Intel Cannonlake, Ice Lake, Tiger Lake & Sapphire rapid thread

Page 225 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

mikk

Platinum Member
May 15, 2012
2,666
559
136
This could explain the high base clock for the final SKUs. I wonder how it's working when the CPU is set to TDP down 15W, will the base clock change in this case? Intel must be confident they can go really high on clock speeds, otherwise a default 28W TDP doesn't make a lot of sense to me. And of course, this is not confirmed.
 

uzzi38

Senior member
Oct 16, 2019
742
842
96
This could explain the high base clock for the final SKUs. I wonder how it's working when the CPU is set to TDP down 15W, will the base clock change in this case? Intel must be confident they can go really high on clock speeds, otherwise a default 28W TDP doesn't make a lot of sense to me. And of course, this is not confirmed.
?

If a laptop is TDP-downed to 15W then of course the standard base clock won't apply. Sustained clocks in a workload that can actually fully utilise the CPU core (without AVX512 instructions) will very likely be lower than 2.8GHz.

Weird alternative theory (I'm throwing things out there just for the sake of it if I'm honest), couldn't it also be because they're trying to ensure the iGPU has ample power budget? Take Renoir for example, to ensure max clock rates on the Vega 8 iGPU Renoir requires approximately 15W (the full SoC will use closer to 20W, but some of that still needs to be for I/O and the small portion of the CPU being utilised). Or alternatively like Coffee Lake-U. Much larger iGPU, and 28W to ensure that iGPU had the power budget to stretch its legs in a standard configuration.
 
  • Like
Reactions: spursindonesia

mikk

Platinum Member
May 15, 2012
2,666
559
136
By the way their Xe presentation from GDC isn't online yet despite Intels claim it would come out in waves through March and April. This is the result, we get half baked and half wrong leaks which I think is not good for Intel because obviously some of the stuff is wrong or contains outdated infos. I bet we will get the full specs from certain leaks before Intels revealing. When Intel reveals Tigerlake/Xe, they can't surprise us anymore because we knew everything from leaks.
 
  • Like
Reactions: Dayman1225

mikk

Platinum Member
May 15, 2012
2,666
559
136
You won't get full architectural coverage from leaks.
But a lot of other news and bits! We have seen the first examples recently. I would prefer to hear from Intel exclusively about the new features like AV1 decoding or a first performance snapshoot etc. but looks like Intel is waiting for the leaks.
 

IntelUser2000

Elite Member
Oct 14, 2003
6,673
1,235
126
But a lot of other news and bits! We have seen the first examples recently. I would prefer to hear from Intel exclusively about the new features like AV1 decoding or a first performance snapshoot etc. but looks like Intel is waiting for the leaks.
These leaks could mean we're not that far off from official unveiling. Perhaps we'll see it in June even without Computex?
 
  • Like
Reactions: french toast

mikk

Platinum Member
May 15, 2012
2,666
559
136
Scheduling seems to be similar to Icelake last year +- 1 month. In the middle of August there is a presentation of Tigerlake mobile during the Hot Chips event, also there is a GDC digital event scheduled for August 4-6.
 

jpiniero

Diamond Member
Oct 1, 2010
7,681
1,051
126
These leaks could mean we're not that far off from official unveiling. Perhaps we'll see it in June even without Computex?
I would think the launch would have to be before Apple's WWDC (which starts on June 22nd), if only for Shareholder PR purposes over the Apple to ARM talk.
 

IntelUser2000

Elite Member
Oct 14, 2003
6,673
1,235
126
Actually Bob Swan himself said it'll debut middle of this year.

No matter. The increase in base clocks is still really noteworthy and still pretty darn good being 500mhz higher than the base clock of the 1068G7. Couple on some smaller IPC gains - even just 5% or so and TGL-U should still be a pretty nice uplift gen-on-gen.
700MHz if you count the 3GHz chip.

But the 28W ICL parts are useless.


The Surface Laptop 3 with 1065G7 set at 25W outperforms it in graphics. The CPU is maybe 5% faster in sustained.

All that and released 7 months later. It's almost like Apple is purposely making it crap since they'll abandon it for ARM next year. What's the point of exclusivity then?
 
Last edited:

IntelUser2000

Elite Member
Oct 14, 2003
6,673
1,235
126

Zucker2k

Golden Member
Feb 15, 2006
1,053
454
136
Rocketlake PL2 by Sharkbay: https://videocardz.com/newz/intel-alder-lake-s-rumored-to-support-ddr5-memory

8C/6C
-PL1: 95W
-PL2: 251/173W
-Tau 56s

8C/6C/4C
-PL1: 80W
-PL2: 251W/191W/146W
-Tau 28s

8C/6C/4C
-PL1: 65W
-PL2: 251W/177W/128W
-Tau 28s

This explains why Rocketlake only has 8 cores. Potentially 25% faster per clock, but with 25% core count difference.
I suspected this. Yeah, this backport is going to need all the juice it can get. Unfortunately, this is the best 14nm can offer at the moment.
 

uzzi38

Senior member
Oct 16, 2019
742
842
96
Rocketlake PL2 by Sharkbay: https://videocardz.com/newz/intel-alder-lake-s-rumored-to-support-ddr5-memory

8C/6C
-PL1: 95W
-PL2: 251/173W
-Tau 56s

8C/6C/4C
-PL1: 80W
-PL2: 251W/191W/146W
-Tau 28s

8C/6C/4C
-PL1: 65W
-PL2: 251W/177W/128W
-Tau 28s

This explains why Rocketlake only has 8 cores. Potentially 25% faster per clock, but with 25% core count difference.
Just one thing. With SNC's cache layout I don't think we'll see it at 25% higher perf per core.

And this ties in a little with something Sharkbay has said in the past - that Rocket Lake won't have an i9 version. Only up to an i7.

Makes sense because there probably will be a small deficit in multithreaded performance, so marketing another i9 SKU would be difficult.
 

Exist50

Member
Aug 18, 2016
76
69
61
Just one thing. With SNC's cache layout I don't think we'll see it at 25% higher perf per core.

And this ties in a little with something Sharkbay has said in the past - that Rocket Lake won't have an i9 version. Only up to an i7.

Makes sense because there probably will be a small deficit in multithreaded performance, so marketing another i9 SKU would be difficult.
Multicore deficit or not, I doubt marketing will them get away with a "regression" to the i7 branding. They've already oriented their marketing around low core count loads, since AMD demolishes them in multicore, so it shouldn't be that hard to double down.

Also, where is this 25% number coming from? Even for Willow Cove, that number doesn't really seem supported anywhere.
 
  • Like
Reactions: spursindonesia

Zucker2k

Golden Member
Feb 15, 2006
1,053
454
136
Also, where is this 25% number coming from? Even for Willow Cove, that number doesn't really seem supported anywhere.
Sunnycove is already 18% ahead of Coffee Lake, clock for clock. Some believe, based on some GB scores, that Willowcove is another 5% or so higher than Sunnycove. If this pans out, Rocket Lake still needs to bring all Willowcove's performance, plus it must clock to 5GHz for that 25% to hold.
 

IntelUser2000

Elite Member
Oct 14, 2003
6,673
1,235
126
If this pans out, Rocket Lake still needs to bring all Willowcove's performance, plus it must clock to 5GHz for that 25% to hold.
It probably won't change things either way. Cometlake is already a lost cause in very parallel applications.

But in other applications it may turn out even better. Lots of multithreaded applications where core scaling is poor. Also the new uarch will bring gains in the areas Skylake is weak in.

Games are going to love the new uarch as they always do. If you see GN tests, they scale poorly with clocks, and worse with cores. Just 20% gain there will wreck most chips even at 4.5GHz.
 
  • Like
Reactions: pcp7

uzzi38

Senior member
Oct 16, 2019
742
842
96
It probably won't change things either way. Cometlake is already a lost cause in very parallel applications.

But in other applications it may turn out even better. Lots of multithreaded applications where core scaling is poor. Also the new uarch will bring gains in the areas Skylake is weak in.

Games are going to love the new uarch as they always do. If you see GN tests, they scale poorly with clocks, and worse with cores. Just 20% gain there will wreck most chips even at 4.5GHz.
IPC vs core clocks doesn't change much for games. They both have the same effect more or less. Also, a 20% IPC gain at 4.5GHz is still roughly equal in performance to 5.4GHz on the original uArch, so I'm not sure what your point is there? That's higher ST performance than Comet Lake.

Rocket Lake would need to clock higher than that and still provide that IPC upgrade of a notable improvement in games over Comet Lake.

Memory latency is something that drastically improves performance on both AMD and Intel systems right now. If you want to improve gaming performance significantly, lower latency access to memory is the best way to do it.
 
  • Like
Reactions: spursindonesia

IntelUser2000

Elite Member
Oct 14, 2003
6,673
1,235
126
@uzzi38 That's not the case. Perf/clock(the more accurate term) not only improves pretty much as is, but also addresses weaknesses of the previous architecture. You can clearly see that from the Zen 2 review.

Clock scaling is far poorer in games than applications. You can typically get 80% in SpecCPU or Cinebench. You're lucky to get 50-60% in games, and that's only when you are running on settings and tests that show the difference.


Overclocking it from 3.6GHz boost to 4.2GHz gets you 5-7% gains resulting in 30-40% scaling. This is typical. I don't know why anyone believes otherwise. Games are somewhat GPU bound even in the 100 fps+ range. You'll never get the scaling like applications because majority of compute is run by the GPU. Before Turbo wasn't as complex, I used to sit down for hours and analyze data.

Now this assumes a 14nm port of Willow/Sunny Cove cores are going to be identical to the 10nm version. It might not. But if it does its a much needed change.

Skylake is a core that's going to be in kindergarten soon. It's going to suck.
 

ASK THE COMMUNITY