice Lake 4+2 LP die (stepping D1) and one Ice Lake PCH-LP die (stepping D0) from which Intel currently makes 4 platforms (ICL-U/UN/Y/YN) and at least 15 SKUs. The N models use smaller "nano" packages, and it's amazing to me that to date there does not appear to be a single photo of those packages anywhere on the internet. For some reason, iFixit didn't bother to publish pictures as part of their usual teardowns. I almost disassembled some of my customers' machines to snap a few pics but couldn't justify pulling off the heat sinks for the sake of my own curiosity. Anyway, mostly similar specs, but not the same product:

ICL-Y, BGA1377, 26.5 mm x 18.5 mm
ICL-YN, BGA1044, 22 mm x 16.5 mm
ICL-U, BGA1526, 50 mm x 25 mm
ICL-UN, BGA1344, ?? mm x ?? mm

 

[LightningZ71]

Can't use base frequencies because according to that Icelake is at half the clocks of Tigerlake and we know that's not true.



I suggested why not a doubled up Tigerlake with 8 cores and 192EUs? But they chose one with a 32EU.

Now if you take Tigerlake-U and double up the iGPU, you'd end up at 190mm2. But that's not what they are doing. No signs of higher EU counts even with Alderlake!

And clearly if you look at profitability between the two companies AMD switched their strategies because a faster CPU has much better earnings potential, unless you are leading on CPU and you can use the GPU as a selling point like Intel did with Iris parts and only for few. Manufacturers charged $150-200 on TOP of the normal i7 variants, and not all models offered it.

If you think the two companies will put a radically faster iGPU and not charge you a lot for it I think you need to adjust expectations a bit.

Oh, I certainly expect them to charge dearly for it. They would be able to obviate the need for the MX450 completely and would certainly expect to get that upcharge for themselves. For the device builders, though, they would have simplified designs to manage, and wouldn't need to pass on as much of the cost. On the end, the existence of the mobile 1650 pretty much sets the ceiling on how much they can charge in any sort of notebook. It's tablets and 2 in 1s where it would be the biggest benefit.

 

[DrMrLordX]

Depends on how high they can clock/draw. There was even some rumors that Intel was looking at even going to 140-145 W for the base TDP on K, presumably to keep the base clock high while having the small cores active. Either way I think they are sticking with 250 W for the default PL2.

145 TDP for the top-end Alder Lake-S? Have they learned nothing?

 

[Zucker2k]

145 TDP for the top-end Alder Lake-S? Have they learned nothing?

It's 142w for anything greater than 6 cores in Ryzenworld,

 

[coercitiv]

It's 142w for anything greater than 6 cores in Ryzenworld,

O yeah, let's talk more about Ryzen power consumption... in the Intel thread where we barely stopped talking about AMD iGPU performance.

 

[jpiniero]

145 TDP for the top-end Alder Lake-S? Have they learned nothing?

For the K parts, and Intel was only "investigating". Either way PL2 is gonna be 250 W for sure. I bet it's more about setting the base clock of the small cores higher.

 

[coercitiv]

For the K parts, and Intel was only "investigating". Either way PL2 is gonna be 250 W for sure. I bet it's more about setting the base clock of the small cores higher.

Normally I would have no issue with them increasing TDP on K parts only, as long as the rest stays on 95W and 65W respectively. The problem is this would still affect all motherboards aiming for K CPU compatibility. (wattage is going up, voltages keep going down... VRM needs to be even beefier).

Is there any talk of Alder Lake or any future consumer gen going back to FIVR or something similar?

 

[DrMrLordX]

It's 142w for anything greater than 6 cores in Ryzenworld,

Incorrect. None of the AM4 CPUs have a TDP over 105W. They have a PPT of 142W. There is a difference. If you look at the 125W TDP of Intel's current CPUs, you should immediately know what I mean.

Alder Lake-S resorting to a PL2 of 250W just like the 10900k and 11900k is a very bad sign.

For the K parts, and Intel was only "investigating". Either way PL2 is gonna be 250 W for sure. I bet it's more about setting the base clock of the small cores higher.

10nm hasn't even had a successful desktop launch yet, and Intel is already pushing their products well out of the optimal v/f curve.

 

[Hulk]

10nm hasn't even had a successful desktop launch yet, and Intel is already pushing their products well out of the optimal v/f curve.

That's a keen observation if it comes to fruition. But there currently are no 10SF desktop parts so I don't think we should get too far ahead of ourselves. Frequency and power look pretty good with Tiger Lake so until proven otherwise I'm giving Intel the benefit of the doubt that they will keep those parameters under control for Alder Lake.

Actually I'm more interested in the final "throughput" or efficiency/clock of the combined Golden Cove/Gracemont design because ultimately that is the hill they will live or die on.

 

[Magic Carpet]

Is there any talk of Alder Lake or any future consumer gen going back to FIVR or something similar?

AFAIK, FIVR was back with Ice Lake on the mobile. Sucks that desktop is so far behind :/

I’ve liked the simplicity of 1150 mobos in that regard.

 

[Mopetar]

O yeah, let's talk more about Ryzen power consumption... in the Intel thread where we barely stopped talking about AMD iGPU performance.

Not to get off topic again or anything, but what do you think the power consumption is going to be like for AMD's integrated graphics?

 

[jpiniero]

Alder Lake-S resorting to a PL2 of 250W just like the 10900k and 11900k is a very bad sign.

I think 250 W for PL2 is here to stay for K parts. There really isn't any downside since it will just drop frequencies if the cooler can't handle it.

 

[jpiniero]

https://twitter.com/x/status/1381595758844055555

Ice Lake Server's HCC die has an stepping of M-1 which seems kind of excessive.

 

[Hulk]

O yeah, let's talk more about Ryzen power consumption... in the Intel thread where we barely stopped talking about AMD iGPU performance.

Good point but I think that since AMD is arguably Intel's biggest competitor/threat any discussion of Intel technology inextricably and implicitly compares their tech to AMD's competing tech. But yes, in-depth discussion of AMD is not the point of this thread I agree.

 

[eek2121]

It's 142w for anything greater than 6 cores in Ryzenworld,

You can get 8 cores @ 65W. Ryzen 3700X, APUs, etc.

For the K parts, and Intel was only "investigating". Either way PL2 is gonna be 250 W for sure. I bet it's more about setting the base clock of the small cores higher.

Their obsession with 5+ GHz is killing them. That being said, this chip is either going to be a beast when it comes to performance or it isn’t going to pull 250+ watts. This is not another 14nm upgrade. We should see significant power savings unless golden cove is much bigger than current.

 

[DrMrLordX]

I think 250 W for PL2 is here to stay for K parts. There really isn't any downside since it will just drop frequencies if the cooler can't handle it.

There is when it comes to motherboards. We've already seen problems with Z490 boards not supporting the 10900k (for instance) due to the high PL2 values and long tau periods. An OEM that wants to supply a bare minimum spec board for AM4 never has to worry about sustained power draw of more than 142W through the socket. And that's for a 16-core monstrosity. Plus we haven't seen what happens when Intel tries to dissipate that much power on 10nm on a desktop-sized die (40c Ice Lake-SP is massive . . . and 380W).

this chip is either going to be a beast when it comes to performance or it isn’t going to pull 250+ watts. This is not another 14nm upgrade. We should see significant power savings unless golden cove is much bigger than current.

You would think so. But if that is the case, why is Intel shooting straight for 250W PL2? Is that what they have to do to beat AMD in some benchmarks? Ice Lake-SP is already pulling monstrous amounts of power while still losing benchmarks to competitor's server CPUs from 2019.

 

[Exist50]

https://twitter.com/x/status/1381595758844055555

Ice Lake Server's HCC die has an stepping of M-1 which seems kind of excessive.

While Ice Lake took an obscene number of steppings, Intel starts the lettering at different points for derivative products.