Question What happened with Intel?

[Endre]

Hello!

I’m on an Intel platform and for the majority of my life I’ve been an Intel fanboy.
But what happened with Intel?

AMD is boasting about launching CPUs with 16-32-64 cores, being run smartly by Lisa Su, while Intel continues to release 14nm CPUs with 8-10 cores that won’t be able to compete against new CPUs from AMD!
Intel 10th gen seems to be a failour right from the beginning!
Intel is no longer relevant!
Intel seems to be run by weak minded leaders (to say the least).

What are your thoughts on this matter?

 

[Endre]

Think is.Intel has brand loyalty fans doesn't matter what s@#t they sell people will buy it.Intel charge you a premium for Hyperthreading, 5ghz etc etc.Where amd all cpu's are unlocked.Intel actually made people believe you can't have more than 4c on a desktop cpu.Ryzen beg them differ.On a neutral point i wasn't impressed with the 3xxx models cept for the 3950x and the new threadripper as i see no need to upgrade from a r5 2600 to an r5 3600 in 1440p.Like i say the wheels have turned for all there scaly tactics some years ago.I won't count them out But...They losing to a company at the moment that has 5% of there R&D budget

Not all is bad as a 8700k are still a solid performer(i hate the 9900k and ks)not worth the price then you might as well get a 3950x.Even on there budget options you can get a 9100f and a 9400f for next to nothing these days.But again i will add a few bucks and get a r5 2600 which has HT.So intel only has themselves to blame for releasing products with high prices.Intel even said amd's not real world performance.They actually measured performance on MS excvel.Like really.....Intels hurt and they are desperate and they will feed people s@#t and people will eat it.....That's just my point of view.

Intel must come back “in a big way”!
The 10th gen desktop CPUs won’t do that!
We know that they’ve been working on different products: discrete graphics cards, optane memory etc.
But their main products are CPUs, and they neglected them!

We’ve seen other giants falling down in the past because of bad management:
Nokia, Polaroid, Kodak, Akai etc.
I hope that doesn’t happen to Intel.

 

[drea.drechsler]

...
Nokia, Polaroid, Kodak, Akai etc.
I hope that doesn’t happen to Intel.

At least two of those four have markedly different stories than Intel's. Polaroid and Kodak were, and still are, both inseparably linked with chemistry based photography in the consumer mind. So much so that even their involvement in other technologies couldn't help as all anyone thought of them as was film photography, so hard to get attention in new areas. Not saying they didn't make mistakes, but it's hard to just manage your way out of a perception problem like that.

Intel's management problem is one that's pretty logical, very different, and from an investors perspective at the time very forgiveable. Why take on more risk with more heavy investments when you're already failing at one major advancement (to 10nm) and your only possible competitor is so obviously on the ropes? Answer: you don't. You pay back to the investors, that's what you do.

Bad call, especially with 5 year design/dev cycles meaning you'll have to be clever with what you got until then. But did Intel really know what AMD had up it's sleeve in 2016, or especially 2015?

But it's pointed out in several articles I've read: AMD does not have any way to swap positions with Intel's market dominance, i.e. turn 80% Intel/20% AMD to 80% AMD/20% Intel, even in 5 years. There simply is not sufficient 7nm wafer fab capacity in the world to do that. Even if TSMC were to increase capacity to enable that feat it would take 5 years to build and qualify the fabs...and what of THEIR risks if AMD should fail? Wafer agreements are only as good as the companies' bottom lines.

At any rate, in the time it takes to build out sufficient capacity, Intel will have new products and then we see what we got.

 

[USAFRet]

At least two of those four have markedly different stories than Intel's. Polaroid and Kodak were, and still are, both inseparably linked with chemistry based photography in the consumer mind. So much so that even their involvement in other technologies couldn't help as all anyone thought of them as was film photography, so hard to get attention in new areas. Not saying they didn't make mistakes, but it's hard to just manage your way out of a perception problem like that.

Kodak screwed themselves.
They invented the digital camera. Literally.
But, "people will always use film, so that digital thing is just a curiosity." (or words to that effect).

Canon, Nikon, Fuji...all were also heavily film based. It's called Fujifilm for a reason. They managed to change direction.

 

[Endre]

Kodak screwed themselves.
They invented the digital camera. Literally.
But, "people will always use film, so that digital thing is just a curiosity." (or words to that effect).

Canon, Nikon, Fuji...all were also heavily film based. It's called Fujifilm for a reason. They managed to change direction.

Think is.Intel has brand loyalty fans doesn't matter what s@#t they sell people will buy it.Intel charge you a premium for Hyperthreading, 5ghz etc etc.Where amd all cpu's are unlocked.Intel actually made people believe you can't have more than 4c on a desktop cpu.Ryzen beg them differ.On a neutral point i wasn't impressed with the 3xxx models cept for the 3950x and the new threadripper as i see no need to upgrade from a r5 2600 to an r5 3600 in 1440p.Like i say the wheels have turned for all there scaly tactics some years ago.I won't count them out But...They losing to a company at the moment that has 5% of there R&D budget

Not all is bad as a 8700k are still a solid performer(i hate the 9900k and ks)not worth the price then you might as well get a 3950x.Even on there budget options you can get a 9100f and a 9400f for next to nothing these days.But again i will add a few bucks and get a r5 2600 which has HT.So intel only has themselves to blame for releasing products with high prices.Intel even said amd's not real world performance.They actually measured performance on MS excvel.Like really.....Intels hurt and they are desperate and they will feed people s@#t and people will eat it.....That's just my point of view.

Yes. Every brand is as good as its last lineup of products is!
If a brand doesn’t innovate and loses market share, it’ll go extinct.

Maybe Intel needs a CEO like Lisa Su?

 

[USAFRet]

Kodak got into consumer digital far too late. And poorly implemented.
Yes, it might have required a different brand name. But that could have easily been done.

Canon/Nikon/Fuji have managed the conversion. Kodak didn't.

Just like Sears. They could have been Amazon. Indeed, they were that big. They had an absolute lock on the mail order channel. You could buy everything to fill a house, and the house to put it in.
They dropped the ball.

Oh well.

I don't see Intel going down that same fail road.

 

[Zizo007]

Hello!

I’m on an Intel platform and for the majority of my life I’ve been an Intel fanboy.
But what happened with Intel?

AMD is boasting about launching CPUs with 16-32-64 cores, being run smartly by Lisa Su, while Intel continues to release 14nm CPUs with 8-10 cores that won’t be able to compete against new CPUs from AMD!
Intel 10th gen seems to be a failour right from the beginning!
Intel is no longer relevant!
Intel seems to be run by weak minded leaders (to say the least).

What are your thoughts on this matter?

Its because Intel is having issues with 10nm on desktop, they overheat and cannot clock high and yields are very low. AMD are on 7nm without any issue and are already preparing 5nm.

Also Intel is behind AMD as AMD has PCIE 4.0 but not Intel.

Intel are launching GPUs now so they might improve. It will take Intel a year or two to catch up AMD.

 

[drea.drechsler]

...Just like Sears....

Sears/KMart has been taken apart by a CEO with quite purposeful intent. Looked at in the right way it's not 'failed' or bad management at all. In fact, it could be said that considering how much it's made him (and shareholders) it's been a major success story. It's literally a story book case of how to keep a torture victim alive in order to prolong the agony and continue getting sick thrills. Hannibal Lecter would be impressed.

The one I'm not sure of is Walmart. Why they can't leverage their distribution network to challenge Amazon is beyond me. But so long as their brick and mortar stores are doing well they'll continue trudging on I suppose.

 

[Endre]

Its because Intel is having issues with 10nm on desktop, they overheat and cannot clock high and yields are very low. AMD are on 7nm without any issue and are already preparing 5nm.

Also Intel is behind AMD as AMD has PCIE 4.0 but not Intel.

Intel are launching GPUs now so they might improve. It will take Intel a year or two to catch up AMD.

Yeah! But that’s crazy! Because Intel as a company is hundreds of times larger than AMD!
How is it possible that the richer company didn’t solve the nm issue, while a “bug” like AMD did?

Maybe their CEO left the company for a 5 year holiday trip? 🤣

 

[InvalidError]

How is it possible that the richer company didn’t solve the issue nm issue, while a “bug” like AMD did?

AMD didn't solve anything nm-wise as it is fab-less, most of the credit for that goes to ASML who makes the EUV equipment used by Samsung and TSMC.

Intel wanted to do 10nm on DUV years ago because EUV didn't exist at the time and now Intel is too deeply invested into 10nm DUV to just give up on it just as it has finally managed to overcome most of its issues.

As for why Intel isn't skipping to 7nm, that is simple: Intel would have to order EUV equipment from ASML too, which would put it behind Samsung, TSMC and others' existing orders, meaning Intel isn't going to have enough EUV equipment for mass-production any time soon and still needs 10nm to get up to speed in the meantime.

 

[Endre]

AMD didn't solve anything nm-wise as it is fab-less, most of the credit for that goes to ASML who makes the EUV equipment used by Samsung and TSMC.

Intel wanted to do 10nm on DUV years ago because EUV didn't exist at the time and now Intel is too deeply invested into 10nm DUV to just give up on it just as it has finally managed to overcome most of its issues.

As for why Intel isn't skipping to 7nm, that is simple: Intel would have to order EUV equipment from ASML too, which would put it behind Samsung, TSMC and others' existing orders, meaning Intel isn't going to have enough EUV equipment for mass-production any time soon and still needs 10nm to get up to speed in the meantime.

Yeah! But don’t you think that Intel is being run in a wrong way?
They should’ve figured it out years ago.

 

[InvalidError]

They should’ve figured it out years ago.

Figure what out years ago? EUV didn't exist, so pushing DUV further was the only way to stay ahead. If you mean figuring out EUV, the fundamental research that goes into lithography equipment has been delegated by pretty much everyone in the chip fab business including Intel to ASML because no single chip manufacturer can afford to do the science required to build those machines from scratch at their own expense.

 

[NightHawkRMX]

Im no expert, but here's what i saw happen.

Intel didn't have much in the way of competition from AMD for a span of around 10 year ranging from roughly 2007 to 2017. In this time AMD was releasing underwhelming performance phenoms and FX CPUs that just couldn't compete with Intel in any meaningful performance metric.

Intel knee they had little competition and people had no choice but to buy their chips, so they just chose not to spend much resources into research and development. Intel just kept refreshing providing noticable but not major improvements generation to generation.


Once AMD hit them with Ryzen on 2017 and especially with Ryzen 3xxx in 2019, Intel's many years of being lazy caught up to them and intel didn't have any new architectures or node shrinks upcoming to compete.

That leads Intel to where they are today.

Intel's 10nm node is not ready for desktop soon, so Intel currently is milking everything possible out of their 14nm coffee lake architecture.

The 14nm 9900ks has 8 cores plus ht and hits 5ghz on all cores. It really delivers on performance. However, the chip uses nearly 250w of power and requires a massive cooler.
If intel were to release a new CPU with tuned up clocks and or add more cores, the chips would be unsuitable for desktop system due to the insane power consumption and cooler requirements.

So basically, untill intel gets 10nm out, they cannot do a lot to get a chip with much more performance.

Meanwhile while Intel is standing around, AMD just keeps whacking them, even now with zen 2 laptop CPUs.

TLDR: Intel got lazy and greedy.

 

[drea.drechsler]

... Intel's many years of being lazy caught up to them and intel didn't have any new architectures or node shrinks upcoming to compete.
...
TLDR: Intel got lazy and greedy.

I really can't agree with the lazy part.

To be sure, the board got greedy and kept on milking the cash cow because what's the risk in that, especially with AMD mired in FX and apparently focused on shoring up their gaming console market. I believe a lot of scientists and engineers at Intel were working fast and furious with many late night reviews explaining why 10nm wasn't working and formulating plans on top of plans to get it back on track. I'm sure they had themselves convinced the cure was just around the corner, but with no competitive threat and few good reasons to stop but a lot of better ones to continue.

I don't see lazy in that, just not understanding what your competition was really up to.

 

[NightHawkRMX]

I'd agree with everything except the lazy part.

To be sure, the board got greedy and kept on milking the cash cow because what's the risk in that, especially with AMD mired in FX and apparently focused on shoring up their gaming console market. I believe a lot of scientists and engineers at Intel were working fast and furious with many late night reviews explaining why 10nm wasn't working and formulating plans on top of plans to get it back on track. I'm sure they had themselves convinced the cure was just around the corner, but with no competitive threat there were few good reasons to stop and a lot of better ones to continue.

I don't see lazy in that, just not understanding what your competition was really up to.

I have a feeling until threatened by Ryzen, Intel wasn't exactly rushing to shrink the node.

 

[drea.drechsler]

I have a feeling until threatened by Ryzen, Intel wasn't exactly rushing to shrink the node.

That brings up a question: what comes first? the processing node or the architecture if you want it solidly optimized for the node? I know people have said you can scale any arch to any node, but I'm talking about fully optimizing the whole process.

I have in mind what AMD's done with 7nm and modular approach (chiplets) that maximize wafer yields while leaving the I/O section on 14nm (12?). Even with it's inherent performance inefficiencies the approach seems truly innovative to me for it's manufacturing efficiencies and cost control.

EDIT add: and oh yeah. I have no reason to disagree that Intel wasn't working on a node shrink, or much of anything else, until Ryzen came along. But not only that, they probably didn't get serious until 2nd Gen nor worry before Zen2. But put yourself in position of an Intel Board member, the board would have to approve any really major development project like new arch dev't and especially processing node dev't effort. When some silly guys come in and propose a $5billion (just a guess) effort to do that they have to present a business case. Well...marketing projections are that you can continue selling all the current lines of processors you can sell at any price point you really want. Just release them with new numbers and call it a refresh. So why do we need this? To maintain leadership? Sorry, such nice marketing doesn't work for them.

Next question: what's my Risk in this, Mr. Engineer?
Answer: Of course, we're Intel Engineering...why none, of course.
Response: BULLSH*T, we just had your friends crying about the 10nm debacles! Point being: in their mind, risk is VERY HIGH.

I don't think they'd get their money, the board has a responsibility to the shareholder and they see a low-risk way to make them very happy for quite a while.

I'd put Intel's current problems at the feet of the Board of Directors.

 

[USAFRet]

That brings up a question: what comes first? the processing node or the architecture if you want it solidly optimized for the node?

Both at the same time.
Shrinking was/is always happening.

"Let's design and build to the next size down. Be that 22, 14, 10, 7, 5, whatever."

 

[TerryLaze]

Hello!

I’m on an Intel platform and for the majority of my life I’ve been an Intel fanboy.
But what happened with Intel?

AMD is boasting about launching CPUs with 16-32-64 cores, being run smartly by Lisa Su, while Intel continues to release 14nm CPUs with 8-10 cores that won’t be able to compete against new CPUs from AMD!
Intel 10th gen seems to be a failour right from the beginning!
Intel is no longer relevant!
Intel seems to be run by weak minded leaders (to say the least).

What are your thoughts on this matter?

Intel has optane ram DIMMS
https://www.tomshardware.com/news/intel-optane-dimm-pricing-performance,39007.html
They have laptop CPUs on M.2 slots
https://www.tomshardware.com/news/i...network-processor-ai-inference-m.2,40204.html
They also put them on PCI slots
https://www.anandtech.com/show/1517...-intel-nervana-nnpt-servers-8way-pcie-and-oam
Now they teased a whole PC on a PCI slot
https://www.anandtech.com/show/1495...-new-element-brings-project-christine-to-life

The age of having a CPU with a bunch of cores and being locked to that is pretty much over, intel is building the future if your CPU hits it's limits you will be able to just stick more cores in it in a variety of ways.

 

[InvalidError]

That brings up a question: what comes first? the processing node or the architecture

You need both.

New architectures pack more transistors into most functions (that's how IPC improves) and you need a smaller faster process to afford those extra transistors without blowing up the timing, area and power budgets. If you don't have a more advanced process, then you either can't produce a significantly better architecture or will have to lower clock frequencies.

 

[drea.drechsler]

.....
The age of having a CPU with a bunch of cores and being locked to that is pretty much over, intel is building the future if your CPU hits it's limits you will be able to just stick more cores in it in a variety of ways.

Indeed, Intel's hardly on the ropes. Intel will still be outselling AMD in 5 years, no doubt, even if 10nm's a complete bust and nothing follows. All they have to do is keep refreshing and lowering prices. Or, rather, selling more cores/threads at the same price to say they didn't. For one thing it's probably impossible for AMD to get enough wafers to challenge Intel's position. Remember also that AMD will have 7nm gaming consoles to equip from their allotments.

Here's a fascinating and informative video on the subject from AdoredTV. Really well supported analysis.

I also wonder how the building-block chiplet approach AMD's used with Zen2 would fit in with what Intel's presenting. At first blush, it appears a shoo-in for them.

 

[bioax]

In my opinion, AMD provides CPUs that are a bang for you buck. They have almost identical performance, but are a lot cheaper.

 

[NightHawkRMX]

Indeed, Intel's hardly on the ropes. Intel will still be outselling AMD in 5 years, no doubt, even if 10nm's a complete bust and nothing follows. All they have to do is keep refreshing and lowering prices. Or, rather, selling more cores/threads at the same price to say they didn't. For one thing it's probably impossible for AMD to get enough wafers to challenge Intel's position. Remember also that AMD will have 7nm gaming consoles to equip from their allotments.

Here's a fascinating and informative video on the subject from AdoredTV. Really well supported analysis.

I also wonder how the building-block chiplet approach AMD's used with Zen2 would fit in with what Intel's presenting.

I agree. I have heard my fare share of "AMD is better than intel hands down" and "Intel is doomed for bankruptcy" comments.

Intel has very little to worry about in the immediate future, although if they stagnate while AMD continues to innovate, they will have issues later on.

Intel had a net income of 21 BILLION dollars in 2018, but AMD only had around 6 Billion in the same year. Remember this was 2018, before Zen 2 desktop or laptop CPUs.

For enthusiast sales, AMD does take a very high percentage of Intels marketshare, however laptops and prebuilts are the real money makers.

Intel still creams AMD in prebuilt sales, probably due to most buyers purchasing Intel due to habit, and thus OEMs don't make many Ryzen PCs since they just dont sell well.

Intel flatout has had better laptop CPUs for as long as I can remember. However, AMD did just unveil new Zen 2 based Ryzen mobile CPUs at CES 2020. I feel these will be genuinely fantastic competitors to Intel's laptop skews and will nibble at Intels share in the laptop market, but AMD will have a hard time gaining a real foothold in the laptop market.

 

[Zizo007]

I agree. I have heard my fare share of "AMD is better than intel hands down" and "Intel is doomed for bankruptcy" comments.

Intel has very little to worry about in the immediate future, although if they stagnate while AMD continues to innovate, they will have issues later on.

Intel had a net income of 21 BILLION dollars in 2018, but AMD only had around 6 Billion in the same year. Remember this was 2018, before Zen 2 desktop or laptop CPUs.

For enthusiast sales, AMD does take a very high percentage of Intels marketshare, however laptops and prebuilts are the real money makers.

Intel still creams AMD in prebuilt sales, probably due to most buyers purchasing Intel due to habit, and thus OEMs don't make many Ryzen PCs since they just dont sell well.

Intel flatout has had better laptop CPUs for as long as I can remember. However, AMD did just unveil new laptop Zen 2 mobile CPUs at CES 2020. I feel these will be genuinely fantastic competitors to Intels laptop skews and will nibble at Intels share in the laptop market, but Intel will still have a massive lead there for long time to come.

AMD are currently taking the lead from Intel in the mobile segment too.
They have an 8 core 15W Zen2 mobile CPU and a 35W 8 core with the same clocks as the 4800H. They can easily make a 4900H with faster clocks at 45W.

The 4800H is already 5% faster in ST loads and 46% faster in MT loads than an i7 9750H. A 4900H will be a lot faster than an i7 9750H. Intel is struggling to get an 8 core mobile CPU as their 6 core already run pretty hot and never reach their turbo speed of 4.5Ghz, many users report a max speed of 4Ghz as temps reach 98C under load.

Also, the 4800H will be much cheaper than Intel. AMD in CES showed a 4800H/5600M laptop for 799$ which is a crazy deal.

Edit: 4800H vs i7 9750H https://www.notebookcheck.net/AMD-Ryzen-7-4800HS-Laptop-Processor-Benchmarks-and-Specs.449912.0.html

 

[NightHawkRMX]

AMD are currently taking the lead from Intel in the mobile segment too.
They have an 8 core 15W Zen2 mobile CPU and a 35W 8 core with the same clocks as the 4800H. They can easily make a 4900H with faster clocks at 45W.

The 4800H is already 5% faster in ST loads and 46% faster in MT loads than an i7 9750H. A 4900H will be a lot faster than an i7 9750H. Intel is struggling to get an 8 core mobile CPU as their 6 core already run pretty hot and never reach their turbo speed of 4.5Ghz, many users report a max speed of 4Ghz as temps reach 98C under load.

They are taking the performance lead.

AMD already has the performance lead in desktop CPUs (in general) but yet Intel prebuilts outsell AMD.
So even if AMD has the performance lead in laptop CPUs, I still feel Intel laptops will outsell AMD for awhile yet.

 

[Zizo007]

They are taking the performance lead.

AMD already has the performance lead in desktop CPUs (in general) but yet Intel prebuilts outsell AMD.
So even if AMD has the performance lead in laptop CPUs, I still feel Intel laptops will outsell AMD for awhile yet.

Exactly, Apple and Dell will always favor Intel and they will never switch to AMD. Lots of people buy Apple PCs and laptops; Apple only uses Intel. That's a market that AMD will never be able to get even if they have the performance crown.

 

[NightHawkRMX]

Exactly, Apple and Dell will always favor Intel and they will never switch to AMD. Lots of people buy Apple PCs and laptops; Apple only uses Intel. That's a market that AMD will never be able to get even if they have the performance crown.

Are you being sarcastic or agreeing with me? hard to tell lol.

Apple does have a long history of using Intel CPUs instead of AMD, but they also have a history of using AMD GPUs.

Dell has plenty of Ryzen machines. They even released Ryzen 5 1400 prebuilts right when ryzen came out like this one LTT reviewed.

An ALL-AMD Gaming PC from DELL??

Buy Dell Inspiron Gaming 5675 on Amazon: http://geni.us/XfBtit Pre-built desktops - they've always been less than ideal, but can AMD's comeback bring them back to relative relevance?

linustechtips.com