Intel's Humbling • Hacker News Board

Re: Intel's Humbling

#301

Post by formerly_proven » Wed, Jan 31, 2024, 1:21 PM UTC

adrian_b wrote: post #276

tgtweak wrote: post #174

I'm kind of bullish on Intel right now. They've moved up so many process nodes so quickly and have made some earnest headway in being an actual fab. Let's ignore the elephant in the room which is taiwan and it's sovereignty, and only focus on the core r&d. Intel flopped so hard on process nodes for 4 years up until Gelsinger took the reigns... it was honestly unprecedented levels of R&D failure. What happened over th…

There is a good chance for Intel to recover, but that remains to be proven. From their long pipeline of future CMOS manufacturing processes with which Intel hopes to close the performance gap between them and TSMC, for now there exists a single commercial product: Meteor Lake, which consists mostly of chips made by TSMC, with one single Intel 4 die, the CPU tile. The Meteor Lake CPU seems to have finally reached the…

> TSMC 5-nm process of almost 4 years ago

N5 is interesting because it's the first process fully designed around EUV and because it was pretty much exclusive to Apple for almost two years. It launched in Apple products in late 2020, then crickets until about late 2022 (Zen 4, RTX 4000, Radeon 7000). Launches of the other vendors were still on N7 or older processes in 2020 - RTX 3000 for example used some 10nm Samsung process in late 2020. All of those were DUV (including Intel 7 / 10ESF). That's the step change we are looking at.

Re: Intel's Humbling

#302

Post by whynotminot » Wed, Jan 31, 2024, 1:29 PM UTC

timschmidt wrote: post #166

whynotminot wrote: post #158

Earlier quoted context omitted.

I think I have a better sense of some of the misunderstanding here. I think you're taking Intel's TDP figures as honest and straightforward, when they really truly aren't. Anandtech has talked about some of Intel's TDP shenanigans before, even for Tiger Lake: https://www.anandtech.com/show/16084/intel-tiger-lake-review...

I wouldn't call them shenanigans. At least Intel publishes numbers (Apple does not). There are simply a lot of power states Intel CPUs are capable of engaging, and many configuration options about when and under what circumstances to do so. Much of this is configurable by the OEM and is not hard-set by Intel, or even under Intel's control after the processor leaves the fab. The Anandtech article seems to indicate tha…

So you’re aware of this Intel TDP chicanery but trying to say with a straight face that these chips achieve similar performance and efficiency to the M1. I think that’s pretty disingenuous.

Re: Intel's Humbling

#303

Post by ajross » Wed, Jan 31, 2024, 1:32 PM UTC

adrian_b wrote: post #276

tgtweak wrote: post #174

I'm kind of bullish on Intel right now. They've moved up so many process nodes so quickly and have made some earnest headway in being an actual fab. Let's ignore the elephant in the room which is taiwan and it's sovereignty, and only focus on the core r&d. Intel flopped so hard on process nodes for 4 years up until Gelsinger took the reigns... it was honestly unprecedented levels of R&D failure. What happened over th…

There is a good chance for Intel to recover, but that remains to be proven. From their long pipeline of future CMOS manufacturing processes with which Intel hopes to close the performance gap between them and TSMC, for now there exists a single commercial product: Meteor Lake, which consists mostly of chips made by TSMC, with one single Intel 4 die, the CPU tile. The Meteor Lake CPU seems to have finally reached the…

> The Meteor Lake CPU [...] has obvious difficulties in reaching high clock frequencies,

Not sure where that's coming from? The released parts are mobile chips, and the fastest is a 45W TDP unit that boosts at 5.1GHz. AMD's fastest part in that power range (8945HS) reaches 5.2GHz. Apple seems to do just fine at 4GHz with the M3.

I'm guessing you're looking at some numbers for socketed chips with liquid cooling?

Re: Intel's Humbling

#304

Post by ajross » Wed, Jan 31, 2024, 1:43 PM UTC

jillesvangurp wrote: post #281

tgtweak wrote: post #174

I'm kind of bullish on Intel right now. They've moved up so many process nodes so quickly and have made some earnest headway in being an actual fab. Let's ignore the elephant in the room which is taiwan and it's sovereignty, and only focus on the core r&d. Intel flopped so hard on process nodes for 4 years up until Gelsinger took the reigns... it was honestly unprecedented levels of R&D failure. What happened over th…

There are a few areas where they are under pressure: - The wintel monopoly is losing its relevance now that ARM chips are creeping into the windows laptop market and now that Apple has proven that ARM is fantastic for low power & high performance solutions. Nobody cares about x86 that much any more. It's lost its shine as the "fastest" thing available. - AI & GPU market is where the action is and Intel is a no-show f…

> Apple has proven that ARM is fantastic for low power & high performance solutions

Apple has proven that Apple Silicon on TSMC's best process is great. There are no other ARM vendors competing well in that space yet. SOCs that need to compete with Intel and AMD on the same nodes are still stuck at the low margin end of the market.

Re: Intel's Humbling

#305

Post by Macha » Wed, Jan 31, 2024, 1:49 PM UTC

bmer wrote: post #254

llamaLord wrote: post #187

Earlier quoted context omitted.

To be fair, they should have seen Ryzen coming, any long-term AMD user knew years before Ryzen landed that it was going to be a good core because AMD were very vocal about how badly wrong they bet with Bulldozer (previous core family). AMD bet BIG on the software industry leaning in heavily on massive thread-counts over high throughput, single-threaded usage... But it never happened so the cores tanked. It was never…

(I'm curious about this story, as I am unfamiliar with it.) Why did that generation of core (Bulldozer) suck? What was it that AMD needed to do to fix the problem? (Links to relevant stories would be sufficient for me!)

Chips and Cheese has probably the most in depth publicly available dive for the tech reasons why Bulldozer was the way it was:

https://chipsandcheese.com/2023/01/22/bulldozer-amds-crash-m...

https://chipsandcheese.com/2023/01/24/bulldozer-amds-crash-m...

---

From a consumer perspective, Bulldozer and revisions as compared to Skylake and revisions were:

+ comparable on highly multi-threaded loads

+ cheaper

- significantly behind on less multi-threaded loads

- had 1 set of FPUs per 2 cores, so workloads with lots of floating point calculations were also weaker

- Most intensive consumer software was single or a very small number of thread focused still (this was also a problem for Intel in trying to get people to buy more expensive i7s/i9s over i5s in those days)

Re: Intel's Humbling

#306

Post by bonton89 » Wed, Jan 31, 2024, 1:56 PM UTC

bmer wrote: post #254

llamaLord wrote: post #187

Earlier quoted context omitted.

To be fair, they should have seen Ryzen coming, any long-term AMD user knew years before Ryzen landed that it was going to be a good core because AMD were very vocal about how badly wrong they bet with Bulldozer (previous core family). AMD bet BIG on the software industry leaning in heavily on massive thread-counts over high throughput, single-threaded usage... But it never happened so the cores tanked. It was never…

(I'm curious about this story, as I am unfamiliar with it.) Why did that generation of core (Bulldozer) suck? What was it that AMD needed to do to fix the problem? (Links to relevant stories would be sufficient for me!)

There's been lots written about this but this is my opinion.

Bulldozer seemed to be designed under the assumption heavy floating point work would be done on the GPU (APU) which all early construction cores had built in. But no one is going to rewrite all of their software to take advantage of the iGPU that isn't present in existing CPUs and isn't present in the majority of CPUs (Intel) so it sort of smelt like Intel's itantic moment, only worse.

I think they were desperate to see some near term return on the money they spent on buying ATI. ATI wasn't a bad idea for a purchase but they seemed to heavily overpay for it which probably really clouded management's judgement.

Re: Intel's Humbling

#307

Post by adrian_b » Wed, Jan 31, 2024, 1:59 PM UTC

ajross wrote: post #303

adrian_b wrote: post #276

Earlier quoted context omitted.

There is a good chance for Intel to recover, but that remains to be proven. From their long pipeline of future CMOS manufacturing processes with which Intel hopes to close the performance gap between them and TSMC, for now there exists a single commercial product: Meteor Lake, which consists mostly of chips made by TSMC, with one single Intel 4 die, the CPU tile. The Meteor Lake CPU seems to have finally reached the…

> The Meteor Lake CPU [...] has obvious difficulties in reaching high clock frequencies, Not sure where that's coming from? The released parts are mobile chips, and the fastest is a 45W TDP unit that boosts at 5.1GHz. AMD's fastest part in that power range (8945HS) reaches 5.2GHz. Apple seems to do just fine at 4GHz with the M3. I'm guessing you're looking at some numbers for socketed chips with liquid cooling?

The 5.1 GHz Intel Core Ultra 9 processor 185H is the replacement for the 5.4 GHz Intel Core i9-13900H Processor of previous year. Both are 45-W CPUs with big integrated GPUs and almost identical features in the SoC.

No liquid cooling needed for either of them, just standard 14" or 15" laptops without special cooling, or NUC-like small cases, because they do not need discrete GPUs.

Both CPUs have the same microarchitecture of the big cores.

If Intel had been able to match the clock frequencies of their previous generation, they would have done that, because it is embarrassing that Meteor Lake wins only the multi-threaded benchmarks, due to the improved energy efficiency, but loses in the single-threaded benchmarks, due to lower turbo clock frequency, when compared to the last year's products.

Moreover, Intel could easily have launched a Raptor Lake Refresh variant of i9-13900H, with a clock frequency increased to 5.6 GHz. They have not done this only to avoid an internal competition for Meteor Lake, so they have launched only HX models of Raptor Lake Refresh, which do not compete directly with Meteor Lake (because they need a discrete GPU).

During the last decade, the products made at TSMC with successive generations of their processes had a continuous increase of their clock frequencies.

On the other hand Intel had a drop in clock frequency at all switches in the manufacturing processes, at 14-nm with the first Broadwell models, then at 10-nm with Cannon Lake and Ice Lake (and even Tiger Lake could not reach clock frequencies high enough for desktops), and now with Meteor Lake in the new Intel 4 process.

With the 14-nm and 10-nm (now rebranded as Intel 7), Intel has succeeded to greatly increase the maximum clock frequencies after many years of tuning and tweaking. Now, with Meteor Lake, this will not happen, because they will pass immediately to different better manufacturing processes.

According to rumors, the desktop variant of Arrow Lake, i.e. Arrow Lake S, will be manufactured at TSMC in order to ensure high-enough clock frequencies, and not with the Intel 20A, which will be used only for the laptop products.

Intel 18A is supposed to be the process that Intel will be able to use for several years, like their previous processes. It remains to be seen how much time will pass until Intel will become able to reach again 6.0 GHz in the Intel 18A process.

Re: Intel's Humbling

#308

Post by drumhead » Wed, Jan 31, 2024, 2:03 PM UTC

With the US re-industrialising, semi-conductors are a strategic priority so Intel will be at the heart of the process. They're going to be a major beneficiary of the process, here and in Europe. They'll be a significant player.

They got lazy and sat on their laurels when AMD was struggling and they didnt view ARM as a threat. TSMC was a probably a joke to them...until everyone brought out killer products and Intel had no response to them. They could have been way ahead of the pack by now but they decided to harvest the market instead of innovating aggresively. Right now they're worth less than $200bn, which is less than half of Broadcom or TSMC, its 30% less than AMD and 10% of Nvidia. Is it intrisically worth that little? Probably not, I think its a buy at this price.

Re: Intel's Humbling

#309

Post by bonton89 » Wed, Jan 31, 2024, 2:07 PM UTC

klelatti wrote: post #244

BirAdam wrote: post #50

Everyone acts as though Intel should have seen everything coming. Where was AMD? Was AMD really competitive before Ryzen? Nope. Core 2 series blew them out of the water. Was ARM really competitive until recently? Nope. Intel crushed them. The problem for Intel is the inertia of laziness due to a lack of competition. I wouldn’t count them out just yet, however. The company’s first true swing at modern GPU was actually…

> Nope. Intel crushed them. The problem is that Intel has had a defensive strategy for a long time. Yes, they crushed many attempts to breach the x86 moat but failed completely and then gave up attempts to reach beyond that moat. Mobile, foundry, GPUs etc have all seen half-hearted or doomed attempts (plus some bizarre attempts to diversify - McAfee!). I think that, as Ben essentially says, they put too much faith in…

Part of the problem is Intel is addicted to huge margins. Maybe of the areas they have tried to enter are almost commodity products in comparison so it would take some strong leadership to convince everyone to back off those margins for the sake of diversification.

They should have been worried about their process leadership for a long time. IIRC even the vaunted 14nm that they ended up living on for so long was pretty late. That would have had me making backup plans for 10nm but it looked more like leadership just went back to the denial well for years instead. It seemed like they didn't start backport designs until after Zen1 launched to me.

Re: Intel's Humbling

#310

Post by jillesvangurp » Wed, Jan 31, 2024, 2:15 PM UTC

hardware2win wrote: post #287

jillesvangurp wrote: post #281

Earlier quoted context omitted.

There are a few areas where they are under pressure: - The wintel monopoly is losing its relevance now that ARM chips are creeping into the windows laptop market and now that Apple has proven that ARM is fantastic for low power & high performance solutions. Nobody cares about x86 that much any more. It's lost its shine as the "fastest" thing available. - AI & GPU market is where the action is and Intel is a no-show f…

They will be manufacturing those ARM cpus

Has that been announced? Or is it more a matter of Intel producing some unannounced product on an unannounced timeline with a feature set that has yet to be announced on an architecture that may or may not involve arm? Intel walking away from x86 would be a big step for them. First they don't own arm and second all their high end stuff is x86.