Intel Faces Potential Class-Action Lawsuit over Chip Bug Controversy
uk.pcmag.com
uk.pcmag.com
Think back to the Pentium 4 and Athlon days... think about all the anti-competitive, illegal and/or unsavory business practices Intel employed during that era to gain dominance. They literally shut out AMD via illegal behavior - and received a slap on the wrist many years later. Their illegal behavior led to AMD being forced (via unfairly reduced revenue by way of anti-competitive Intel practices) to spin off Global Foundries, and forced to employ only second-tier engineers, managers, executives, etc.
Where would we be if Intel and AMD had actually competed freely during that era. It seems likely Intel would not have been the darling of so many PC enthusiasts, and would not be dominant in business computing either.
Here we are today, with Intel seemingly unable to find their way out of a paper bag...
It's hard to overstate Intel's importance in regards to the modern CPU: they invented the microprocessor itself. (Though I'd like to state it's a bit controversial whether they only were the first to come up with a commercially available one or whether they actually invented the concept itself)
Also, there are many things outside the CPU that we take for granted today that were of Intel's innovation. A great example would be PCIe that Intel invented and essentially standardized. More recent example would be USB4.
Regarding CPUs they have some truly legendary chips, like the Intel 8086. This chip from the 70s is still in use today in your everyday products (as a clone though). It can be found TODAY in common products like computer mice, keyboards, AC, TV remotes and so on. Anything that is a commodity and needs some simple processing power has usually some derivative of 8086 inside it, granted ESP32 and such are taking over nowadays.
Of course the x86 instruction set comes from that very chip. I guess the name speaks for itself how big of a thing that was.
There's too much to fit inside one comment about how innovative Intel truly used to be, but they have a pretty good page for their history here: https://timeline.intel.com/
Look at the Core series - moderately incremental improvements for nearly two decades now? Many of which have significant, unfixable design flaws. AMD got their act together, and with significantly fewer resources ended up totally leap-frogging Intel in nearly all CPU metrics. How did this not happen at Intel after all these years?
AMD had some objectively better CPU's during the Athlon era - until they ran out of money mostly due to Intel's anti-competitive behavior.
The "what-if" scenario is interesting to ponder...
AMD outsourced foundry issues to TSMC.
Who's upgrading their cpus after 1-2 generations?
I have seen more innovation happen in the CPU space in the last 5 years than the 20 years prior to that. This has been thanks to AMD, Apple, ARM, TSMC etc; Intel has seemingly attempted to only slow down the innovation to keep cozy at the top spot.
Personal example: I rented an AMD machine last month and I seriously thought there was a bug with fastfetch when it showed the CPU having 5,7GHz clocks with 32 threads. I didn't believe such was possible. I had to double check because it felt so far fetched seeing such monstrous increase in clocks and cores, when upgrading from a few years old Intel machine to a new AMD one. That's innovation.
However to Intel's credit they have made major innovations in other areas, like peripherals, interconnects and so on. I am extremely grateful for Thunderbolt/USB4 existing today compared to the myriad of vendor specific docking connectors of the past.
>How did this not happen at Intel after all these years?
They replaced engineers with accountants after gaining dominant position on the market. Short term it gave more profits but long term most innovation was lost in the process.
You had CPUs before that, but they were not single chips.
> like the Intel 8086.
You mean the 8051? It is a nice embedded chip, mostly because it has a lot of bit-level instructions and has separate execution and data address spaces if I'm remembering the details right.
> Of course the x86 instruction set comes from that very chip
It does come from the 8086, which I do believe was at least somewhat based on the even earlier 8080. I could be wrong though.
But yeah Intel did a lot. In the 70's though you had lots of semiconductor companies: Fairchild, Signetics, Motorola, MOS/CSG branching off of them. I really wonder what Intel would be if IBM selected a 68000 for the 5150 though.
These chips with 70s design are very common in ordinary "non computing" products, granted most are Chinese clones with varying levels of modification. I spend a lot of time reverse engineering regular commodity devices, their electronics and their firmware. For example the portable AC I just opened had 2 of these 8051s(?) inside, with an ESP32 for networking which I find rather fascinating. The first one controls physical inputs (buttons) and IR input from a remote, the second controls the AC compressor, fans and lower level electrical inputs (sensors etc) while taking input from the first and the ESP32 handles wireless communication sending input to the first one.
These old low performance chips are found inside mice, keyboards, remote controllers, dehumidifiers, air fryers and almost any other "simple" electronics. It's fascinating how a 70s chip design is still so prevalent in our everyday products.
Current execution stumbles do not diminish historical triumphs -- Pentium II and Nehalem were and are amazing.
Crazy remembering a time when there was that much 16-bit code in the wild.
Not for those who have "$THING was never good" mindset after repeated, multiple stumbles in a row. You see this with anything that closed/died in disgrace.
I imagine there are people out there who think that Intel might file for bankruptcy/insolvency if they can't right themselves from this issue.
Intel, in the past 20 years, has floundered and misfired constantly. It's not a stretch to assert the only reason Intel is still around is due to their illegal business practices in the early and mid 2000's. Intel became synonymous with CPU's because they were literally (and illegally) the only option for 99% of the computer buying market.
Intel didn't invent/discover FinFET, nor were they the first to commercialize it.
[1] https://en.wikipedia.org/wiki/Fin_field-effect_transistor
>> Commercially produced chips at 22 nm and below have generally utilised FinFET gate designs (but planar processes do exist down to 18 nm, with 12 nm in development). Intel's tri-gate variant were announced at 22 nm in 2011 for its Ivy Bridge microarchitecture.[30] These devices shipped from 2012 onwards. From 2014 onwards, at 14 nm (or 16 nm) major foundries (TSMC, Samsung, GlobalFoundries) utilised FinFET designs.
Even if we examine Intel's Tri-gate Transistor contribution to existing FinFET technology, we can see even there it was not an Intel invention/discovery.
>caught the attention of attorneys looking to capitalize on it.
Yes, I will gladly contact some suave gentlemen in business suits so I can make 10 cents while they make bank if I register with so-and-so entity after a dozen years of wasting the judiciary's time, all the while crapping on a company I like.
Not.
I don’t like lawyers but I guarantee there are equally if not more suave Intel lawyers who would tell customers to shove it if they could.
So this supposed lawsuit isn't even serving that dubious purpose.
As an aside, I have a 14700K. I've not filed for an RMA because it's been working perfectly fine for the better part of a year now.
I guess my point is that it’s better if Intel behaves, but if they don’t we have some recourse. Lawyers are a last resort thing.