• Guest, The rules for the P & N subforum have been updated to prohibit "ad hominem" or personal attacks against other posters. See the full details in the post "Politics and News Rules & Guidelines."

Discussion Intel current and future Lakes & Rapids thread

Page 462 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

coercitiv

Diamond Member
Jan 24, 2014
4,546
6,258
136
I think the rules of engagement for making fun of the competition are very simple: you're only allowed to make fun of the competition when YOU have the the better product. Try to pull this stunt in a context such as Intel's "glued Epyc" or AMD's "poor Volta" and you get a boomerang effect.

I still remember the first Tiger Lake launch when Intel marketing reps managed to pronounce AMD product names more often than they did for Tiger Lake SKUs. Intel effectively ran the biggest AMD commercial because their marketing department never realized TGL SKU line cannot be remembered and pronounced by normal humans (including themselves).

When marketing does it's job right a company gets tremendous benefits, but when they fail the company gets this:

I hope Intel (marketing) cleans up their act with Alder Lake, both in terms of claims and product naming.
 

scannall

Golden Member
Jan 1, 2012
1,808
1,307
136
Intel gets a little salty when they aren't in the lead. After the M1 was introduced they referred to Apple as a "Lifestyle Company'. I find it more amusing than anything really. Nothing to be upset about.
 

eek2121

Golden Member
Aug 2, 2005
1,287
1,364
136
What cracked me up at the time was that same marketing material quoting a WCCFtech article to support their arguments against Zen.
AMD and Intel both have quoted WCCFTech. Both times I facepalmed.
Remember Phenom true quad core...?
I think the funny thing is marketing pokes about the glue or the likes always end up as a failure.
Not only sounding like a smartass but also beeing a jackass. Marketing Hybris consistently leads to Nemesis.
One have to wonder if its the last time we hear this stuff or some jackass in marketing is still alive after all those failures.
The “true” quad core debate started with the Q6600 and related chips. IIRC the Q6600 actually consisted of two dual core chips. I never understood that argument, because the Q6600 still walked all over AMD stuff.
 

eek2121

Golden Member
Aug 2, 2005
1,287
1,364
136
I think the rules of engagement for making fun of the competition are very simple: you're only allowed to make fun of the competition when YOU have the the better product. Try to pull this stunt in a context such as Intel's "glued Epyc" or AMD's "poor Volta" and you get a boomerang effect.

I still remember the first Tiger Lake launch when Intel marketing reps managed to pronounce AMD product names more often than they did for Tiger Lake SKUs. Intel effectively ran the biggest AMD commercial because their marketing department never realized TGL SKU line cannot be remembered and pronounced by normal humans (including themselves).

When marketing does it's job right a company gets tremendous benefits, but when they fail the company gets this:

I hope Intel (marketing) cleans up their act with Alder Lake, both in terms of claims and product naming.
Maybe with Alder Lake we will see a return to sanity. Probably not, however.
 

IntelUser2000

Elite Member
Oct 14, 2003
7,602
2,462
136
The “true” quad core debate started with the Q6600 and related chips. IIRC the Q6600 actually consisted of two dual core chips. I never understood that argument, because the Q6600 still walked all over AMD stuff.
Don't make a mistake, AMD's approach with Phenom quad core was sound. But Core architecture was so much better it didn't matter.

When Nehalem was introduced, it really showed it's muscle. Of course mostly for servers since for client it was mostly the SMT. For ST it was really only about 5-10% better normalizing for Turbo.

Intel gets a little salty when they aren't in the lead. After the M1 was introduced they referred to Apple as a "Lifestyle Company'. I find it more amusing than anything really. Nothing to be upset about.
I actually thought of it as "We only make CPUs and we're losing to a company that makes an entire system plus people view it as a fashion accessory of sorts".

To me it sounded like it was embarassed. Maybe just difference in viewpoint.
 

moinmoin

Platinum Member
Jun 1, 2017
2,767
3,657
136
I actually thought of it as "We only make CPUs and we're losing to a company that makes an entire system plus people view it as a fashion accessory of sorts".

To me it sounded like it was embarassed. Maybe just difference in viewpoint.
That was my take as well. Also Gelsinger said this at an internal all-hands company meeting back in January, so the target were his (at that point future) employees. Essentially an appeal to their honor.
 

dullard

Elite Member
May 21, 2001
23,044
1,301
126
I think you are going out of your way to find a way to be offended. In your mind you've created this scenario that when AMD said glue no one cared but when Intel did, *GASP* it hit the fan. How about the fact that Intel called Epyc glued together "Desktop" Die? It was much more smartass and just not factual.
Where did I say that I was offended? I said that it is humorous.
 

dullard

Elite Member
May 21, 2001
23,044
1,301
126
Please show me where the term chiplet was used further back than a couple of years ago. Again, putting multiple die on 1 substrate does not equal chiplets. You are completely discounting the work/tech behind AMD's chiplet solution.
I didn't say a thing about AMD's solution. AMD did quite a good job. I'm typing on a Ryzen right now. I'm simply stating that I find it humorous that people hated chiplets until they love them. They loved it when AMD said glue, but hate it when Intel says glue. There is quite a double standard and I don't understand your defense of it.

Do I really have to google "Chiplets" for you? Does 2011 count as "further back than a couple of years ago"? https://researcher.watson.ibm.com/researcher/files/us-lefurgy/micromag2011.pdf How about 2009? https://www.pnas.org/content/107/3/993 How far back until it counts? But that is a false goose chase anyways. Chiplets isn't the only term. Heterogeneous and coprocessor are also commonly used terms for non-monolythic CPUs.
 
Last edited:

Asterox

Senior member
May 15, 2012
630
979
136
Intel gets a little salty when they aren't in the lead. After the M1 was introduced they referred to Apple as a "Lifestyle Company'. I find it more amusing than anything really. Nothing to be upset about.
To put it mildly, or when Intel marketing use "insanely fast" you now that "someone ate too much poisonous garlic". :mask:


 
Last edited:
  • Haha
Reactions: Tlh97 and clemsyn

Hitman928

Diamond Member
Apr 15, 2012
3,644
4,079
136
I didn't say a thing about AMD's solution. AMD did quite a good job. I'm typing on a Ryzen right now. I'm simply stating that I find it humorous that people hated chiplets until they love them. They loved it when AMD said glue, but hate it when Intel says glue. There is quite a double standard and I don't understand your defense of it.
The problem is you're just completely missing context and accuracy around all your posts.

Do I really have to google "Chiplets" for you? Does 2011 count as "further back than a couple of years ago"? https://researcher.watson.ibm.com/researcher/files/us-lefurgy/micromag2011.pdf How about 2009? https://www.pnas.org/content/107/3/993 How far back until it counts? But that is a false goose chase anyways. Chiplets isn't the only term. Heterogeneous and coprocessor are also commonly used terms for non-monolythic CPUs.
I meant more in terms of on these forums or in computer enthusiast circles in general as you said specifically:

In this forum terminology going back the 20 years that I've actively posted here, Chiplet = glue, monolithic = no glue
Obviously there will be tech and research papers talking about chiplets as these things don't come out of thin air but no one even in enthusiast forums such as this one was talking about them and especially not in context of mainstream or even general server CPUs.

I don't know why you are so confused as pretty much everyone else seems to get it. It's pretty simple, when Intel is behind and taking inaccurate pot shots at AMD, Intel's messaging deserves ridicule. When Intel is behind and people are criticizing their lack of vision and rush jobs to compete with inferior tech, the criticism is valid. The same applies to AMD or Nvidia or whomever. If AMD literally just put multiple die on a substrate with the traditional links with terrible latency, bandwidth issues, power balancing issues, traffic conflicts, etc., leading to an inferior and poorly planned product, then sure, Intel's slide would have been valid and everyone would have agreed. That's not what happened, though, and so that slide just makes Intel look foolish.
 
  • Like
Reactions: Tlh97

Ajay

Diamond Member
Jan 8, 2001
9,482
3,947
136
Oh, well, the tiny chips will make some pipe cleaners. Small low perf cores would be useful in embedded designs that requires some general compute ability. Maybe RISC-V + ASIC? Although, doesn't Intel already have ARM cores for ASIC/FPGA designs?

Thanks.
 

eek2121

Golden Member
Aug 2, 2005
1,287
1,364
136
Oh, well, the tiny chips will make some pipe cleaners. Small low perf cores would be useful in embedded designs that requires some general compute ability. Maybe RISC-V + ASIC? Although, doesn't Intel already have ARM cores for ASIC/FPGA designs?

Thanks.
Definitely not a successor to golden cove, but also not as simple as you imply. This should help Intel ramp up 7nm quickly.
 

maddie

Diamond Member
Jul 18, 2010
3,513
2,497
136
Definitely not a successor to golden cove, but also not as simple as you imply. This should help Intel ramp up 7nm quickly.
How do you reason that? ASML production is the limiting factor for worldwide EUV ramps and I haven't seen info to suggest Intel is getting or have a large enough quantity of these litho machines.
 

gdansk

Senior member
Feb 8, 2011
696
387
136
Intel gets a little salty when they aren't in the lead. After the M1 was introduced they referred to Apple as a "Lifestyle Company'. I find it more amusing than anything really. Nothing to be upset about.
The CEO said that. In an internal email. That wasn't marketing. It was meant for internal consumption. They're currently getting outdone by a company who happens to make processors but for whom it is not their entire business. It should be a motivating comparison.
 

Hulk

Diamond Member
Oct 9, 1999
3,113
559
126
Although I don't have the knowledge of this business that many around here do I feel as though Intel has been so used to not having competition that they don't really know how to compete. Let me explain. For years they would release small upgrade at huge prices on THEIR schedule more so than at the need of staying competitive.

Apple, AMD, TSMC, Samsung, etc... they are putting out the best technology they can get out the door as fast as possible. Intel is having a hard time coming to terms with the fact that they indeed have competition and what seems like a minor loss of market or technological lead could become problematic rather quickly and suddenly if they don't move the tiller full port and avoid the iceberg a couple miles in front of them.
 

Ajay

Diamond Member
Jan 8, 2001
9,482
3,947
136
How do you reason that? ASML production is the limiting factor for worldwide EUV ramps and I haven't seen info to suggest Intel is getting or have a large enough quantity of these litho machines.
Well, sadly, that's Intel's own fault. From my understanding they had options/contracts to buy more EUV machines from ASML. They cancelled those as they were behind on 7nm. This didn't bother ASML because demand outstripped supply anyway. Now, Intel has come up begging, but ASML can only expand at a relatively low rate because of the complexity of both the assembly and test, and the complexity of their supply chain.

Such an amazing win - saving on production equipment and offering over $8B to investors instead - freaking corrupt morons. //rant
 

Saylick

Golden Member
Sep 10, 2012
1,056
965
136
Well, sadly, that's Intel's own fault. From my understanding they had options/contracts to buy more EUV machines from ASML. They cancelled those as they were behind on 7nm. This didn't bother ASML because demand outstripped supply anyway. Now, Intel has come up begging, but ASML can only expand at a relatively low rate because of the complexity of both the assembly and test, and the complexity of their supply chain.

Such an amazing win - saving on production equipment and offering over $8B to investors instead - freaking corrupt morons. //rant
Intel dropped the ball so hard from 2015 onward that it didn't even know where the ball was for a few years. To their credit, their timing was god awful when it came to development of their 10nm node. Back in 2015, EUV wasn't quite mature/ready for HVM so they pushed forward with using SAQP and DUV for the 10nm node, which as we all now know had developmental issues for being overly aggressive. By the time they resigned to relaxing the node a few years later, EUV was mature enough such that TSMC and Samsung started ordering them from ASML so that they could develop their future nodes. Intel was still stuck trying to fix 10nm that they put 7nm on the backburner, letting TSMC and Samsung hog up EUV machines for the foreseeable future. So here we are today with Intel in the process of ramping up 10n, after Intel finally admitted that their 7nm node was going to be behind schedule with only a fraction of the world's EUV machine supply. Intel is going to have to pull the strings of the US government if they want to get more machines, because I wouldn't be surprised if ASML just tells them to get to the back of the line.

I mean, just look at the projected volume ramp for 7nm... and then compare it to TSMC.


Source:
 
Last edited:

maddie

Diamond Member
Jul 18, 2010
3,513
2,497
136
Intel dropped the ball so hard from 2015 onward that it didn't even know where the ball was for a few years. To their credit, their timing was god awful when it came to development of their 10nm node. Back in 2015, EUV wasn't quite mature/ready for HVM so they pushed forward with using SAQP and DUV for the 10nm node, which as we all now know had developmental issues for being overly aggressive. By the time they resigned to relaxing the node a few years later, EUV was mature enough such that TSMC and Samsung started ordering them from ASML so that they could develop their future nodes. Intel was still stuck trying to fix 10nm that they put 7nm on the backburner, letting TSMC and Samsung hog up EUV machines for the foreseeable future. So here we are today with Intel in the process of ramping up 10n, after Intel finally admitted that their 7nm node was going to be behind schedule with only a fraction of the world's EUV machine supply. Intel is going to have to pull the strings of the US government if they want to get more machines, because I wouldn't be surprised if ASML just tells them to get to the back of the line.

I mean, just look at the projected volume ramp for 7nm... and then compare it to TSMC.


Source:
The future appears quite clear for anyone without bias.

2023 production for nodes using EUV
TSMC = 370 Kwpm
Intel = 20 Kwpm
 

JasonLD

Senior member
Aug 22, 2017
352
317
136
The future appears quite clear for anyone without bias.

2023 production for nodes using EUV
TSMC = 370 Kwpm
Intel = 20 Kwpm
Wouldn't that be enough for whatever Intel has planned for 2023? Since logic chiplets for Meteor Lake is going to be the only thing Intel will be using their 7nm.
 

Saylick

Golden Member
Sep 10, 2012
1,056
965
136
Wouldn't that be enough for whatever Intel has planned for 2023? Since logic chiplets for Meteor Lake is going to be the only thing Intel will be using their 7nm.
Possibly, Mizuho Securities shows Intel 7nm being risk production all the way out to the end of 2022, so maybe it would be ready for HVM in 2023. Also, whether or not 20 kwpm is adequate depends on how much of their line-up they need to shift over to Meteorlake to be competitive with AMD by 2023, who will most likely be at the tail end of Zen 4, with Zen 5 on TSMC N3 likely coming in H1 2024. N3 looks to ramp faster than N5, which itself ramped faster than N7, so Apple could potentially move onto N3 faster than normal, leaving a glut of N5P volume for AMD in 2023. If AMD is pushing out 50 kwpm of Zen 4 in 2023, 20 kwpm for Meteorlake may not be sufficient.
 

ASK THE COMMUNITY