This is Intel’s Top secret weapon against AMD

erek

[H]F Junkie
Joined
Dec 19, 2005
Messages
10,785
An incredibly hardcore process technology is coming from Intel to dominate the platform!

"For 2021, Intel is preparing 10+++ as well as a new process node (7nm), and we know this to be the case given that Intel has a contract with the Aurora supercomputer for Argonne, which is due to have Intel’s new Ponte Vecchio HPC accelerator inside, which is partly built on 7nm. That system has to be delivered in 2021, or at least Intel has to hit a number of minimum milestones for that system in 2021, so we might see the first inklings of 7nm at that time. The slide above also says 7nm on EUV.

Assuming all goes to plan, we have 2022 with 7+, then 2023 with 7++ and another new process node, believed to be 5nm. This is expected to be the inflection point with TSMC’s 3nm in terms of density, which could suggest that this is where GAA technology is likely to be.

Going beyond that, 2024 is 5+, then 2025 is 5++ and 3nm. If I were a betting man, and I predicted that Intel’s timeline over the course of five years might slip the best part of 6-12 months, then 2025 might still be in the 5nm / 5+ era. So with Dr. Mayberry saying within 5 years for high volume, the smart money would be GAA coming at 5nm, in 2023-2024."


https://www.anandtech.com/show/1586...anoribbon-transistors-in-volume-in-five-years
 
Is it? Just because Intel says it coming does not mean it will. Intel loss a lot of credibility with their still mostly non present 10nm process that was supposed to have been transitioned to years ago. I hope they can make the above projections/goals on time, better yet advance those times. Could just be another ploy to try to keep their stock from dropping out to the bottom.
 
Buzzword buzzword hashtag buzzword buzzword dominate!

Honestly, I don't even know what you mean by that. What makes it "incredibly hardcore" and what platform exactly is it dominating?
PC / x86_64
 
Intel is just trying SO hard.

But why is UserBenchmark unusable? Did Intel pay them to manipulate results?
Why is Intel's marketing so tricky and unreliable?
Why do they abandon benchmarks, except benchmarks that support Intel?

It sounds like they're a den of thieves and vermin, and that although ONCE UPON A TIME, they had a good engineering idea, it is now a company that is run by evil marketers.

There is no innovation there. If I was a genius engineer with the most supreme CPU enhancement idea, I highly doubt I'd ever just give it to Intel.
Maybe I'd be an engineer at Intel, but they'd NEVER get my best, most valuable ideas. They'd probably pay me a cheap salary to do what they say as some sort of submissive underling, and not credit me properly even if I did contribute massively.

The dark truth that no one likes to admit is that the most glorious ideas are thought of by ONE singular person. Not a team.
Teams refine glorious ideas, but in the end, it's just the fruit of the one single person who thinks it first.

And Intel simply does not have fruit-bearing people at this point.
 
It has taken Intelt three years now to get to 10nm ++, with no solid roadmap releases for 7nm. yet.

That THREE years from introduction of BROKEN dual-cores to feeling confident enough to make 38-core server part (still not available yet).

It's going to take Intel just as long to make something as complex as a massive compute GPU. So, if they release 7nm next year, I'd expect this thing to ship by 2024.


Best case, they fix their new process node do it in half that time, but they still can't ship until the middle of 2023.
 
Last edited:
I think the community is over hearing about news releases. Let me know when it's actually possible to have product in hand in quantity that people can actually use.

"Hey guys, 1nm is coming in 20XV6 and it's going to be amaaaaaaaazing! It will blow your mind it's sooooooo fast!"

Really, every Intel press release should be accompanied by someone saying: "Cool story bro."
 
I think the community is over hearing about news releases. Let me know when it's actually possible to have product in hand in quantity that people can actually use.

"Hey guys, 1nm is coming in 20XV6 and it's going to be amaaaaaaaazing! It will blow your mind it's sooooooo fast!"

Really, every Intel press release should be accompanied by someone saying: "Cool story bro."

heh
 
It has taken Intelt three years now to get to 10nm ++, with no solid roadmap releases for 7nm. yet.

That THREE years from introduction of BROKEN dual-cores to feeling confident enough to make 38-core server part (still not available yet).

It's going to take Intel just as long to make something as complex as a massive compute GPU. So, if they release 7nm next year, I'd expect this thing to ship by 2024.


Best case, they fix their new process node do it in half that time, but they still can't ship until the middle of 2023.
Intel’s 7nm is already on track and they have started that they have been hitting their marks with it and it is on track. Their 10nm fiasco has taught them a lot, unless they have been lying on their investor reports.
 
I haven't paid attention top management theory at Intel, but it might be the same as Boeing in which case they are irredeemably fucked.

Boeing literally cannot make new aircraft or spacecraft because they have anti-engineering management. Management of these projects is make or break. Boeing was doing well until they imported the broken management of McDonnell Douglas when they acquired the failed company. Since then they have been asset stripping the golden goose assuming the golden eggs will magically keep appearing. There are a number of bluechip American companies like this that need to be allowed to die. Keeping them alive is drowning America.

Is Intel one of them? Maybe. It seems like engineers are doing their every best with management thwarting their every step. Eventually the talent leaves and its a zombie running on legacy tech.
 
So intel’s plan to compete is to catch up to AMD in process modes by 2023. Seems legit.
 
  • Like
Reactions: raz-0
like this
Intel is down, but not out. They are a juggernaut, and will come back at some point.

It was just a perfect storm of Ryzen coming though, and a barrage of security and design missteps from Intel.

AMD does not look like they are letting up either, there is no stopping them right now but things can change.
 
  • Like
Reactions: c3k
like this
Intel is down, but not out. They are a juggernaut, and will come back at some point.

I don't believe this.

There is no reason to believe that most people are able to essentially reinvent modern computing from scratch and do it better than what Intel already has.
A person who CAN invent a CPU architecture better than what Intel or AMD has will form their own company.

Geniuses who can invent new CPU architectures will have too much pride for one, but most importantly, why would you give yourself to Intel, considering their ethical stance?
They'd just screw you over. I would never give Intel my magnum opus.

And no one else with those skills would either.

Intel is over. They will live only for as long as their marketing team can keep tricking people into buying it, and as long as they can make incremental improvements to an architecture that is getting older and showing nothing but stagnation.

They have nothing.
 
Intel is down, but not out. They are a juggernaut, and will come back at some point.

It was just a perfect storm of Ryzen coming though, and a barrage of security and design missteps from Intel.

AMD does not look like they are letting up either, there is no stopping them right now but things can change.
I'm not sure I entirely agree with idiomatic or Aegir's analysis of Intel: but we can say for sure just looking at the history that Intel is far more stuck than they care to admit.

Sandy Bridge from 2011 was basically the last real piece of innovation from them. And it has been incredible that they've essentially been able to ride a few die shrinks and the same process technologies, but basically they've been coasting for at least 5+ years. And that number is generous and conservatively low.
They don't need press releases. They need to put up or shut up. Same as AMD before the Ryzen 1 series launch.
 
I mean, I don't think they will just give up. AMD is ahead, but not so far ahead to put Intel out of biz.

And I wasn't saying that will happen right away. It may be a few years down the pipe. But I think they'll be back.
 
I mean, I don't think they will just give up. AMD is ahead, but not so far ahead to put Intel out of biz.

And I wasn't saying that will happen right away. It may be a few years down the pipe. But I think they'll be back.
Here's hoping. I would hate for the future of processors essentially to be down to just AMD (as any monopoly there would be bad for all of us) and Apple (whom will continue to horde all of their own processors). The only other people large enough and with the capitol to enter the market are companies like Qualcomm and Samsung that could also contribute ARM processors to the consumer space.
 
Intel had the wherewithal to push out new processes and chip designs, and they have, but they have all been shit remakes of earlier products, like the 737 max, or the SLS.

Intel certainly has the cash and clout to fix their shit, but if it is a management problem then only the board can fix it. Like Boeing would need such a huge reboot its far easier to liquidate it and start six new companies with the talent and capital. Like I said earlier I have no idea about Intel, but its starting to smell like if I looked into it I would find the same syndrome.

If Intel did die AMD would probably have to roll into ARM and then its a free for all with plenty of competition.
 
ITT: A ton of delusional folks that are greatly exaggerating Intels position.

What reality do you live in if you actually believe intel is just dead? Because we dont exist in the same reality. They are not even really hurting as far as i can tell. They just hit a 5 year high in stock price in January. Im really confused how anyone can look at Intel and really think that they are dying at all...

They just lost the stranglehold they had on the market and there is now some competition. Its not killing them, they are not "done". Thats pretty ridiculous, Intel is still dominating the market contrary to what youd think reading these comments... AMD is barely starting to compete on a real level again and people somehow think intel is now just dead??? The fuck??
 
Is it? Just because Intel says it coming does not mean it will. Intel loss a lot of credibility with their still mostly non present 10nm process that was supposed to have been transitioned to years ago. I hope they can make the above projections/goals on time, better yet advance those times. Could just be another ploy to try to keep their stock from dropping out to the bottom.
Everyone likes to jump on them for being stuck on 14nm but I do not think that their 14nm is directly comparable to TSMC's 7nm. If it was, AMD chips should be way faster per core than Intel chips and that isn't the case.
 
ITT: A ton of delusional folks that are greatly exaggerating Intels position.

What reality do you live in if you actually believe intel is just dead? Because we dont exist in the same reality. They are not even really hurting as far as i can tell. They just hit a 5 year high in stock price in January. Im really confused how anyone can look at Intel and really think that they are dying at all...

They just lost the stranglehold they had on the market and there is now some competition. Its not killing them, they are not "done". Thats pretty ridiculous, Intel is still dominating the market contrary to what youd think reading these comments... AMD is barely starting to compete on a real level again and people somehow think intel is now just dead??? The fuck??


They're not dead, but if Intel's poor management continues for another 20 years, AMD will become a bigger company than Intel. The constant bleeding will have a poor end.

IF AMD continues to own Intel on server CPU + compute GPU, not even government supercomputer contracts are going to keep Intel afloat.

The saddest thing is, both AMD and Intel announced supercomputer jobs within 3 months of each other: AMD is giving us 1.5 PF for $600 million, while Intel is charging $500 million for a measly 1 PF :rolleyes:

https://www.theverge.com/2019/5/7/1...er-amd-cray-doe-oak-ridge-national-laboratory

That kinda mismanagement is going to (EVENTUALLY) cost you high-margin customers. That would mean yet-another round of x86 fab-selling!
 
Last edited:
Everyone likes to jump on them for being stuck on 14nm but I do not think that their 14nm is directly comparable to TSMC's 7nm. If it was, AMD chips should be way faster per core than Intel chips and that isn't the case.
Looks like way faster will be coming up shortly with Zen 3 if AMD 15% IPC claim holds out. So far AMD IPC claims with Zen have been less than actual so Zen 3 maybe the Intel killer chip :D. It is not just that Intel perfects 10nm and then 7nm but converting or building all the foundries to support that. TSMC Arizona 5nm foundry, is estimated to cost 14billion dollars with a paltry 20,000 wafers per month, fab #18 in Taiwan does over 1 million wafers/month. Intel appears to hit a wall for costs going down to the lower nodes is what I am seeing, it is not even a question if Intel can (they have not shown this) it is now more (my opinion) if they can actually afford the number of conversions/building foundry's needed to sustain their previous market production numbers. One may have thought Intel chips were expensive before, what happens when they actually are on 7nm node (if they ever are)?
 
  • Like
Reactions: raz-0
like this
Here's me with my i7 2600k @ 4.6 Ghz still waiting for Intel to release a new desktop CPU architecture that's worth upgrading to. If they'd released one 3 years ago, I would have upgraded back then. Same with 2 years ago, and with last year.

While my Sandy Bridge CPU has help-up phenomenally over the years, I can't put it off much longer. And if Intel doesn't deliver something new that's good, I'll go with AMD like so many others have now. My main reason for not going with AMD is simply unfamiliarity - I don't know if I'll encounter issues or limitations with an AMD CPU that I don't have with an Intel one. Notably, I want to be able to still run Windows 7 without problems created by the CPU and available drivers. I also don't want a power-inefficient and hot-running CPU - which everything Intel is releasing on 14 nm is.
 
  • Like
Reactions: N4CR
like this
Well Windows 7 support is over. I think if you use a new CPU it will still "work" but give you constant nagging messages and maybe some stuff could be unsupported.

Honestly, any of the high end chips from Intel or AMD would probably be better than a 2600k. But if you waited this long, I guess maybe get AMD since it will be a bigger jump (especially for non-gaming workloads, though gaming performance is still good).

I like to constantly upgrade my machine, though the 8700k I got is still quite good. It stands up well even to the newer chips. But I will see what AMD has with Zen 3, maybe I will just sell my machine and start from scratch with AMD.
 
erek does a great job finding fun and interesting articles.
If you don't like them, then feel free to find and post your own articles of interest. :meh:

Agreed, ever since AMD got back on track with Ryzen the AMD bois come out in full force and shitting on anything Intel related, good thing there are other forums that actually have people who discuss these sort of things from a neutral stand-point though.
 
Back
Top