I hope he is in good health. Jim Keller is a semiconductor rock star if ever there was one, and he led the design of the CPU that keeps my car from hitting squirrels, the CPU that sits in my PC, and the CPU that sits in my phone.
Most likely this means that Keller has done the hard part of the next gen Intel CPU already and he is seeking for something new and more interesting... Again :)
Or he thinks its a Phuckfest beyond repair? Personally, I think he's got health issues.. Intel would be his greatest challenge , and they really need someone like him atm.
Keep in mind before Intel he only needed two years at Tesla. Before that, he produced Zen for AMD in just three years. Thus, besides the "personal reasons", there's a good chance he was able to deliver (or at least set the course) for at least one product.
I have a feeling that he butted heads with too many people at Intel and wasn't able to do the work that he needed to do. Intel does have a very arrogant attitude that they're the best at everything for no good reason. Perhaps he was unwilling to drink the Intel Kool-Aid.
Here’s why I feel he did: he probably couldn’t get results he wanted having to always fight tooth and nail with the bureaucratic monstrosity Intel has become. Honestly, I am surprised he lasted as long as he did because I personally know of employees who have left because of the sluggard of a rotting corporation it has become.
I work with just such a previous employee. His prediction was that unless they rebuilt the management teams and their whole culture, no fixer was going to be able to right the ship that is Intel.
My bet he have realized the lead in cpu development is right now in the large hands of Haifa Team. Honestly at Haifa there are more talented cpu engineers than Keller, so he had only a job of coordination. Done its job he have resigned. He have not a future in Intel.
You and the other commenters have NO IDEA what you are talking about. Go and look at their senior VP hirings from outside intel over the last 5 years. There is nobody from Intel left. It's all Qualcomm, Apple, Google, AMD, etc. Anandtech is way more "arrogant" than Intel, you guys are almost to Apple levels.
I have been frequently visiting this site since it was one of the only tech sites. I'm talking back in the Celly 300a days. Long, long ago...... I haven't noticed any excessive arrogance coming from this site, or more specifically, Anand La Shimpi himself.
Do you have even a tiny bit of evidence to back your claim up? I don't even see how (Unless they are bad-mouthing other tech sites, which they don't) a tech website CAN be arrogant in the manner you have suggested.
You compared Intel's arrogance to AnandTech's "arrogance" (I quote because one of these is not like the other, lol).... Seriously? Intel might as well be Nvidia when it comes to arrogance.
You shouldn't have even attempted to make such a ridiculous comparison. Yes, in both cases, arrogance would be the similarity. That only works if you're comparing something at least a little bit similar. Comparing Intel's arrogance (One of 2 of the biggest CPU manufacturers in the world) to AnandTech's (A fucking WEBSITE) makes you look like a goofball. I'm sorry, it truly does.
People always have this impression that microarchitectures are implemented in the space of a year or two or something from conception to actual silicon. It's not even close to being that fast
He was there as leadership to help organize the engineering team more than being THE architect. I'm assuming he couldn't run things the way he wanted and maybe some high level architectural decisions were not welcome by non engineer corporate leadership.
I don't think that that will show his greatest influence. Willow Core was effectively done when he arrived. Rocket Lake, being a 14 nm backport, he could have influenced but that was clearly a plan B due to a massive hole in the desktop market, something you don't distract your A team with. Rather it'll be the parts next year and in 2022 that'll be a show case for his performance.
The ship could not be turned around. Sounds like he decided to cut his losses. It's difficult trying to reform a system much in need of reform when the folks in that system try to thwart your efforts.
When he joined Intel Willow Cove's design was almost certainly "locked" already. Perhaps he had some input in Rocket Lake, more input in Golden Cove, but apparently his greatest focus was meant to be on Ocean Cove (the core of what's currently known as Meteor Lake), Golden Cove's successor.
I recall that Ocean Cove was pre-pre-announced just a few weeks after he joined Intel as a "clean sheet" design, and Jim Keller was to lead its design. Perhaps its design is done or almost done already, but it will take a while before Intel has engineering samples and it taped out. On top of a brand new design (as "brand new" as can be done nowadays) Meteor Lake is going to be the first Intel CPU series fabbed on their 7nm node, which will employ EUV extensively.
Meteor Lake should also have quite a beefy (Gen14) iGPU, which will be fed by DDR5 and LPDDR5. Release date? Q1 2023 at the absolute earliest, Q4 2022 if a miracle occurs.
Clean slate architectures takes years to make. Zen took from about 2011 to 2016 tape out. Development time only takes longer as process nodes shrink and the amount of transistors you have to work with.
Intel has had 5 years of break to create this miracle architecture. They most probably were in the middle of it in 2018 when they hired Keller to give his input on the matter. Currently Ocean Cove must be in design/verification phase.
given that most of those transistors have gone to building caches/buffers and the like rather than new logic, it's difficult to see why tape out should be taking longer. at least from an arch point of view. to the extent that multi-core communication demands new designs, that too ought to be a solved problem.
I doubt he had anything to do with Willow Cove, as that architecture most likely already went into validation when he joined Intel. Maybe he was too late even for Golden Cove, so we will not see his contribution until Ocean Cove.
Rocket Lake is the LAST 14nm desktop and Cooper Lake (4-8S Xeon) is last 14nm period. Both of those are launching this year. Time to update your little skit.
Actually, sounds like the joke's good for at least the rest of the year. ;) TBH, if anything it's a grudging compliment since they've actually shown how solid their 14nm is, and frankly how great their underlying architecture's been.
On the contrary, I am happy if it means Keller is free of the nonsense and numbskullery he probably had to deal with there on a daily basis. They probably couldn’t handle his free-thinking, maverick style and it was probably for the best of everyone that he parted ways: for him to breath creatively again and for Intel to continue waning. By the looks of Lakefield which I think he had some hand in with its evolutionary package, it just does not have the evolutionary performance it was hyped up to have. Remember that Lakefield is using Sunny Cove which has an 18% IPC improvement over Skylake so Lakefield’s 3 GHz boost is only roughly equivalent to 3.54 GHz on Skylake. So when Intel compared Lakefield to the Core i7-8100Y which has a 4.2 GHz boost clock, they must be speaking of sustained loads in thermally or energy throttled situations and not actual boost performance. This potentially makes it significantly slower than the Core i7-8100Y it was compared to unless its Sunny Cove is heavily modified and suddenly gained more IPC than advertised, which I highly doubt. Keller likes his name associated with success so good for him for walking and leaving this dinosaur corporation to fall into obscurity from their idiotic internal operations.
Intel purported a “12% SINGLE threaded” performance uplift for Lakefield over the Core i78500Y. But it does not compute since that 8500Y has a 4.2 GHz boost clock. I really think that number is for a very specific, thermally and energy constrained scenario. Honestly, the more I look at Lakefield, the more I see “over-engineering” and “falling short“ in the design goals.
I hope all's well with him and his family, since the fact that he's sticking around as a consultant makes it unlikely that "personal reasons" is being for some of the more common euphemisms: If its because of a disagreement with management on direction, they wouldn't want him as a consultant, if its because he had his eyes set on another opportunity he wouldn't agree to stick around, and if it was for any kind of alleged (or substantiated) misconduct on his part, Intel would have promptly shown him the door (as they did with ex-CEO Brian Krzanich when it came to light that he had a consensual relationship with another employee).
Oooor, "staying on as a consultant" is a nice way of saying "gardening leave". In other words, not allowed to work for any competitor in the 6 months following his departure.
I'm going to guess that Jim just ran into to much institutional inertia at Intel and decided that he'd be wasting his time trying to turn the Intel super-tanker in the direction of more design automation. Intel is a very insular culture that doesn't take much to outsiders coming in and telling them what to do. Of course, they desperately need some input from outside to help them, but it's generally not appreciated.
AMD, Arm, Apple, Google, MS, Nvidia, Tesla all got their boards and chips generations ahead of Intel. Jim Keller arrived at Intel. Old chips, phantasy product pipeline, costly and very limited production, no shrink path. This means that the Intel herd is headless. Was headless for a long time. The herd built social constructs rather than chips corps and consumers want. If we treat corporations like individuals - then we analyze them a partially sentient beeings. Corporate psychiatry. In case of Intel this must follow with treatment. Jim Keller is not what Intel needs. Intel needs to clear the stable. Bring its moohing herd to the slaugtherhouse and grow a new herd. Then Jim Keller has a chance of reform. Because of this I moved our hardware procurement to AMD, Nvidia. Dr. Lisa Su and Jim Keller are exceptional, Elon Musk too. To me Intel is done. Its over. Expensive, greedy, slow innovations. No sell to me, no security - they moved hw design into the warzone Israel. Israel is running a Gulag. I want ethical sourced hardware. Intel cant.
Funny. Intel has a lot more talented engineers besides Jim. Intel did Sandy bridge without Jim. Intel did optane without Jim. Intel did a lot of innovation without Jim. They needed Jim now because they wanted the best product to come out after this long stagnation period. And it will come. Jim resigned for actual personal reasons, you can read about it in the press, about his family. You're just a conspiracy believer and should quit that. The fact that Intel core from 2015, that is Skylake, competes with Zen 2 parts which are much much newer talks clearly about who has the upper hand on innovation. Sure, Intel has been hand cuffed because of their fab issues, but when it comes to ip, I don't think they have a talent problem.
Intel is publicly traded. Before they go down, you can bet your ass there will be a complete change of management, if need be. They're still doing quite well, though, so there's no real pressure.
Intel's fab has a big ongoing cultural problem and is an albatross around Intel's neck. Unfortunately current CTO inherited a trail of cronies in TMG (TD) management chain from the previous one who was unceremoniously "retired". Nothing much has changed in last 2 years for TMG. The list is long.. SVP heading overall PTD, yield VP, 10nm yield manager - who oddly has kept his job/promoted in spite of behavioral issues and consistently poor performance - all Rennsaeler grads btw (some kind of favoritism going on there as well) , litho manager calling shots in 10nm, process integration manager in charge of COAG,.....
Let's not forget that Intel's budget is orders of magnitude higher than AMD's. Let's not forget that Skylake doesn't compete with Zen2, not even close, in metrics that matter such as performance per watt, or compute density (cores per socket) / MT performance and I/O (for servers).
Your comment is spoken like a clueless gamer who thinks that a few extra % average FPS in a latency-sensitive workload on a dead-end architecture without chiplets and any future is the last word in microrprocessor technology. Let's also not forget the security vulnerabilities.
AMD has innovated more in the processor industry between 2017 and today, than Intel has in the last decade. And you need to accept that.
Between 20217 and now i can see only an Ark----->Zen, Zen 2 is only a small revamping (comparable with Skylake in per core raw performance) and Zen 3 that is is only a better arrangement of the cache stack of Zen 2. More or less have done Intel with Sunny Cove (better SMT), Willow cove (even better SMT), and soon Golden Cove ( nice strong surprise for many) and Ocean Cove (another strong news).
Accept this :)
The only contribute of AMD is to have joined with TSMC ( so kudos to TSMC, definitively NO to AMD ). The main AMD sin is the chiplet desing without a really good interconnection.
"comparable with Skylake in per core raw performance" only when the Skylake CPU has a 10+% clock speed advantage and using a lot more power. We have seen the i7-7700k vs Ryzen 3 3300X review and the 7700k loses far more often than it wins. Both are 4c/8t CPUs but the 7700k has less than a 10% clock speed advantage. That means that Zen 2's IPC advantage over Skylake allows its lower clocks to overtake the Skylake chip.
Sunny Cove has a better IPC on paper than Zen 2, however, when running none SPEC applications it seems to falter. Look at the Acer Swift 3 review that has the best Ice Lake i7 and the near best Ryzen 4700U. Yes the 4700U has double the cores and a 100MHz boost clock speed advantage. However, Ice Lake is supposed to have a 7ish% IPC advantage over Zen 2 but loses on many ST applications by more than the 2.5% boost clock disadvantage should allow.
I just read jibberish about it chiplet and node and blah blah blah here. What you all need to accept are just two simple facts:
- Intel's comfortable market position in the past decade made them seek architecture evolution, not design evolution. Now their game is milked up. - AMD went through the last decade with a fundamentaly flawed design (bulldozer/vishera) that turned up also as an architecture dead-end.
Now it is just the same history, with sides exchanged.
And yes, fanboys will be fanboys, both sides. You suck.
Most people of Jim Keller's seniority and stature have significant non-compete clauses in their contracts, which often explains why they remain "consultants" for up to and over a year after they either resign or get fired. Basically, one has to "age out" of the inside knowledge before moving on, but that time is paid for.
Just because they'll lose in the end doesn't mean a company with deep pockets and lots of corporate lawyers can't make your life a living hell. Non-competes are not enforceable in many places, especially since with specialized jobs it amounts to slave labor if it could be enforced that you can essentially never work in your industry again.
There are two main situations with such non-competes: 1. You (or Jim Keller) have another gig lined up, and you or they are willing to cover the legal costs and potential liability 2. You don't, but are paid handsomely for not running to your next gig (as a "consultant"). So, unless he already has a new employer who is willing to cough those legal fees up, basically sitting at home or in your office reading Anandtech and other things and getting paid for it is not a bad gig. Plus, many companies in the Valley will tie the full vesting of stock grants and options to adherence to the non-compete. Lastly, 6 months is harmless compared to some others; can be years. Those are the ones that get challenged in court.
Intel has massive problems at the moment having become complacent and uncompetitive after many years of milking a high profitable x86 market. It's mind-boggling to me that supposedly the most brilliant corporate minds let so many opportunities pass by over the last ten years.
> It's mind-boggling to me that supposedly the most brilliant corporate minds let so many opportunities pass by over the last ten years.
Greed is one heck of a powerful drug. It will can even blind the most brilliant Ivy League financial whizzes if they don’t constantly check themselves which history has proven time and again. I really think history is going to be repeated itself here again very soon and quite possibly not to Intel’s satisfaction.
disagree they got complacent or let things pass by - they were just beaten in each endeavor outside their core competency: 2010-2013 - smartphone/tablet SOCs that went nowhere. Shame, because baytrail was a great chip, just got no wins beyond ones they bought. 2013-2019 - cellular modems that put them in the predicament they've been in with shortages for the past 2 years (because they MUST fab apple modems before they can fab their own stuff. This was supposed to be overhead absorption and a way to keep their n-1 fabs near full utilization. Great idea, but 10nm's failure meant they were stuck with low margin parts and a contractual obligation on their top process) IoT - 2013-2017ish? - this was probably the most optimistic of their splashes out, how they thought they could make any kind of margin on commodity microcontrollers is beyond me. Maybe this was n---- fab utilization boosts? (I'm not sure how far back they keep fabs, if their 45 or 32nm stuff is even still running)
>disagree they got complacent or let things pass by - they were just beaten in each endeavor outside their core competency:
They were ramming atom down everyone's throat and they lost. It is obvious they had that coming and it is their fault.
They practically gave the modem away to Apple and they were still not that great (power, signal strength, stability, etc). They just didn't spend enough resources to catch up where Infineon throw in their towel on. They think as long as they have something and if they bundle it with x86 they can sell it.
IoT? You got to be kidding me. Marvell, Qualcomm, and Broadcom would rule that long before Intel would. They actually spend time and resources to get things done while Intel spend its time trying to force an atom down any project. Yup, the same as smart phone and tablet. They don't want to go down the market of $5 SoC so they shove an atom down and charge people $20, or more if they want a celeron.
Yeah "Lisa" Su let a huge opportunity pass her by by not being able to capitalize in a once every 12-15 year opportunity where Intel had issues. She is supposedly brilliant - terrible at contracts and capitalizing on a competitors failures...
Are... Are you high? Sue joined as CEO of AMD only in 2014, and if you're not aware of her brilliance, go google her. It's easy enough, even for a blind fanboy like you.
She was the highest-paid CEO last year precisely because she has been doing an incredible job. And is relatively humble about it too. Oh, and people know her name, unlike whoever is running Intel. That says a lot.
And new, innovative, better products can't just be turned out. They need years of development to come to fruitition. Su and AMD don't really have much control over when something will be ready.
You cannot just divest from your own fab overnight, Lisa did the unthinkable and turn the ship right. Both AMD's old fabs and AMD are now on the right track after the divorce, instead of following the "real men have fabs" policy from Jerry Sanders.
> Tesla, AMD, Apple, AMD (again), and PA Semiconductor
This is not even in chronological order (though it is nearly, but not quite, reverse chronological order), so I don’t know why AMD merited two mentions.
This is a big one. Intel with their recent "Woke" drivel and the nonsensical M&A instead of their bread and butter x86 industry are =going sideways, I guess Keller hit an immovable rock in his path to make them great again but the corporate PR didn't allow.
Highly doubt Intel is going to do anything good. Woke fantasy land of that CA state and it's ridiculous BS nowadays, I think it's irreparable.
I just wish all the best for the Hero of this era of Semiconductor engineering.
I'm sure Lisa has already asked him if she could have his old office unlocked, cleaned, and the couch pillows fluffed for him when the 6month clock ticks its final tock.
Zen 6-10 or even Zen' aren't gonna design themselves.
"This is a big one. Intel with their recent "Woke" drivel and the nonsensical M&A instead of their bread and butter x86 industry are =going sideways, I guess Keller hit an immovable rock in his path to make them great again but the corporate PR didn't allow."
I don't think you understand how a business work. It is true that the x86 is one of the core businesses, but you can't stick around and not diversify. Perfect example of corporates that do note diversify much, Nokia and Blackberry. They are not chip producers for sure, but if you only survive on what you think is not going to go away, then you are just living in your own world. As it stands, ARM, AMD and Nvidia are chipping away at their stronghold. Corporates are making their own processors, so its a sign that they need to look elsewhere to ensure they don't sink.
check out a recent youtube clip from Jim @ AdoredTV about some insights he got about whats its like to to work at Intel, the clip is called "Inside Intel" if the link doesnt paste correctly... https://www.youtube.com/watch?v=agxSclh27uo
Considering Intel has all the replacements ready this isn't really "whoa out of the blue" departure like Intel's CEO recently. There might be something bad that forced his leave, or his work is done and he is retiring. Or maybe moving on to another company after a break. I guess IBM to finish his tour of the relevant US semi companies.
makes you wonder if like many brilliant folks he may be on the edge of the autistic spectrum and just needs a specific challenge to keep him motivated, seems his pattern at least for an outsider looking in
Good luck to him and hope the personal matters are not too serious.
From my perspective Intel does not have a design or performance issue but a issue of manufacturing and security.
Manufacturing could be a shareholder issue as they are not willing to invest in the capital expenditure to move to lower nm fabs as they “feel” they still have the performance advantage. It could be a “ego” thing where they just do not see the need to invest.
Another advantage Intel has is in the business enterprise space where their Microsoft partnership is benefitting them a lot. It is really expensive to by software especially database software for AMD. Their core count actually counts against them and per core performance is better in that space which makes it more economical to go Intel, however the power efficiency lag is starting to catch up with Intel.
The other area of concern for Intel is the security space as they just seem to be hit with issue after issue as we go along with not end in sight. Their predictive performance algorithms that gives them that IPC advantage seems to be prone to security exploits and this could be causing some serious conflicts inside Intel. Address the security and loose IPC performance crown?
From my perspective AMD is smoking at the moment. If Intel can however address their fab and security soon they will stay on top for the future.
The AMD / INTC dynamic is different this time as INTC doesn't control or dominate in MFG. INTC has historically just killed AMD due to their MFG scale and performance. If AMD had a good design INTC would create design and improve MFG process well ahead. AMD was right to spin off Glofo and now they could be competitive or more in the next 10 years so I don't think INTC will just come back and dominate like before. I see a more evenly distributed market as the next few years unfold.
The problem is, AMD share the lower nm cost with Apple, Qualcomm, and others while Intel is doing it all by themselves. There is a different economy of scale and AMD is now able to hitch hike with Apple and Qualcomm to beat Intel. This is new and has never really happened before, when foundry was always behind Intel for one or two generation. Now Apple and Qualcomm will pay for this due to power consumption reason rather than cost performance reason.
Intel didn't have the wireless dominance to justify the lower nm cost, so they are now losing the edge to AMD. If you think AMD catching up is bad, wait till Apple starts switching their mac to their own CPU on ARM, that could really get Intel out of consumer space over time, starting with chrome book then gradually to gaming laptop.
Intel has had an exodus of talent in the last 1 or 2 years, The CEO needs to sell off its fabs and manufacturing unit to the highest bidder (Gloflo or Samsung?) ASAP and buy one of ARM/Nvidia/AMD/Qualcomm with the proceeds for Intel to stay relevant.
We’ve updated our terms. By continuing to use the site and/or by logging into your account, you agree to the Site’s updated Terms of Use and Privacy Policy.
Facebook
Twitter
RSS
127 Comments
Back to Article
Sivar - Thursday, June 11, 2020 - link
I hope he is in good health. Jim Keller is a semiconductor rock star if ever there was one, and he led the design of the CPU that keeps my car from hitting squirrels, the CPU that sits in my PC, and the CPU that sits in my phone.Hifihedgehog - Thursday, June 11, 2020 - link
+1supdawgwtfd - Friday, June 12, 2020 - link
To bad that CPU can't stop you from hitting a flip stationary truck directly in front though...Alejandropoi - Friday, June 12, 2020 - link
Grow up.maxnix - Friday, June 12, 2020 - link
You should! Adultery as an adult is much more fun than infancy as an infant.Opencg - Saturday, June 13, 2020 - link
I don't think he meant it as a rip on Jim Keller. Just a legit concern over self driving systems as a whole.Noodle-Naut - Friday, June 12, 2020 - link
He is a hardware guy. You are talking software glitch.mukiex - Wednesday, March 17, 2021 - link
That wasn't his CPU tho?haukionkannel - Friday, June 12, 2020 - link
Most likely this means that Keller has done the hard part of the next gen Intel CPU already and he is seeking for something new and more interesting... Again :)dromoxen - Sunday, June 14, 2020 - link
Or he thinks its a Phuckfest beyond repair? Personally, I think he's got health issues.. Intel would be his greatest challenge , and they really need someone like him atm.bug77 - Monday, June 15, 2020 - link
Keep in mind before Intel he only needed two years at Tesla. Before that, he produced Zen for AMD in just three years. Thus, besides the "personal reasons", there's a good chance he was able to deliver (or at least set the course) for at least one product.YB1064 - Saturday, June 13, 2020 - link
I wonder how the HardOCP guy is doing at Intel? Butting heads was his specialty.bug77 - Sunday, June 14, 2020 - link
Not well, I'm afraid: https://www.reddit.com/r/hardware/comments/bvhxs1/...drexnx - Thursday, June 11, 2020 - link
I guess his work is done and Rocket Lake/Willow Cove will be Sandy Bridge levels of good, or the ship couldn't be turned around after alltrparky - Thursday, June 11, 2020 - link
I have a feeling that he butted heads with too many people at Intel and wasn't able to do the work that he needed to do. Intel does have a very arrogant attitude that they're the best at everything for no good reason. Perhaps he was unwilling to drink the Intel Kool-Aid.realneil - Thursday, June 11, 2020 - link
This sounds about right.Hifihedgehog - Thursday, June 11, 2020 - link
Here’s why I feel he did: he probably couldn’t get results he wanted having to always fight tooth and nail with the bureaucratic monstrosity Intel has become. Honestly, I am surprised he lasted as long as he did because I personally know of employees who have left because of the sluggard of a rotting corporation it has become.FullmetalTitan - Thursday, June 11, 2020 - link
I work with just such a previous employee. His prediction was that unless they rebuilt the management teams and their whole culture, no fixer was going to be able to right the ship that is Intel.watzupken - Thursday, June 11, 2020 - link
I agree. Unlike in AMD, Intel may not be as easy to navigate and manage.Gondalf - Friday, June 12, 2020 - link
My bet he have realized the lead in cpu development is right now in the large hands of Haifa Team. Honestly at Haifa there are more talented cpu engineers than Keller, so he had only a job of coordination.Done its job he have resigned. He have not a future in Intel.
Tams80 - Friday, June 12, 2020 - link
Haha, someone sounds bitter.I think with that language you might be better of being playwright than an electrical engineer though.
Spunjji - Monday, June 15, 2020 - link
Big lol @ this one.jjjag - Friday, June 12, 2020 - link
You and the other commenters have NO IDEA what you are talking about. Go and look at their senior VP hirings from outside intel over the last 5 years. There is nobody from Intel left. It's all Qualcomm, Apple, Google, AMD, etc. Anandtech is way more "arrogant" than Intel, you guys are almost to Apple levels.nexxusty - Friday, June 12, 2020 - link
AnandTech arrogant? How so?I have been frequently visiting this site since it was one of the only tech sites. I'm talking back in the Celly 300a days. Long, long ago...... I haven't noticed any excessive arrogance coming from this site, or more specifically, Anand La Shimpi himself.
Do you have even a tiny bit of evidence to back your claim up? I don't even see how (Unless they are bad-mouthing other tech sites, which they don't) a tech website CAN be arrogant in the manner you have suggested.
You compared Intel's arrogance to AnandTech's "arrogance" (I quote because one of these is not like the other, lol).... Seriously? Intel might as well be Nvidia when it comes to arrogance.
You shouldn't have even attempted to make such a ridiculous comparison. Yes, in both cases, arrogance would be the similarity. That only works if you're comparing something at least a little bit similar. Comparing Intel's arrogance (One of 2 of the biggest CPU manufacturers in the world) to AnandTech's (A fucking WEBSITE) makes you look like a goofball. I'm sorry, it truly does.
Morky - Saturday, June 13, 2020 - link
I think he means Anandtech commenters, not Anandtech itself.mikato - Sunday, June 14, 2020 - link
"There is nobody from Intel left."Left at where? At Intel? Did they normally hire senior VPs from within in the past?
hafizmajid - Sunday, June 14, 2020 - link
Don't forget recent hiring from GloFlowut - Thursday, June 11, 2020 - link
People always have this impression that microarchitectures are implemented in the space of a year or two or something from conception to actual silicon. It's not even close to being that fastdrexnx - Thursday, June 11, 2020 - link
okay, golden cove then.(he's been there over 2 years already)
Santoval - Thursday, June 11, 2020 - link
More like Ocean Cove. This was already reported as the primary reason he was hired.Santoval - Thursday, June 11, 2020 - link
p.s. Sorry, I just noticed you are the same person I replied above. Ignore that second reply.jaker788 - Thursday, June 11, 2020 - link
He was there as leadership to help organize the engineering team more than being THE architect. I'm assuming he couldn't run things the way he wanted and maybe some high level architectural decisions were not welcome by non engineer corporate leadership.yeeeeman - Friday, June 12, 2020 - link
Ocean Cove. Golden Cove is Alder Lake and Intel already has prototypes for it in the labs.Kevin G - Thursday, June 11, 2020 - link
I don't think that that will show his greatest influence. Willow Core was effectively done when he arrived. Rocket Lake, being a 14 nm backport, he could have influenced but that was clearly a plan B due to a massive hole in the desktop market, something you don't distract your A team with. Rather it'll be the parts next year and in 2022 that'll be a show case for his performance.Deicidium369 - Thursday, June 11, 2020 - link
Cove not core. Rocket Lake has been on the slides for ages - unamed then, the last 14nm based on next gen arch.Truthy - Thursday, June 11, 2020 - link
The ship could not be turned around. Sounds like he decided to cut his losses. It's difficult trying to reform a system much in need of reform when the folks in that system try to thwart your efforts.name99 - Thursday, June 11, 2020 - link
Certainly looks that way to me.Let's see if those "personal reasons" prevent him from popping up at some other company (NUVIA seems a good bet...) in a few months...
Santoval - Thursday, June 11, 2020 - link
When he joined Intel Willow Cove's design was almost certainly "locked" already. Perhaps he had some input in Rocket Lake, more input in Golden Cove, but apparently his greatest focus was meant to be on Ocean Cove (the core of what's currently known as Meteor Lake), Golden Cove's successor.I recall that Ocean Cove was pre-pre-announced just a few weeks after he joined Intel as a "clean sheet" design, and Jim Keller was to lead its design. Perhaps its design is done or almost done already, but it will take a while before Intel has engineering samples and it taped out. On top of a brand new design (as "brand new" as can be done nowadays) Meteor Lake is going to be the first Intel CPU series fabbed on their 7nm node, which will employ EUV extensively.
Meteor Lake should also have quite a beefy (Gen14) iGPU, which will be fed by DDR5 and LPDDR5. Release date? Q1 2023 at the absolute earliest, Q4 2022 if a miracle occurs.
jaker788 - Thursday, June 11, 2020 - link
Clean slate architectures takes years to make. Zen took from about 2011 to 2016 tape out. Development time only takes longer as process nodes shrink and the amount of transistors you have to work with.yeeeeman - Friday, June 12, 2020 - link
Intel has had 5 years of break to create this miracle architecture. They most probably were in the middle of it in 2018 when they hired Keller to give his input on the matter. Currently Ocean Cove must be in design/verification phase.FunBunny2 - Friday, June 12, 2020 - link
given that most of those transistors have gone to building caches/buffers and the like rather than new logic, it's difficult to see why tape out should be taking longer. at least from an arch point of view. to the extent that multi-core communication demands new designs, that too ought to be a solved problem.Deicidium369 - Thursday, June 11, 2020 - link
Rocket Lake is Tiger Lake backported. Willow Cove and a much smaller iGPU based on Xe LPajp_anton - Saturday, June 13, 2020 - link
I doubt he had anything to do with Willow Cove, as that architecture most likely already went into validation when he joined Intel. Maybe he was too late even for Golden Cove, so we will not see his contribution until Ocean Cove.Luminar - Sunday, August 2, 2020 - link
I agreeraywin - Thursday, June 11, 2020 - link
this makes me sad for intelMakaveli - Thursday, June 11, 2020 - link
Don't be sad for a companies worth $270.43B, they are not going anywhere anytime soon.sing_electric - Thursday, June 11, 2020 - link
Neither are their 14nm chips, apparently.Deicidium369 - Thursday, June 11, 2020 - link
Rocket Lake is the LAST 14nm desktop and Cooper Lake (4-8S Xeon) is last 14nm period. Both of those are launching this year. Time to update your little skit.sing_electric - Friday, June 12, 2020 - link
Actually, sounds like the joke's good for at least the rest of the year. ;) TBH, if anything it's a grudging compliment since they've actually shown how solid their 14nm is, and frankly how great their underlying architecture's been.pepone1234 - Friday, June 12, 2020 - link
Every year is the last year 14nm is going to be used. And still, here we are.AndrewJacksonZA - Friday, June 12, 2020 - link
Don't you mean that every year is THE year of Linux on the desktop? ;-)FunBunny2 - Friday, June 12, 2020 - link
"THE year of Linux "it's been that way for years. it's just that you carry your desktop in your pocket.
shabby - Thursday, June 11, 2020 - link
Rip intel, bankruptcy coming soon 😂Deicidium369 - Thursday, June 11, 2020 - link
You spelled AMD wrongpepoluan - Friday, June 12, 2020 - link
Rip intel, AMD coming soon 😂There, FTFY
Hifihedgehog - Thursday, June 11, 2020 - link
On the contrary, I am happy if it means Keller is free of the nonsense and numbskullery he probably had to deal with there on a daily basis. They probably couldn’t handle his free-thinking, maverick style and it was probably for the best of everyone that he parted ways: for him to breath creatively again and for Intel to continue waning. By the looks of Lakefield which I think he had some hand in with its evolutionary package, it just does not have the evolutionary performance it was hyped up to have. Remember that Lakefield is using Sunny Cove which has an 18% IPC improvement over Skylake so Lakefield’s 3 GHz boost is only roughly equivalent to 3.54 GHz on Skylake. So when Intel compared Lakefield to the Core i7-8100Y which has a 4.2 GHz boost clock, they must be speaking of sustained loads in thermally or energy throttled situations and not actual boost performance. This potentially makes it significantly slower than the Core i7-8100Y it was compared to unless its Sunny Cove is heavily modified and suddenly gained more IPC than advertised, which I highly doubt. Keller likes his name associated with success so good for him for walking and leaving this dinosaur corporation to fall into obscurity from their idiotic internal operations.brantron - Thursday, June 11, 2020 - link
Lakefield does not have hyperthreading or AVX, so that 18% IPC is not apples to apples.In the specific case of Lakefield vs. the 8100Y, it's still 4 Tremont cores vs. 2 Skylake cores, though, so YMMV.
Hifihedgehog - Thursday, June 11, 2020 - link
Intel purported a “12% SINGLE threaded” performance uplift for Lakefield over the Core i78500Y. But it does not compute since that 8500Y has a 4.2 GHz boost clock. I really think that number is for a very specific, thermally and energy constrained scenario. Honestly, the more I look at Lakefield, the more I see “over-engineering” and “falling short“ in the design goals.Deicidium369 - Thursday, June 11, 2020 - link
You mean the CPU for the Samsung Note type device? it's not for main line laptops... that's Tiger Lake - the one smoking the AMD 4800Hifihedgehog - Friday, June 12, 2020 - link
Haha. No.DaveLT - Saturday, July 25, 2020 - link
Tiger lake smoking 4800?sure. One exists and one doesn't. Choose which one ;)
Hifihedgehog - Friday, June 12, 2020 - link
I accidentally wrote 8100Y above.sing_electric - Thursday, June 11, 2020 - link
I hope all's well with him and his family, since the fact that he's sticking around as a consultant makes it unlikely that "personal reasons" is being for some of the more common euphemisms: If its because of a disagreement with management on direction, they wouldn't want him as a consultant, if its because he had his eyes set on another opportunity he wouldn't agree to stick around, and if it was for any kind of alleged (or substantiated) misconduct on his part, Intel would have promptly shown him the door (as they did with ex-CEO Brian Krzanich when it came to light that he had a consensual relationship with another employee).sing_electric - Thursday, June 11, 2020 - link
Edit, since I can't read, "being used for some of the common euphemisms."PrimalNaCl - Thursday, June 11, 2020 - link
Or his contract stipulates that he has to remain for some agreed upon turnover period should he wish to leave as a requirement for some final payout.xype - Friday, June 12, 2020 - link
Oooor, "staying on as a consultant" is a nice way of saying "gardening leave". In other words, not allowed to work for any competitor in the 6 months following his departure.Tams80 - Friday, June 12, 2020 - link
A bit like Koduri at AMD.And look where he ended up.
Only in this case the person being put in the garden is a decent person; so the opposite.
I do hope it's not because Keller or one of his relatives are unwell though.
mikato - Sunday, June 14, 2020 - link
"Only in this case the person being put in the garden is a decent person; so the opposite."Haha, nice
PeterCollier - Thursday, June 11, 2020 - link
Does Ian not have spell check and grammar check on his computer?willis936 - Thursday, June 11, 2020 - link
"Shorted" is a real world.Tams80 - Friday, June 12, 2020 - link
Off course it is.Truthy - Thursday, June 11, 2020 - link
I'm going to guess that Jim just ran into to much institutional inertia at Intel and decided that he'd be wasting his time trying to turn the Intel super-tanker in the direction of more design automation. Intel is a very insular culture that doesn't take much to outsiders coming in and telling them what to do. Of course, they desperately need some input from outside to help them, but it's generally not appreciated.Deicidium369 - Thursday, June 11, 2020 - link
Or he got accomplished what he wanted to and wanted to move on - as is his pattern808Hilo - Thursday, June 11, 2020 - link
AMD, Arm, Apple, Google, MS, Nvidia, Tesla all got their boards and chips generations ahead of Intel.Jim Keller arrived at Intel. Old chips, phantasy product pipeline, costly and very limited production, no shrink path. This means that the Intel herd is headless. Was headless for a long time. The herd built social constructs rather than chips corps and consumers want. If we treat corporations like individuals - then we analyze them a partially sentient beeings. Corporate psychiatry. In case of Intel this must follow with treatment. Jim Keller is not what Intel needs. Intel needs to clear the stable. Bring its moohing herd to the slaugtherhouse and grow a new herd. Then Jim Keller has a chance of reform. Because of this I moved our hardware procurement to AMD, Nvidia. Dr. Lisa Su and Jim Keller are exceptional, Elon Musk too. To me Intel is done. Its over. Expensive, greedy, slow innovations. No sell to me, no security - they moved hw design into the warzone Israel. Israel is running a Gulag. I want ethical sourced hardware. Intel cant.
yeeeeman - Friday, June 12, 2020 - link
Funny. Intel has a lot more talented engineers besides Jim. Intel did Sandy bridge without Jim. Intel did optane without Jim. Intel did a lot of innovation without Jim. They needed Jim now because they wanted the best product to come out after this long stagnation period. And it will come. Jim resigned for actual personal reasons, you can read about it in the press, about his family. You're just a conspiracy believer and should quit that. The fact that Intel core from 2015, that is Skylake, competes with Zen 2 parts which are much much newer talks clearly about who has the upper hand on innovation. Sure, Intel has been hand cuffed because of their fab issues, but when it comes to ip, I don't think they have a talent problem.yeeeeman - Friday, June 12, 2020 - link
I agree that they have a management/culture problem and hopefully that will be fixed soon, cause otherwise is a pity for so many talented engineers.willis936 - Friday, June 12, 2020 - link
Does anyone know of any company that had a culture problem that didn’t end in the dismemberment of the company?xype - Friday, June 12, 2020 - link
Intel is publicly traded. Before they go down, you can bet your ass there will be a complete change of management, if need be. They're still doing quite well, though, so there's no real pressure.dromoxen - Sunday, June 14, 2020 - link
Jim keller *is* the Ip-manjoejohnson293 - Friday, June 19, 2020 - link
Intel's fab has a big ongoing cultural problem and is an albatross around Intel's neck. Unfortunately current CTO inherited a trail of cronies in TMG (TD) management chain from the previous one who was unceremoniously "retired". Nothing much has changed in last 2 years for TMG. The list is long.. SVP heading overall PTD, yield VP, 10nm yield manager - who oddly has kept his job/promoted in spite of behavioral issues and consistently poor performance - all Rennsaeler grads btw (some kind of favoritism going on there as well) , litho manager calling shots in 10nm, process integration manager in charge of COAG,.....AshlayW - Friday, June 12, 2020 - link
Let's not forget that Intel's budget is orders of magnitude higher than AMD's. Let's not forget that Skylake doesn't compete with Zen2, not even close, in metrics that matter such as performance per watt, or compute density (cores per socket) / MT performance and I/O (for servers).Your comment is spoken like a clueless gamer who thinks that a few extra % average FPS in a latency-sensitive workload on a dead-end architecture without chiplets and any future is the last word in microrprocessor technology. Let's also not forget the security vulnerabilities.
AMD has innovated more in the processor industry between 2017 and today, than Intel has in the last decade. And you need to accept that.
Gondalf - Friday, June 12, 2020 - link
Between 20217 and now i can see only an Ark----->Zen, Zen 2 is only a small revamping (comparable with Skylake in per core raw performance) and Zen 3 that is is only a better arrangement of the cache stack of Zen 2.More or less have done Intel with Sunny Cove (better SMT), Willow cove (even better SMT), and soon Golden Cove ( nice strong surprise for many) and Ocean Cove (another strong news).
Accept this :)
The only contribute of AMD is to have joined with TSMC ( so kudos to TSMC, definitively NO to AMD ). The main AMD sin is the chiplet desing without a really good interconnection.
schujj07 - Friday, June 12, 2020 - link
"comparable with Skylake in per core raw performance" only when the Skylake CPU has a 10+% clock speed advantage and using a lot more power. We have seen the i7-7700k vs Ryzen 3 3300X review and the 7700k loses far more often than it wins. Both are 4c/8t CPUs but the 7700k has less than a 10% clock speed advantage. That means that Zen 2's IPC advantage over Skylake allows its lower clocks to overtake the Skylake chip.Sunny Cove has a better IPC on paper than Zen 2, however, when running none SPEC applications it seems to falter. Look at the Acer Swift 3 review that has the best Ice Lake i7 and the near best Ryzen 4700U. Yes the 4700U has double the cores and a 100MHz boost clock speed advantage. However, Ice Lake is supposed to have a 7ish% IPC advantage over Zen 2 but loses on many ST applications by more than the 2.5% boost clock disadvantage should allow.
iranterres - Friday, June 12, 2020 - link
I just read jibberish about it chiplet and node and blah blah blah here. What you all need to accept are just two simple facts:- Intel's comfortable market position in the past decade made them seek architecture evolution, not design evolution. Now their game is milked up.
- AMD went through the last decade with a fundamentaly flawed design (bulldozer/vishera) that turned up also as an architecture dead-end.
Now it is just the same history, with sides exchanged.
And yes, fanboys will be fanboys, both sides. You suck.
AshlayW - Friday, June 12, 2020 - link
Don't bring politics into the tech discussion please. Because the "Israel is running a Gulag" thing is subjective and political in nature.Intel's technology inferiority, however, is objective.
willis936 - Friday, June 12, 2020 - link
It’s impossible to separate them. It would be nice and clean if you could, but that isn’t reality.Oxford Guy - Thursday, June 11, 2020 - link
Effective immediately = six months from now.eastcoast_pete - Thursday, June 11, 2020 - link
Most people of Jim Keller's seniority and stature have significant non-compete clauses in their contracts, which often explains why they remain "consultants" for up to and over a year after they either resign or get fired. Basically, one has to "age out" of the inside knowledge before moving on, but that time is paid for.MooseNSquirrel - Friday, June 12, 2020 - link
Non compete is basically non enforceable in California.rrinker - Friday, June 12, 2020 - link
Just because they'll lose in the end doesn't mean a company with deep pockets and lots of corporate lawyers can't make your life a living hell. Non-competes are not enforceable in many places, especially since with specialized jobs it amounts to slave labor if it could be enforced that you can essentially never work in your industry again.eastcoast_pete - Friday, June 12, 2020 - link
There are two main situations with such non-competes: 1. You (or Jim Keller) have another gig lined up, and you or they are willing to cover the legal costs and potential liability 2. You don't, but are paid handsomely for not running to your next gig (as a "consultant"). So, unless he already has a new employer who is willing to cough those legal fees up, basically sitting at home or in your office reading Anandtech and other things and getting paid for it is not a bad gig. Plus, many companies in the Valley will tie the full vesting of stock grants and options to adherence to the non-compete. Lastly, 6 months is harmless compared to some others; can be years. Those are the ones that get challenged in court.MikeMurphy - Thursday, June 11, 2020 - link
Intel has massive problems at the moment having become complacent and uncompetitive after many years of milking a high profitable x86 market. It's mind-boggling to me that supposedly the most brilliant corporate minds let so many opportunities pass by over the last ten years.Hifihedgehog - Thursday, June 11, 2020 - link
> It's mind-boggling to me that supposedly the most brilliant corporate minds let so many opportunities pass by over the last ten years.Greed is one heck of a powerful drug. It will can even blind the most brilliant Ivy League financial whizzes if they don’t constantly check themselves which history has proven time and again. I really think history is going to be repeated itself here again very soon and quite possibly not to Intel’s satisfaction.
drexnx - Thursday, June 11, 2020 - link
disagree they got complacent or let things pass by - they were just beaten in each endeavor outside their core competency:2010-2013 - smartphone/tablet SOCs that went nowhere. Shame, because baytrail was a great chip, just got no wins beyond ones they bought.
2013-2019 - cellular modems that put them in the predicament they've been in with shortages for the past 2 years (because they MUST fab apple modems before they can fab their own stuff. This was supposed to be overhead absorption and a way to keep their n-1 fabs near full utilization. Great idea, but 10nm's failure meant they were stuck with low margin parts and a contractual obligation on their top process)
IoT - 2013-2017ish? - this was probably the most optimistic of their splashes out, how they thought they could make any kind of margin on commodity microcontrollers is beyond me. Maybe this was n---- fab utilization boosts? (I'm not sure how far back they keep fabs, if their 45 or 32nm stuff is even still running)
PandaBear - Wednesday, June 17, 2020 - link
>disagree they got complacent or let things pass by - they were just beaten in each endeavor outside their core competency:They were ramming atom down everyone's throat and they lost. It is obvious they had that coming and it is their fault.
They practically gave the modem away to Apple and they were still not that great (power, signal strength, stability, etc). They just didn't spend enough resources to catch up where Infineon throw in their towel on. They think as long as they have something and if they bundle it with x86 they can sell it.
IoT? You got to be kidding me. Marvell, Qualcomm, and Broadcom would rule that long before Intel would. They actually spend time and resources to get things done while Intel spend its time trying to force an atom down any project. Yup, the same as smart phone and tablet. They don't want to go down the market of $5 SoC so they shove an atom down and charge people $20, or more if they want a celeron.
Deicidium369 - Thursday, June 11, 2020 - link
Yeah "Lisa" Su let a huge opportunity pass her by by not being able to capitalize in a once every 12-15 year opportunity where Intel had issues. She is supposedly brilliant - terrible at contracts and capitalizing on a competitors failures...Avoton - Friday, June 12, 2020 - link
Are... Are you high? Sue joined as CEO of AMD only in 2014, and if you're not aware of her brilliance, go google her. It's easy enough, even for a blind fanboy like you.Tams80 - Friday, June 12, 2020 - link
She was the highest-paid CEO last year precisely because she has been doing an incredible job. And is relatively humble about it too. Oh, and people know her name, unlike whoever is running Intel. That says a lot.And new, innovative, better products can't just be turned out. They need years of development to come to fruitition. Su and AMD don't really have much control over when something will be ready.
PandaBear - Wednesday, June 17, 2020 - link
You cannot just divest from your own fab overnight, Lisa did the unthinkable and turn the ship right. Both AMD's old fabs and AMD are now on the right track after the divorce, instead of following the "real men have fabs" policy from Jerry Sanders.londedoganet - Thursday, June 11, 2020 - link
> Tesla, AMD, Apple, AMD (again), and PA SemiconductorThis is not even in chronological order (though it is nearly, but not quite, reverse chronological order), so I don’t know why AMD merited two mentions.
Lord of the Bored - Thursday, June 11, 2020 - link
Because the first time he did it, we got Sledgehammer, and the second time we got Zen.Deicidium369 - Thursday, June 11, 2020 - link
2 stints at AMD - that's why.Meteor2 - Friday, June 12, 2020 - link
“AMD (twice)” would have been better wording.Quantumz0d - Thursday, June 11, 2020 - link
This is a big one. Intel with their recent "Woke" drivel and the nonsensical M&A instead of their bread and butter x86 industry are =going sideways, I guess Keller hit an immovable rock in his path to make them great again but the corporate PR didn't allow.Highly doubt Intel is going to do anything good. Woke fantasy land of that CA state and it's ridiculous BS nowadays, I think it's irreparable.
I just wish all the best for the Hero of this era of Semiconductor engineering.
PrimalNaCl - Thursday, June 11, 2020 - link
I'm sure Lisa has already asked him if she could have his old office unlocked, cleaned, and the couch pillows fluffed for him when the 6month clock ticks its final tock.Zen 6-10 or even Zen' aren't gonna design themselves.
Kamen Rider Blade - Friday, June 12, 2020 - link
I hope he goes back to AMD, that should be his "Final Home" where he would be "Welcomed".Irata - Friday, June 12, 2020 - link
+1Deicidium369 - Thursday, June 11, 2020 - link
Nonsensical like AI? and buying the AI company with the most impressive product?Yeah that's just crazy.
watzupken - Friday, June 12, 2020 - link
"This is a big one. Intel with their recent "Woke" drivel and the nonsensical M&A instead of their bread and butter x86 industry are =going sideways, I guess Keller hit an immovable rock in his path to make them great again but the corporate PR didn't allow."I don't think you understand how a business work. It is true that the x86 is one of the core businesses, but you can't stick around and not diversify. Perfect example of corporates that do note diversify much, Nokia and Blackberry. They are not chip producers for sure, but if you only survive on what you think is not going to go away, then you are just living in your own world. As it stands, ARM, AMD and Nvidia are chipping away at their stronghold. Corporates are making their own processors, so its a sign that they need to look elsewhere to ensure they don't sink.
WatcherCK - Thursday, June 11, 2020 - link
check out a recent youtube clip from Jim @ AdoredTV about some insights he got about whats its like to to work at Intel, the clip is called "Inside Intel" if the link doesnt paste correctly... https://www.youtube.com/watch?v=agxSclh27uoazfacea - Thursday, June 11, 2020 - link
if i am amazon right now, i am thinking "what a pity"supdawgwtfd - Friday, June 12, 2020 - link
"silicon for self driving which analysts have Tesla’s competitors have said put the company up to seven years ahead."Uh. What?
You know an easy way to figure out if what you have written makes sense?
Try to read it out loud.
Basic proof reading 101 which this site still can't even grasp.
Ryan Smith - Friday, June 12, 2020 - link
It was written on the fly as breaking news. Unfortunately that means we have to rush a bit to get things up in a timely manner, and clean it up later.Zizy - Friday, June 12, 2020 - link
Considering Intel has all the replacements ready this isn't really "whoa out of the blue" departure like Intel's CEO recently. There might be something bad that forced his leave, or his work is done and he is retiring. Or maybe moving on to another company after a break. I guess IBM to finish his tour of the relevant US semi companies.realrealtruth - Friday, June 12, 2020 - link
This website is written by amateurs doctors that are trying to get a piece of pie of audience with bankrupt articles..WaWaThreeFIVbroS - Friday, June 12, 2020 - link
Well i bet Dr. Ian Cutress isn't more amateur than username "realrealtruth" on the internetdevione - Friday, June 12, 2020 - link
Don't forget to ask for a refund before you leaveNoodle-Naut - Friday, June 12, 2020 - link
He quits when he thinks he can do more somewhere else. He needs to start his own company before he is too old.alufan - Friday, June 12, 2020 - link
makes you wonder if like many brilliant folks he may be on the edge of the autistic spectrum and just needs a specific challenge to keep him motivated, seems his pattern at least for an outsider looking invision33r - Friday, June 12, 2020 - link
Intels problem is mostly manufacturing and capacity issue. Because they are greedy. They need 1 chip stamped 100 different ways to milk the skus.Nictron - Saturday, June 13, 2020 - link
Good luck to him and hope the personal matters are not too serious.From my perspective Intel does not have a design or performance issue but a issue of manufacturing and security.
Manufacturing could be a shareholder issue as they are not willing to invest in the capital expenditure to move to lower nm fabs as they “feel” they still have the performance advantage. It could be a “ego” thing where they just do not see the need to invest.
Another advantage Intel has is in the business enterprise space where their Microsoft partnership is benefitting them a lot. It is really expensive to by software especially database software for AMD. Their core count actually counts against them and per core performance is better in that space which makes it more economical to go Intel, however the power efficiency lag is starting to catch up with Intel.
The other area of concern for Intel is the security space as they just seem to be hit with issue after issue as we go along with not end in sight. Their predictive performance algorithms that gives them that IPC advantage seems to be prone to security exploits and this could be causing some serious conflicts inside Intel. Address the security and loose IPC performance crown?
From my perspective AMD is smoking at the moment. If Intel can however address their fab and security soon they will stay on top for the future.
stockolicious - Saturday, June 13, 2020 - link
The AMD / INTC dynamic is different this time as INTC doesn't control or dominate in MFG. INTC has historically just killed AMD due to their MFG scale and performance. If AMD had a good design INTC would create design and improve MFG process well ahead. AMD was right to spin off Glofo and now they could be competitive or more in the next 10 years so I don't think INTC will just come back and dominate like before. I see a more evenly distributed market as the next few years unfold.PandaBear - Wednesday, June 17, 2020 - link
The problem is, AMD share the lower nm cost with Apple, Qualcomm, and others while Intel is doing it all by themselves. There is a different economy of scale and AMD is now able to hitch hike with Apple and Qualcomm to beat Intel. This is new and has never really happened before, when foundry was always behind Intel for one or two generation. Now Apple and Qualcomm will pay for this due to power consumption reason rather than cost performance reason.Intel didn't have the wireless dominance to justify the lower nm cost, so they are now losing the edge to AMD. If you think AMD catching up is bad, wait till Apple starts switching their mac to their own CPU on ARM, that could really get Intel out of consumer space over time, starting with chrome book then gradually to gaming laptop.
joejohnson293 - Friday, July 17, 2020 - link
Intel has had an exodus of talent in the last 1 or 2 years, The CEO needs to sell off its fabs and manufacturing unit to the highest bidder (Gloflo or Samsung?) ASAP and buy one of ARM/Nvidia/AMD/Qualcomm with the proceeds for Intel to stay relevant.FPGAs4Life - Tuesday, July 28, 2020 - link
Does anyone have any predictions on where Jim Keller might end up post Intel or has anyone heard any rumors on the topic?alumine - Wednesday, July 29, 2020 - link
It'd be a big kick in the guts for Intel if he moved to Apple to do their new silicon....