• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Intel 18A Process Node Clocks an Abysmal 10% Yield: Report

in the fantasy world where there are "good" and "evil" corporations and if you can just eliminate those darn evil ones everything is ok...
What you call "fantasy" is precisely why antitrust laws exist.

If all corporations were equally evil, then no single one should be allowed to reach market dominance. But no, it is only monopoly abuse which runs afoul of antitrust and leads to government intervention. And the remedy is ensuring that the company does not have the ability to repeat such behavior, e.g. through breaking it up in several parts.
 
Newsflash: The PC is no longer commanding sales numbers that justify expensive development of CPUs just for it.
This is because to get another 1% more performance, you'd need a lot of R&D to do so. Most consumers aren't willing to spend hundreds or thousands for a few percent, but data centers would in a heart beat.
AMD uses since 2017 (Zen) repurposed server silicon for their higher performing processors on the desktop. For notebooks and tablet PCs, we are seeing repeated attempts by Microsoft since Windows RT (2013) to prepare for the era of repurposed mobile processors. For Apple macOS (and also Google Chrome OS), this era is already here.

View: https://www.youtube.com/watch?v=_XNLXQ7_bs4

Chrome OS is trash. Apple's ARM chips are not as good as people think. ARM supporters quickly point at Geekbench and Cinebench, but rarely run actual applications and do tests. Phoronix did some tests and while the M4 is extremely power efficient, it was nearly dead last in majority of tests, with maybe a handful where it won against x86 chips with less than 10 cores. It did win in FLAC audio. There's also gaming, which Apple silicon does have a fair amount of native ported games to make direct comparisons to x86 machines.
If/when ARM comes to dominate in servers, then you won't have a choice any more, and live with x86 emulation for your legacy games. This is also why Intel going away would not mean the end of competition in the desktop market.
Right now, AMD is the most efficient and the fastest. The ARM chips are cheaper, so they do win in price performance but it's likely that AMD and Intel will price their products accordingly to compete. In other words, competition is good.
NVIDIA has done really shady stuff and is up there with Intel and Microsoft as the three most evil and anti-competitive companies to ever exist in tech. But at least they innovate and relentlessly work on delivering bigger and better gear to their customers (and make them pay through the nose for it).
Like what? Ray Tracing which kills performance? DLSS which kills image quality? CUDA is about the only thing Nvidia has done right, and I'm not into AI to care.
AMD on the other hand does not meet the definition of shady. Yes they made many stupid moves and are among the least agile in the industry. But even when AMD was in the performance lead, their product prices were always fair, and they never engaged in anti-competitive practices.
You do know that AMD and Nvidia have colluded to price fixing in the past? AMD is no saint.
Intel is evil from the core and used to getting their products into people's hands even if inferior, if necessary by paying money to lock out competitors and making back that money when customers no longer have a choice as result.
You can call Intel evil, but inferior products? That's only something you can say since maybe 2019, when AMD released Zen3. Before that, Intel was ahead.
We can all be happy that this didn't work in mobile (google "Intel contra revenue"), and now this scheme also collapsed in desktop and server markets as Intel went too long without competitive product.
The reason Intel would never work out on mobile is because they charge too much. The allure of ARM is that you can make some really cheap chips. Everything else about ARM was secondary. If Intel wanted to compete then they'd license out x86.
 
Chrome OS is trash. Apple's ARM chips are not as good as people think. ARM supporters quickly point at Geekbench and Cinebench, but rarely run actual applications and do tests. Phoronix did some tests and while the M4 is extremely power efficient, it was nearly dead last in majority of tests, with maybe a handful where it won against x86 chips with less than 10 cores. It did win in FLAC audio. There's also gaming, which Apple silicon does have a fair amount of native ported games to make direct comparisons to x86 machines.
Apples ARM chips are only good because the OS is making use of every single accelerator and extension to optimize every possible aspect of the hardware/software integration.
Apple gets to make use of a degree of hardware optimization other platforms just can’t, so it lets them do a lot more with less.
But you take those same chips into a generic environment with none of those optimizations and suddenly performance is average at best.
 
In the past they publish a lot of pro Korea anti everyone else hit pieces.

I’d believe it to be low but not 10%, likely closer to 40%.

But Toms published this not too long ago.

https://www.tomshardware.com/tech-i...8a-is-healthy-potential-clients-are-lining-up

"I am happy to update the audience that that we are now, for this production process, we are now below 0.4 d0 defect density, this is now a healthy process," said Pat Gelsinger, chief executive of Intel, at the technology conference.


When it comes to defect density (D0), then it is commonly believed that a defect density below 0.5 defects per square centimeter is good (0.5 def/cm^2), so Intel 18A's 0.4 defects per cm^2 is a good result (at least on paper), especially considering the fact that the production technology is a few quarters away from entering mass production, at which point defect density will likely get lower.

Ousted Intel CEO Steps In To Defend The Firm’s 18A Process, Says Yield Rate % Isn’t The Right Metric To Measure Semiconductor Progress​


1733606587634.png


Source: https://wccftech.com/ousted-intel-c...ght-metric-to-measure-semiconductor-progress/
 
Apples ARM chips are only good because the OS is making use of every single accelerator and extension to optimize every possible aspect of the hardware/software integration.
Apple gets to make use of a degree of hardware optimization other platforms just can’t, so it lets them do a lot more with less.
But you take those same chips into a generic environment with none of those optimizations and suddenly performance is average at best.
Another way Intel has stiffed innovation. (half joking). Remember that time in the 90s when Intel started doing software to take advantage of their fancy new MMX extensions and the like. When they started making software that made Microsoft look terrible. Intel caved and said ok ok MS we'll stop making software to show off our hardware.

--- finally an opportunity to also shit on Microsoft. lol
 
Apples ARM chips are only good because the OS is making use of every single accelerator and extension to optimize every possible aspect of the hardware/software integration.
Apple gets to make use of a degree of hardware optimization other platforms just can’t, so it lets them do a lot more with less.
But you take those same chips into a generic environment with none of those optimizations and suddenly performance is average at best.
This has some truth as Asahi Linux tends to run a bit slower compared to MacOS. This is probably one of the most interesting benchmarks I've seen so far. These guys did tests on x86 hardware running Debian 12 and with Apple hardware that could use Asahi Linux and MacOS. Even the guy managing this has said that the gap would only widen if he ran Windows, and you'd get similar performance benefits if you ran MacOS on AMD and Intel hardware.
SPECint2017.png
 
Lakados and other industry insiders. Is Gelsinger's claim credible? But I seem to recall reading that TSMC got a 60% yield in one of their advanced fabs.
That's where it gets fun, TSMC got 60% but the chips they were making were tiny, so we are talking about 60'ish percent on chips like the A17, but the M series chips were so low yield that TSMC was only charging Apple for functional chips instead of the full wafers like they do for others.
If you have a defect rate you can average out to some metric of errors per square inch or something, then you know that chips manufactured below that size are likely to be fine and easily binnable with a low rate of complete failure.
You might have some you need to cut a section off of so maybe 2 fewer cores, or fewer SM's, or maybe a lower clock speed, but still usable, and of course you will have some with absolutely no errors, but you should have very few with 2 or more errors or in a state rendering it completely useless.
So at an engineering level yield rates are going to be specific to what you are making, and what the tolerance of that design.

Going back to that TSMC yield percentage that 60% on the A series was done with a defect rate of 0.33, so lets give Intel the benefit of the doubt and say theirs was 0.399 (as that is less than 0.4) that would put their yields on something small like the A series chips probably in the 50% yield rate, and again abysmal on something larger.

So behind where TSMC was on their initial 3N process, but not by a lot, and it's not yet in production so Intel still has time to bring that defect rate down, and it should only improve over time as they lock their process and design requirements down.

So I would believe his claims to be credible.
 
Exactly. % perfect/usable chips is bullshit if you want to compare one process against another without considering what product is being made. The real measure is defects per unit of wafer area. There is a massive difference between a budget phone chip and a high end NV GPU that's up against the reticle limit. The NV chip will likely have multiple defects, but NV will use it anyway because they have a modular design and planned to be able to disable defective bits and a lot of them will still be 4090s.
 
Exactly. % perfect/usable chips is bullshit if you want to compare one process against another without considering what product is being made. The real measure is defects per unit of wafer area. There is a massive difference between a budget phone chip and a high end NV GPU that's up against the reticle limit. The NV chip will likely have multiple defects, but NV will use it anyway because they have a modular design and planned to be able to disable defective bits and a lot of them will still be 4090s.
So was Gelsinger being scapegoated for failures across the company? To be clear, I use only AMD CPUs, done that for years now, and I certainly don't own any Intel stock. (Don't own any individual stocks any more. I can't outsmart the professionals.)
 
So was Gelsinger being scapegoated for failures across the company? To be clear, I use only AMD CPUs, done that for years now, and I certainly don't own any Intel stock. (Don't own any individual stocks any more. I can't outsmart the professionals.)
Not scapegoated, but I’m betting the board wanted him to cut back on the fab upgrades, increase orders to TSMC and prioritize military contracts for the new Intel advanced nodes.
Even with a terrible failure rate on the military chips there’s enough of a markup to keep it profitable.

Because while Pat is gone, so are more than a few board members.
 
This is because to get another 1% more performance, you'd need a lot of R&D to do so. Most consumers aren't willing to spend hundreds or thousands for a few percent, but data centers would in a heart beat.
Willingness to spend top money for few % performance gain used to be true for most of the PC era until its peak in 2011, and is still true to some degree among gaming enthusiasts. Datacenter was always this way. Nowadays, no more. So this explanation doesn't hold water I think.

Chrome OS is trash.
On the contrary, Chrome OS is excellent. It enabled $200 x86 and ARM notebooks which is the most that many people were willing to spend on a PC, and it would be fine for web browsing, word processing and other basic tasks. Compare this to $200 Windows notebooks which were universally painful to use. For a long time, Chromebooks were the only growing segment in an overall declining PC market.

Apple's ARM chips are not as good as people think. ARM supporters quickly point at Geekbench and Cinebench, but rarely run actual applications and do tests. Phoronix did some tests and while the M4 is extremely power efficient, it was nearly dead last in majority of tests, with maybe a handful where it won against x86 chips with less than 10 cores. It did win in FLAC audio. There's also gaming, which Apple silicon does have a fair amount of native ported games to make direct comparisons to x86 machines.

Right now, AMD is the most efficient and the fastest. The ARM chips are cheaper, so they do win in price performance but it's likely that AMD and Intel will price their products accordingly to compete. In other words, competition is good.
AMD doesn't hold a candle to M4 even in real-world testing, not until Strix Halo (and possibly not even then). Maybe in the sub-$800 laptop market where Apple doesn't compete, AMD and Intel are still king.

Like what? Ray Tracing which kills performance? DLSS which kills image quality? CUDA is about the only thing Nvidia has done right, and I'm not into AI to care.
You don't care about ray tracing and DLSS maybe. (Perhaps you want take another look: The tech press used to be critical too initially, and are now praising DLSS.) But even if you narrow your view into performance only: It is undisputable that NVIDIA used their massive earnings not only to pay shareholders, but to actually make their products faster, and not stagnate like Intel did between Sandy Bridge and Skylake.

This was ATI, not AMD. The acquisition of ATI by AMD was only completed in October 2006, and the reported price fixing happened before that. ATI was a somewhat shady company, but far from NVIDIA level shady.

Intel is evil from the core and used to getting their products into people's hands even if inferior, if necessary by paying money to lock out competitors and making back that money when customers no longer have a choice as result.
You can call Intel evil, but inferior products? That's only something you can say since maybe 2019, when AMD released Zen3. Before that, Intel was ahead.
AMD was also ahead between Thunderbird (1999) and Windsor (2006). Both the reduced consumer choice and the inferior products of AMD between Barcelona and Zen were the result of Intel anti-competitive practices which were detailed in the EU antitrust case. Intel preventing AMD products from reaching the market meant that AMD had less revenue to re-invest in product development.

The reason Intel would never work out on mobile is because they charge too much. The allure of ARM is that you can make some really cheap chips. Everything else about ARM was secondary. If Intel wanted to compete then they'd license out x86.
I told you to google "Intel contra revenue".
Intel could not work out things in mobile even when paying companies to accept their products.

Edit: I just remembered that back in the day, Paul Thurrott did specifically ask HP why they would not use Intel chips in their Elite x3 Windows smartphone, and they replied that Intel technology was behind by years.
 
Last edited:
On the contrary, Chrome OS is excellent. It enabled $200 x86 and ARM notebooks which is the most that many people were willing to spend on a PC, and it would be fine for web browsing, word processing and other basic tasks. Compare this to $200 Windows notebooks which were universally painful to use. For a long time, Chromebooks were the only growing segment in an overall declining PC market.
You know what else would do all that and a lot more? A GNU/Linux based laptop. ChromeOS is just Google trying to lock people into their... web browser ecosystem? It's seriously a stupid OS.
AMD doesn't hold a candle to M4 even in real-world testing, not until Strix Halo (and possibly not even then). Maybe in the sub-$800 laptop market where Apple doesn't compete, AMD and Intel are still king.
Max Tech isn't exactly someone I like looking at for benchmarks. Right now the Ryzen AI HX 370 stupid ass name, is not as good as Apple's M4. Just not the world of difference that these reviewers are claiming. Mostly because these reviewers use stupid benchmark tools, like Max Tech. For Gaben sake, the dude ran Geekbench's GPU test and Apple's M4 Pro won. Then when he ran 3D Mark and Blender, which were both much faster on the RTX 4070. 3D Mark is not a good representation for gaming performance even still. Just get a game and run that. How hard is it? Apple's M series chips now have a decent enough games ported where you can stop using 3D Mark and Geekbench.
You don't care about ray tracing and DLSS maybe. (Perhaps you want take another look: The tech press used to be critical too initially, and are now praising DLSS.)
Just ignore all the people criticizing DLSS, along with FSR and TAA.
But even if you narrow your view into performance only: It is undisputable that NVIDIA used their massive earnings not only to pay shareholders, but to actually make their products faster, and not stagnate like Intel did between Sandy Bridge and Skylake.
Is that with or without 8GB of VRAM? Is that with or without the RTX 4060? You do know that the RTX 3060 12GB is faster in a lot of titles compared to the RTX 4060 8GB, and costs less too?
I told you to google "Intel contra revenue".
Intel could not work out things in mobile even when paying companies to accept their products.

Edit: I just remembered that back in the day, Paul Thurrott did specifically ask HP why they would not use Intel chips in their Elite x3 Windows smartphone, and they replied that Intel technology was behind by years.
This is the same company that also wanted to make Windows 10 Mobile-based phone. Guess they couldn't see the software was outdated too. I had an Asus Zenphone with an Intel x86 chip and it worked well. I'm actually in the process to replace the battery since it's completely dead. I like keeping my old things running, even if I don't use them.
 
You know what else would do all that and a lot more? A GNU/Linux based laptop. ChromeOS is just Google trying to lock people into their... web browser ecosystem? It's seriously a stupid OS.
It's progressively gotten worse too, fewer plugins, and more reliance on the Android App store and the apps don't always work well at all.
 
Doesn’t matter either way intel is just fucked. And I’m glad.

Definitely. I'm betting AMD cuts prices by 20% on their first day as the sole desktop CPU company. History certainly bears out the fact that corporations are hesitant to take profit opportunities when they can give their customers a good deal instead.
 
Definitely. I'm betting AMD cuts prices by 20% on their first day as the sole desktop CPU company. History certainly bears out the fact that corporations are hesitant to take profit opportunities when they can give their customers a good deal instead.
I have a 7950x So yes I do hope intel fails. I work offshore and i can que up all my cpu tasks via remote desktop and have everything done when i get home. So yes I do hope intel fails and goes under 7950x is under $500 so I am good if i ever wanted another. I am also a 13900k owner and got fucked HARD. so YES I DO HOPE THEY GO UNDER.
 
So was Gelsinger being scapegoated for failures across the company? To be clear, I use only AMD CPUs, done that for years now, and I certainly don't own any Intel stock. (Don't own any individual stocks any more. I can't outsmart the professionals.)
Well Pat is responsible imo for the down fall of Intel dating back to his time as CTO... he put so much into motion that has bit them in the ass for 20 years now.

Anyway I would say the real reason he has been retired was the Fab spin off plan. Pat decided he wanted to take the fab and decouple Intel proper from it. I mean decoupling the internal teams so customers would maybe trust Intel enough to fabricate competing products was the only solution there. Spinning it off financially though would have no doubt been the end of Intel. To believe that was going to work and Intel wasn't going to end up sold off in bits... you would have to be some sort of delusional individual who really truly believed Intel was the best company ever and had some sort of super power of winning. OHH wait ya that is Pat.

The board got rid of him before he tanked the company for real. The US gov gave the board the impetus to get it done. Pretty sure, the requirement for them receive Chips funds was a commitment to NOT split the fabs off. At that point Pat either said I can't go along with that I'm out... or I can't go along with that and the board said OK your out.
 
It's progressively gotten worse too, fewer plugins, and more reliance on the Android App store and the apps don't always work well at all.
Kids in school end up with these Chromebooks and want to know if they can have Windows installed. The devices are locked to the school and they have no plans to ever give them back. The device is now ewaste. These devices are such cheap junk that it isn't worth the time to figure a way around it.
 
The board got rid of him before he tanked the company for real. The US gov gave the board the impetus to get it done. Pretty sure, the requirement for them receive Chips funds was a commitment to NOT split the fabs off. At that point Pat either said I can't go along with that I'm out... or I can't go along with that and the board said OK your out.

would imagine the subsidiary plan were always made with enough Intel ownership to not loose CHIP money ? It was not to decouple what Intel would have become from it, but decouple Intel new design sudivision, from intel foundry new foundry subdivision. A bit like Coca-cola bottling being a bit independent and bottling competitor product from the Coca-cola concentrate and brand subdivision.
 
This was ATI, not AMD. The acquisition of ATI by AMD was only completed in October 2006, and the reported price fixing happened before that. ATI was a somewhat shady company, but far from NVIDIA level shady.
Not to overly defend ATI. The whole price fixing stuff. People need to remember how small ATI... and for that matter Nvidia were back then. Its easy now to think AMD the big x86 player, they power Sony and MS consoles big big. Nvidia they are for real the highest valued company in the world. They colluded EVIL. Accept at the time in question ATI was sturggling to stay alive. Nvidia, AMD, PowerVR, Intel, Matrox were all in the GPU business. Margins were stupid low cause no one really saw the value in crazy expensive GPUs. ATI had bought ARTx which got them a nintendo console contract but the margins were so slim any hicuup in production would cause looses. ATI had to make products like the wonder line to try and eak out some extra profit.

Anyway just saying its not like ATI and Nvidia were the rulers with 100% market share between the two of them of some massive 100s of billions of dollars industry. I mean GPUs are apparently a 75 billion dollar industry in 2024 projections from different sources expect GPUs to grow to 400b-1.4t by 2030 or so. In 2008 it wasn't even a 100 million dollar a year industry.
 
Well Pat is responsible imo for the down fall of Intel dating back to his time as CTO... he put so much into motion that has bit them in the ass for 20 years now.

Anyway I would say the real reason he has been retired was the Fab spin off plan. Pat decided he wanted to take the fab and decouple Intel proper from it. I mean decoupling the internal teams so customers would maybe trust Intel enough to fabricate competing products was the only solution there. Spinning it off financially though would have no doubt been the end of Intel. To believe that was going to work and Intel wasn't going to end up sold off in bits... you would have to be some sort of delusional individual who really truly believed Intel was the best company ever and had some sort of super power of winning. OHH wait ya that is Pat.

The board got rid of him before he tanked the company for real. The US gov gave the board the impetus to get it done. Pretty sure, the requirement for them receive Chips funds was a commitment to NOT split the fabs off. At that point Pat either said I can't go along with that I'm out... or I can't go along with that and the board said OK your out.
At some point an insider from Intel will tell people the reasons. Your explanation would make the most sense.
 
  • Like
Reactions: ChadD
like this
In 2008 it wasn't even a 100 million dollar a year industry.
jpr_q2_2016_mkt_historical_annual_gpu_sales.png


They were selling close to 100 millions graphic cards back in those days... that seem too low of an average even if very low were quite popular before iGPU became the norm.

And by 2008, Matrox-PowerVR-Intel going from memory were done in the desktop gpu world and had been for a good while, it was an AMD vs Nvidia affair, with Intel gaining giant steam in iGPU.

https://arstechnica.com/gadgets/200...st-quarterly-growth-in-six-years/?form=MG0AV3
Total GPU shipments for the third quarter in 2008 was some 111 million units

Desktop GPU shipments are still ruled by Intel with 43.9 percent of the market, an increase in market share from Q2. NVIDIA's share of the desktop GPU market slipped to 32.6 percent and AMD's share climbed to 20.3 percent of the market.


That was already getting close to 100% market share for the big 3 (with I imagine intel being almost purely mobile-igpu sales)
 
would imagine the subsidiary plan were always made with enough Intel ownership to not loose CHIP money ? It was not to decouple what Intel would have become from it, but decouple Intel new design sudivision, from intel foundry new foundry subdivision. A bit like Coca-cola bottling being a bit independent and bottling competitor product from the Coca-cola concentrate and brand subdivision.
That is what they had done internally. After Intel face planted for the first year trying to sell fabrication to third parties... they realized no one was going to trust Intel fabrication with their designs. The possibility of Intel stealing from you was just way to high. So Pat did wisely decouple the design and fab teams. I mean that probably did make life a little harder for Intel, it was required to get real business for the fab.

Its my understanding Pats future plans were not a coke situation where Coke proper and its distributors co own a bottling entity. He wanted to go full split, including a new stock offering for Intel Fabrication. He wanted to take the fab public was my understanding. I'm sure he planned to retain a % of Intel ownership. It wouldn't have mattered. Intel design would have been dealing with mulitple hostile take over attempts and surviving them would have been an Intel reality for a few years until it failed, and the winner of Intel proper would have fully divested from Intel fabrication.
 
View attachment 696884

They were selling close to 100 millions graphic cards back in those days... that seem too low of an average even if very low were quite popular before iGPU became the norm.
I thought all the ATI-Nvidia fixing price mins went back to like 2006-2008 type time frames. Back when ATI was still selling things like wonder cards.. and most PCs didn't even have 3D acceleration. But I could be wrong.
 
I thought all the ATI-Nvidia fixing price mins went back to like 2006-2008 type time frames. Back when ATI was still selling things like wonder cards.. and most PCs didn't even have 3D acceleration. But I could be wrong.
Yes lot of card were really cheap back then, because iGPU did not replace yet all of the not for gaming card.

But the volume of cards sold were so much bigger than today that it would have took an under $2 by cards to be an under 100 millions world business no ?, Nvidia 2007 revenue were above 3 billion, GPU was not there only business but most of it. ATI was around 2.4 billion in 2006.
https://www.sec.gov/Archives/edgar/data/2488/000119312507003669/dex991.htm

In my mind (going from memory) the Parhelia was the last attempt of Matrox to compete directly with ATI-Nvidia and that was in 2002. And the rest of the competition (3dfx (voodoo 5 in 2000 maybe ?), SGI, etc...) were pretty much done or really clear they did not had for long around that 2002-2003 time as well, when the ATI vs Nvidia battle became really clear at the latest, I do not remember having considered seriously a non ATI-Nvidia card around the GeForce 4/9700 pro era, like I would have in the Voodoo 2-3 era.
 
Yes lot of card were really cheap back then, because iGPU did not replace yet all of the not for gaming card.

But the volume of cards sold were so much bigger than today that it would have took an under $2 by cards to be an under 100 millions world business no ?, Nvidia 2007 revenue were above 3 billion, GPU was not there only business but most of it. ATI was around 2.4 billion in 2006.
https://www.sec.gov/Archives/edgar/data/2488/000119312507003669/dex991.htm
Fair. Still was a much much smaller market... and they had nothing close to a monopoly. As the article you linked stated... Of the six GPU makers—Intel, NVIDIA, AMD, Matrox, VIA, and SiS.
Not trying to overly defend ATI... but ya I think the collusion stuff was ATI not AMD so that means prior to 2006. I think even 3DFX was still in play at the time of some of the emails they dug up to bring action on.

https://www.tomshardware.com/news/nvidia-amd-ati-graphics,6311.html this is about the only article I could really find on the actual suits. We are talking pre AMD purchase for sure.
I remember vaguely at the time thinking those suits were stupid... and an attempt to make some money. IMO I didn't see any issue with the two working together in some regards to build out the industry. Those class actions were all based on silly innocuous stuff like this. They were mostly talking about marketing the idea of 3d acceleration and working together to make it a PC defacto standard. Most PCs were still shipping with no acceleration at all.

In another email between Dan Vivoli of Nvidia and Dave Orton of ATI, Dan wrote:
"I really think we should work harder together on the marketing front. As you and I have talked about, even though we are competitors, we have the common goal of making our category a well positioned, respected playing field. $5 and $8 stocks are the result of no respect."


I mean reading that email I guess I could see someone saying COLLUSION. People have to remember this was a time when Nvidia and ATI were trying to get PC OEMS to ship machines with RivaTNT and ATI wonder cards. There is no comparison to todays PC world. They just both needed to convince the industry that 3D acceleration pre installed would move PCs and justify their cost. Neither could afford to let Dell HP and Compaq (cause they were still a big deal lol) squeeze them for cards at cost.
 
Last edited:
Fair. Still was a much much smaller market... and they had nothing close to a monopoly. As the article you linked stated... Of the six GPU makers—Intel, NVIDIA, AMD, Matrox, VIA, and SiS.
Yes but like the article say:
43.9 percent of the market, an increase in market share from Q2. NVIDIA's share of the desktop GPU market slipped to 32.6 percent and AMD's share climbed to 20.3

43.9+32.6+20.3=
96.8%

It was over for all the others, outside medical visualization stuff, workstation, large multimonitor control room, traffic, broadcast, etc... more niche endavour.

In the discrete for regular Desktop tower around that time, ATI-Nvidia was probably getting close to 100% (intel gpu market is bigger now than the 43 of the time, that was I imagine all laptop-iGPU sales)

but ya I think the collusion stuff was ATI not AMD so that means prior to 2006
Yes must have been prior to 2006 and quite a bit let alone 2008 if they were not a duopoly yet, last serious 3dfx attempts was in summer 2000 maybe, Nvidia bought them at the end of 2000.
 
Yes but like the article say:
43.9 percent of the market, an increase in market share from Q2. NVIDIA's share of the desktop GPU market slipped to 32.6 percent and AMD's share climbed to 20.3

43.9+32.6+20.3=
96.8%

It was over for all the others, outside medical visualization stuff, workstation, large multimonitor control room, traffic, broadcast, etc... more niche endavour.

In the discrete for regular Desktop tower around that time, ATI-Nvidia was probably getting close to 100% (intel gpu market is bigger now than the 43 of the time, that was I imagine all laptop-iGPU sales)

Yes must have been prior to 2006 and quite a bit let alone 2008 if they were not a duopoly yet, last serious 3dfx attempts was in summer 2000 maybe, Nvidia bought them at the end of 2000.
Those law suits were brought in 2008. They were not based on sales from 2008. We are talking about sales from 2000-2006. Yes in 2002 when 3dFX was bought they had a massive part of the market. PowerVR was actually still shipping cards. Matrox was still selling gaming cards. Intel hadn't folded on the idea of selling cards yet.

I'm not sure where your math is coming form you just posted "NVIDIA's share of the desktop GPU market slipped to 32.6 percent and AMD's share climbed to 20.3" that is 52.9% of the market between AMD and Nvidia at that point. Yes I understand that Intel was the majority of the rest in iGPU sales. That is my point. OEMs were not be default shipping machines with GPUs. Its common now if you buy a PC for more then a couple hundred bucks it will have some low end Nvidia or ATI GPU in it. In 2008 that wasn't true. AMD and Nvidia were competing for OEM contracts where companies like Dell and HP were negotiating deals for GPUs for OEM machines sometimes asking for cards for single digit margin pricing. (remember all the terrible OEM only cards back then with fans that barely spun if they even had active cooling... and heat sinks from the bargain bin lol) It was in that environment that Nvidia said to ATI hey we are trying to grow a brand new sector here we can't be pushing pricing so hard that neither of us can afford to be in the market. (the min pricing being talked about was for OEMs... who played hardball, and the resulting product both ATI and Nvidia were shipping to them was sort of embarrassing and setting the whole 3D acceleration adoption back)

I'm not even sure what happened with the California class action... I'm pretty sure it ended up getting tossed. (heck I even wonder if one of the Cali OEMs sort of put people up to that suit... the OEM pricing shit that went on in those days on all fronts was insanity)
 
Last edited:
Yes in 2002 when 3dFX was bought they had a massive part of the market.
Nvidia start to buy 3dfx in 2000:
https://www.zdnet.com/article/3dfx-dissolves-sells-assets-to-nvidia/
The vendor of 3D graphics accelerators on Friday said its board has voted to dissolve the company following the sale of most of its assets to competitor Nvidia for cash and stock.
Dec. 15, 2000 at 4:00 p.m. PT


In 2002 is when Nvidia closed 3dfx completely, I think, I do not remember them having more than very marginal market share in 2001-2002, if they were still selling making new regular dGPU chips at all.

I'm not sure where your math is coming form you just posted "NVIDIA's share of the desktop GPU market slipped to 32.6 percent and AMD's share climbed to 20.3" that is 52.9% of the market between AMD and Nvidia at that point

Intel with 43.9 percent of the market, an increase in market share from Q2. NVIDIA's share of the desktop GPU market slipped to 32.6 percent and AMD's share climbed to 20.3 percent of the market.

That is my point
Well that still true today, they are less a duopoly now than in 2006 (intel has now 60% of the gpu market) under that logic, for regular consumer dGPU they were already pretty much all the market.
 
"Intel 18A Yields Are Actually Okay, And The Math Checks Out
Editorial by AleksandarK Today, 01:49 Discuss (26 Comments)
A few days ago, we published a report about Intel's 18A yields being at an abysmal 10%. This sparked quite a lot of discussion among the tech community, as well as responses from industry analysts and Intel's now ex-CEO Pat Gelsinger. Today, we are diving into known information about Intel's 18A node and checking out what the yields of possible products could be, using tools such as Die Yield Calculator from SemiAnalysis. First, we know that the defect rate of the 18A node is 0.4 defects per cm². This information is from August, and up-to-date defect rates could be much lower, especially since semiconductor nodes tend to evolve even when they are production-ready. To measure yields, manufacturers use various yield models based on the information they have, like the aforementioned 0.4 defect density. Expressed in defects per square centimeter (def/cm²), it measures manufacturing process quality by quantifying the average number of defects present in each unit area of a semiconductor wafer.

Measuring yields is a complex task. Manufacturers design some smaller chips for mobile and some bigger chips for HPC tasks. Thus, these two would have different yields, as bigger chips require more silicon area and are more prone to defects. Smaller mobile chips occupy less silicon area, and defects occurring on the wafer often yield more usable chips than wasted silicon. Stating that a node only yields x% of usable chips is only one side of the story, as the size of the test production chip is not known. For example, NVIDIA's H100 die is measuring at 814 mm²—a size that is pushing modern manufacturing to its limits. The size of a modern photomask, the actual pattern mask used in printing the design of a chip to silicon wafer, is only 858 mm² (26x33 mm). Thus, that is the limit before exceeding the mask and needing a redesign. At that size, nodes are yielding much less usable chips than something like a 100 mm² mobile chip, where defects don't wreak havoc on the yield"
 
Kids in school end up with these Chromebooks and want to know if they can have Windows installed. The devices are locked to the school and they have no plans to ever give them back. The device is now ewaste. These devices are such cheap junk that it isn't worth the time to figure a way around it.
I’m literally in the process of phasing out some 1500 Chromebooks because Google’s been pulling features, and rolling back support dates for software updates. But all those Chromebooks were part of the Google Education platform which still uses the Intel N platform. The latest 2024 model being based around the N4500 with at most 8GB ram and 64GB of storage, all soldered.
Maybe if they splurged recently they got the N100 models…. But same memory and storage config options, so they may be 4GB ram and 32GB storage.

Your not running Windows anything on that without wanting to pull your teeth out.

Either way, Google EDU Admin is falling apart and Apple Education is picking up the slack.

Microsoft is currently pissing us off because security updates are mandatory as part of the Entra best practices and compliance config.
But it breaks shit unless you have their Intune driver management configured as applying security updates automatically with no regard for drivers and firmware is going to eventually land you in hot water. But they charge through the ass for the patch management.

Apple Education is looking mighty nice right now for the sole reason of not being either the Google or Microsoft solution…
They are gobbling up contracts with district after district simply by not fucking things up.

I give it 3 years, but at the rate Microsoft and Google are going in the Edu space unless they get their shit together, it’s going to be Apple in most schools.
 
I am a 13900k owner. For months i thought it was something i did in setting and my computer kept crashing and crashing to the point I build a 2nd computer. I am going to send a e-mail with my info and see what they say for a rma. I hope its easy to get a rma. But yes fuck intel for the bull shit chip i paid like $700 for back in the day.
You can get mad about this or you can just do what I did and be happy. Picked up a 12900K for 240 bucks, RMA'ed my 13900K and Intel cut me a check for nearly 700 dollars. Took em about 2 months but they made good on their end. I would say it's a 5-8 week process if they approve you for the processor. Initially they were handing out 14900K's for your trouble.

Granted it took Intel almost a year to admit they were bending all of us over. However, they are making good on their replacement policy. I would skip the replacement and go straight for the cold hard cash and then dump that into a 9800X3D or something and sell your old intel board on ebay or here.

No good can come from Intel going Tits up. We need healthy (or even unhealthy, underperforming shite) competition to give us options and such. Even AMD's FX blunder opened the door for plenty of affordable CPUs and they were barely hanging in there. Intel is in much the same position now, in theory.

So, turn that frown upside down, and get your ass some money (and avoid Arrow Lake too).

Just my two cents.
 
Stating that a node only yields x% of usable chips is only one side of the story, as the size of the test production chip is not known. For example, NVIDIA's H100 die is measuring at 814 mm²—a size that is pushing modern manufacturing to its limits.
And not just size, an Nvidia H100 can have 20% of its cores not working and still being a successful on the worked column of the yield column (14,592 out of 18,432 cores are enabled), size + error tolerance are needed. And his the error rate the same for cache transistor part, logic part.

No good can come from Intel going Tits up.
Depends how and a bit of a long shot and depends what people mean, if it is quick and sold their asset-patents and the state force a deal with the buyer and AMD for the x86_64 licensing (or we can get away from a fully closed ISA....), there could be.
 
Back
Top