Discussion Intel current and future Lakes & Rapids thread

[uzzi38]

You seem to gloss over the fact that the thinnest, lightest, and most efficient laptops you can buy have Tiger Lake U chips in them, not Ryzen.

Define efficiency?

 

[Abwx]

You seem to gloss over the fact that the thinnest, lightest, and most efficient laptops you can buy have Tiger Lake U chips in them, not Ryzen.

That s simply not true, Ryzen has noticeable higher efficency.

At NBC, and for a comparable Cinebench score, the Ryzen laptop use 95W from start to finish while the TGL one start at 135W and then sustain 121W, methink that it does the whole test closer to 135W than 121W.

From the test with Prime 95 wich help measuring sustained power we can conclude that Ryzen use 77W while TGL require 100W or so.

https://www.notebookcheck.net/Razer-Blade-14-Ryzen-9-Laptop-Review-Blade-Pro-17-Performance-At-Half-The-Size.546754.0.html

https://www.notebookcheck.net/Asus-TUF-Gaming-F15-FX506HM-Laptop-Review-No-Extras-Just-Performance.546842.0.html

 

[jpiniero]

https://www.amazon.com/dp/B096KN5N9K?th=1

First DDR5 kit made available to buy. Seems strange to get it out there with no product to put it in, but hey.

 

[repoman27]

IIRC, Intel uses tape in for internal manufacturing and tape out for external manufacturing in the same way. I'm less than convinced by attempts to draw meaningful conclusions from their use of that term. More to the point, perhaps:

Intel: Sorry, But Our 7nm Chips Will Be Delayed to 2022, 2023

In an earnings call, Intel CEO Bob Swan said the company identified a 'defect mode' in the 7nm process that resulted in 'yield degradation' when manufacturing the chips.

www.pcmag.com

If not Meteor Lake, then what would that be? Makes far more sense than a stopgap product pushing things out by a whole year.



No first product on a node tapes out on a 1.0 PDK. But that distinction doesn't matter in this context.

Well, even Gelsinger tripped up and used tape-in and tape-out in the same sentence, and the corresponding tweet from Gregory Bryant is notable for using the progressive form of the verb, i.e. "taping in". I believe the consensus was that "taped-in" indicates that design work has been completed, however the final mask images have not yet been generated (traditional tape-out), nor do they have a complete set of photomasks (which is a prerequisite for wafer production). So the start of Meteor Lake risk production will likely be 1 to 2 quarters out from May 24, 2021, which would be sometime in the second half of this year.

Your other quote is from Bob Swan during the Q2 2020 earnings call on July 23, 2020, nearly a year ago:

We now expect to see initial production shipments of our first Intel-based seven-nanometer product, a client CPU, in late '22 or early '23.
We are also focused on maintaining an annual cadence of significant product improvements independent of our process road map, including the holiday refresh window of 2022. In addition, we expect to see initial production shipments of our first Intel-based seven-nanometer data center CPU design in the first half of '23.

Ignoring the fact that Intel has updated both their CEO and roadmap since then, you still need to parse that statement correctly. The term "initial production shipments" when addressing shareholders means the same thing as "shipping for revenue", which hinges entirely on the Product Release Qualification (PRQ) engineering milestone. What he's actually saying is, "We hope to achieve Meteor Lake lead-die PRQ during Q1 2023." (I don't care how much of an eternal optimist you are, if you expect Intel to ever hit the earlier side of a stated production window, then you're only setting yourself up for disappointment.) RTS will trail that by 1-2 quarters as Intel undertakes the volume ramp and channel fill, which puts you right at an August 2023 launch date. Consider Intel's press embargo dates over the past eight years:

Jun 02, 2013: Haswell U/H/S
Sep 01, 2013: Haswell Y/U
Sep 05, 2014: Broadwell-Y
Aug 05, 2015: Skylake-S (Core 17-6700K and Core i5-6600K only)
Sep 01, 2015: Skylake Y/U/H/S
Aug 30, 2016: Kaby Lake-Y/U
Aug 21, 2017: Kaby Lake Refresh-U
Aug 28, 2018: Amber Lake-Y / Whiskey Lake-U
Aug 01, 2019: Ice Lake-Y/U
Sep 17, 2020: Tiger Lake-UP4/UP3

The small-die Y/U platforms have consistently launched within a 7 week window in Q3 each year, and have represented the lead die for every generation except for Haswell and Skylake. Which is also why I still think we might see an Alder Lake-P announcement in August this year.

Intel 7nm risk production has been underway for quite some time at this point with Ponte Vecchio. And I'll allow my unrelenting optimism show through for a moment and say that I think they'll be able to ramp just fine with much shorter cycle times and better defect densities and overall yields than 10nm. I think part of the reason Intel failed so hard on 10nm was because they were trying to achieve densities that were nearly impossible without EUV. However, the term "compute tile" being used in conjunction with Meteor Lake also indicates advanced packaging techniques will be used. So even if cycle times are improved on 7nm, test, sort and packaging will almost certainly take longer. So I'm expecting a similar amount of time between PRQ and RTS as Ice Lake and Tiger Lake.

 

[eek2121]

Anything in 22 or most of 23 is going to be a test device at best. Like Cannonlake was supposed to be. I got the impression with Cannonlake that they thought (hoped) that yields would improve enough in time that they would be able to sell it with the IGP enabled even if it had to be mostly busted.

You must have missed the announcement.

Define efficiency?

Overall battery life. We don't yet have any decent Ryzen 5x00u laptops for comparison for the size/weight and the size of battery. Several models have been announced, but there are few, if any decent reviews and the laptops themselves are nearly impossible to get. Meanwhile, Intel is featured in great laptops such as the LG Gram, Lenovo Yoga, and more, all of which are readily available. Intel Quick Sync probably helps a lot in the power area.

Outside of that, according to AnandTech's latest article, Sapphire Rapids will see a 1H 2022 release with the Aurora supercomputer getting parts this year. HBM based SPR parts will apparently launch in 2H 2022. It will be interesting to see how things compare to both Ice Lake SP and EPYC Milan (and Milan X if AMD pushes it out to everyone).

 

[jpiniero]

The small-die Y/U platforms have consistently launched within a 7 week window in Q3 each year, and have represented the lead die for every generation except for Haswell and Skylake. Which is also why I still think we might see an Alder Lake-P announcement in August this year.

P isn't a small die. That would be the M (2+8). Intel literally just released Tiger Lake-H. August would be too short. P/M also has an updated display engine versus S so it's farther behind.

You must have missed the announcement.

That's the test chips I am talking about.

 

[mikk]

Wait, where are people getting these late 23-24 numbers for MTL from? If it taped out now, in mid-21, that would be 2 1/2 years from tape out till shipping, which would be extraordinarily long. The industry standard is closer to half that.

Because IntelUser2000 told it and people here are repeating it like it's confirmed.

It didn't tape out it "taped-in" which is sort of an Intel-only term that was already discussed ad-nauseaum in this very thread.

Whatever it is, it certainly isn't the classic tape-out we know, as the process itself is far from ready.

And therefore a late 2023/early 2024 MTL release requires additional delays from now on which is even bigger than back with 10nm and Icelake-U. Icelake-U taped-in in June 2017, MTL in May 2021.

 

[uzzi38]

You must have missed the announcement.

Overall battery life. We don't yet have any decent Ryzen 5x00u laptops for comparison for the size/weight and the size of battery. Several models have been announced, but there are few, if any decent reviews and the laptops themselves are nearly impossible to get.

AMD has a small lead in WiFi based tests (usually about 5%), Intel has a larger lead in video playback tests (around 15%). Overall I'd say TGL-U has slightly better battery life, but the two are pretty close to one another at the end of the day.

Of course, that only applies to TGL-U, a much smaller die that features less I/O in general that makes it much easier to power gate for idle power consumption. TGL-H sports similar battery life to CML-H and in some reviews, worse battery life. Both of the two fall very far behind CZN-H/CZN-U/TGL-U.

So "it's nuanced" is probably the best way of putting that.

Anyway, only the lead in video playback tests is down to actual power efficiency differences and that's thanks to crusty old Vega showing it's downsides once again. Calling it power efficiency isn't the best phrasing, most people take power efficiency to refer to performance under load, in which TGL falls behind significantly.

 

[eek2121]

That s simply not true, Ryzen has noticeable higher efficency.

At NBC, and for a comparable Cinebench score, the Ryzen laptop use 95W from start to finish while the TGL one start at 135W and then sustain 121W, methink that it does the whole test closer to 135W than 121W.

From the test with Prime 95 wich help measuring sustained power we can conclude that Ryzen use 77W while TGL require 100W or so.

https://www.notebookcheck.net/Razer-Blade-14-Ryzen-9-Laptop-Review-Blade-Pro-17-Performance-At-Half-The-Size.546754.0.html

https://www.notebookcheck.net/Asus-TUF-Gaming-F15-FX506HM-Laptop-Review-No-Extras-Just-Performance.546842.0.html

I should clarify I am speaking explicitly about thin and light laptops. around 2 lbs or so. Both of those are "H" laptops and both are around twice that weight or more. Also those are two completely different laptops with different size batteries, and one is much larger/heavier than the other. Finally, Prime95 is NOT an indicator of battery performance. It is a worst case scenario that no user will ever reach. To properly measure battery life, one must take a mix of idle performance, load performance, video performance, etc. Basically a variety of workloads.

https://www.amazon.com/dp/B096KN5N9K?th=1

First DDR5 kit made available to buy. Seems strange to get it out there with no product to put it in, but hey.

The memory has to be available for people to be able to buy it when products using it roll out later this year.

 

[eek2121]

P isn't a small die. That would be the M (2+8). Intel literally just released Tiger Lake-H. August would be too short. P/M also has an updated display engine versus S so it's farther behind.



That's the test chips I am talking about.

AMD has a small lead in WiFi based tests (usually about 5%), Intel has a larger lead in video playback tests (around 15%). Overall I'd say TGL-U has slightly better battery life, but the two are pretty close to one another at the end of the day.

Of course, that only applies to TGL-U, a much smaller die that features less I/O in general that makes it much easier to power gate for idle power consumption. TGL-H sports similar battery life to CML-H and in some reviews, worse battery life. Both of the two fall very far behind CZN-H/CZN-U/TGL-U.

So "it's nuanced" is probably the best way of putting that.

Anyway, only the lead in video playback tests is down to actual power efficiency differences and that's thanks to crusty old Vega showing it's downsides once again. Calling it power efficiency isn't the best phrasing, most people take power efficiency to refer to performance under load, in which TGL falls behind significantly.

I've not seen anything conclusive that the Ryzen "H" models outperform the Intel "H" models in terms of battery life. Every site I've looked at avoids even comparing the two. The closest we get is the Intel reference platform test. Notebook Check has some models, but curiously I was unable to find 2 laptops to compare that have similar size, weight, GPU, GPU Power Limit, resolution, refresh rate, and battery size. Hopefully we'll see more reviews as the Intel "H" series rolls out. Regardless...

 

[uzzi38]

I've not seen anything conclusive that the Ryzen "H" models outperform the Intel "H" models in terms of battery life. Every site I've looked at avoids even comparing the two. The closest we get is the Intel reference platform test. Notebook Check has some models, but curiously I was unable to find 2 laptops to compare that have similar size, weight, GPU, GPU Power Limit, resolution, refresh rate, and battery size. Hopefully we'll see more reviews as the Intel "H" series rolls out. Regardless...

The GS76 Stealth in here is the laptop that first came to mind. It's not a particularly long lasting device, suffice to say.

And well, Jarrod's Tech's charts explain the situation in terms of battery life for -H devices pretty well I'd say.

EDIT: Oh ah heck, the mobile app's killed the image quality on that chart. You'll have to just check the video instead.

 

[Exist50]

You must have missed the announcement.

A RISC-V development platform is not a client device by any definition.

 

[lobz]

You must have missed the announcement.

Overall battery life. We don't yet have any decent Ryzen 5x00u laptops for comparison for the size/weight and the size of battery. Several models have been announced, but there are few, if any decent reviews and the laptops themselves are nearly impossible to get. Meanwhile, Intel is featured in great laptops such as the LG Gram, Lenovo Yoga, and more, all of which are readily available. Intel Quick Sync probably helps a lot in the power area.

Outside of that, according to AnandTech's latest article, Sapphire Rapids will see a 1H 2022 release with the Aurora supercomputer getting parts this year. HBM based SPR parts will apparently launch in 2H 2022. It will be interesting to see how things compare to both Ice Lake SP and EPYC Milan (and Milan X if AMD pushes it out to everyone).

Wow, this has really quickly changed from efficiency (courtesy of architecture + process node) to battery life (courtesy of OEM design and your freely chosen measurements, which you did not disclose by the way). No wonder of course, just funny
Still I challenge you to provide evidence and I wish you a lot of fun trying to do that.
Or maybe just tell us what do you mean by "efficient laptop"... efficient in what?

 

[DrMrLordX]

I don’t understand the claims that Intel is somehow behind.

Just take a look at server, workstation, and desktop.

A RISC-V development platform is not a client device by any definition.

It's a pipecleaner with tiny die sizes.

 

[Exist50]

It's a pipecleaner with tiny die sizes.

Sure, but that doesn't match the quote given. So there must be something else, which like as not is Meteor Lake.

 

[Hougy]

Intel Talks 10nm DTCO, EUV Benefits

Intel talks 10-nanometers DTCO and the benefits of EUV on their future 7 nm and 5 nm nodes.

fuse.wikichip.org

Does anyone have access to these? Are there any interesting news?

 

[Abwx]

I should clarify I am speaking explicitly about thin and light laptops. around 2 lbs or so. Both of those are "H" laptops and both are around twice that weight or more. Also those are two completely different laptops with different size batteries, and one is much larger/heavier than the other. Finally, Prime95 is NOT an indicator of battery performance. It is a worst case scenario that no user will ever reach. To properly measure battery life, one must take a mix of idle performance, load performance, video performance, etc. Basically a variety of workloads.



The memory has to be available for people to be able to buy it when products using it roll out later this year.

So now it s no more about efficency but rather battery life wich is extremely implementation dependant, talk about changing the goalposts...

Well...

Microsoft Surface Laptop 4 13 Laptop Review - Too expensive with Intel CPU?

Notebookcheck reviews the Surface Laptop 4 13.5 with Core i7 CPU and metal chassis. Is the CPU upgrade enough to be competitive?

www.notebookcheck.net

Microsoft Surface Laptop 4 13 Review: Is AMD's Ryzen held back on purpose?

Notebookcheck reviews the Surface Laptop 4 13.5 with the AMD Ryzen Edition with Ryzen 5 CPU, 8 GB RAM, and 256 GB SSD storage.

www.notebookcheck.net

That s not exactly a fair comparison since the AMD model use a previous gen APU but that s still relevant for our purpose.

Other TGLs laptops can be better of course, but overall those tests say it all, that is, that this parameter is of little importance since all 8C TGLs come with a dGPU, and that s logical as it would be a mistake to have second rates iGPUs with the most powerfull CPUs.

Btw, the linked 14" Razer Blade using a a Ryzen 5900 weigh 1.8kg, not much more than those "ultra" thins...

 

[repoman27]

P isn't a small die. That would be the M (2+8). Intel literally just released Tiger Lake-H. August would be too short. P/M also has an updated display engine versus S so it's farther behind.

P and M are packages, not dies. But I wasn't very clear in what I was trying to say. The 2+2 LP and later 4+2 and 6+2 LP dies used for the Y/U/UP4/UP3 platforms are almost always the lead dies. (I forgot to include Comet Lake-U 6+2 V1 which launched Aug 21, 2019.) So I really should have said that I still expect the press embargo to lift in Aug or Sep for M5/U9/U15 platforms based on the 2+8+2 LP die.

H has traditionally been based on an HP die and has consistently launched in Q2 for the past 4 years:

Apr 3, 2018: Coffee Lake-H
Apr 23, 2019: Coffee Lake Refresh-H
Apr 2, 2020: Comet Lake-H
May 11, 2021: Tiger Lake-H

This has not affected the U platform release schedule in the slightest so far, so I'm not sure why you think releasing 45-65W desktop replacement chips in May would cause Intel to not release their 15-28W chips (that sell in far greater volumes) in Q3 as they normally do.

And what makes you say the display engine is different for any of the Alder Lake chips? (I actually have no knowledge of the matter, and was just wondering where that info comes from.)

 

[coercitiv]

We don't yet have any decent Ryzen 5x00u laptops for comparison for the size/weight and the size of battery.

Asus ZenBook 13 laptop review: Core i7-1165G7 versus Ryzen 7 5800U. Which one is the better ZenBook?
Both systems have same battery capacity, very similar (as in almost identical) OLED panels from Samsung, same SSD model. Intel system has 32GB LPDDR4X, AMD system has 16GB DDR4 3733.

AMD 5800U system
WiFi Browsing 9h 10m / Load 3h 54m

Intel i7-1165G7 system
WiFi Browsing 10h 40m / Load 3h 2m

TGL platform wins the browsing test by 16%, the AMD platform wins the load test by 29% while also delivering significantly more performance. Pick your poison.

 

[Gideon]

AMD 5800U system
WiFi Browsing 9h 10m / Load 3h 54m

Intel i7-1165G7 system
WiFi Browsing 10h 40m / Load 3h 2m

DDR4 probably murders that WiFi result, usually AMD does better there when equalizing battery sizes. Intel usually wins in video playback or doing nothing (keeping the laptop awake).
Overall surprisingly good WiFi results for an OLED screen though (my wife's x360 Spectre really dissapointed in this regard because of the 4K OLED screen)

 

[jpiniero]

And what makes you say the display engine is different for any of the Alder Lake chips? (I actually have no knowledge of the matter, and was just wondering where that info comes from.)

Various rumors and the Linux drivers. P/M have the Gen13 version of the display while S is still Gen12. I believe the IGP is virtually identical to Rocket/Tiger Lake otherwise but there might be some bug fixes.

 

[lobz]

Asus ZenBook 13 laptop review: Core i7-1165G7 versus Ryzen 7 5800U. Which one is the better ZenBook?
Both systems have same battery capacity, very similar (as in almost identical) OLED panels from Samsung, same SSD model. Intel system has 32GB LPDDR4X, AMD system has 16GB DDR4 3733.

AMD 5800U system
WiFi Browsing 9h 10m / Load 3h 54m

Intel i7-1165G7 system
WiFi Browsing 10h 40m / Load 3h 2m

TGL platform wins the browsing test by 16%, the AMD platform wins the load test by 29% while also delivering significantly more performance. Pick your poison.

If we're already on the subject, I also don't really know who on Earth buys an extremely cost-inefficient ultrathin laptop for....playing back video.

 

[repoman27]

Various rumors and the Linux drivers. P/M have the Gen13 version of the display while S is still Gen12. I believe the IGP is virtually identical to Rocket/Tiger Lake otherwise but there might be some bug fixes.

Ahh, got it, thanks.

So it looks like Alder Lake-P will support the newer XE_LPD (formerly "Display13") display engine. The GPU core and media engine for Alder Lake-M/P/S are all still Gen12 Xe-LP, same as Tiger Lake, Rocket Lake, and DG1. The updates to the Linux drivers so far only seem to indicate increased granularity for various power settings, which makes perfect sense for P. Not sure why that should affect the Alder Lake release order in any way though.

I know pretty much everybody here cares way more about the desktop platforms, but the last time Intel legitimately put the lead HP die ahead of the lead LP die on the production schedule was 8 years ago with Haswell. What would make them change that up at this point?

 

[jpiniero]

I know pretty much everybody here cares way more about the desktop platforms, but the last time Intel legitimately put the lead HP die ahead of the lead LP die on the production schedule was 8 years ago with Haswell. What would make them change that up at this point?

In practice it's more likely to be the same time. This year is probably going to be just K. Some of it is probably embarrassment over Rocket Lake.

Also the P die almost has to be way bigger than the U one.

 

[Zucker2k]

If we're already on the subject, I also don't really know who on Earth buys an extremely cost-inefficient ultrathin laptop for....playing back video.

Cringe, as usual.