EU legal eagle: Euro court should review Intel's €1.6bn fine An EU High Court Advocate General has recommended a review of the case that saw Intel slapped with a record fine after it was found to have coerced OEMs to avoid using rival companies' x86 CPUs. Back in 2009, the EC imposed a €1.06bn penalty - 4.15 per cent of Intel's turnover - following a multi-year legal battle but …
At the beginning of the period reviewed in the court's decision there used to be 5:
Cyrix
Transmeta
Via
AMD
and Intel
Most of the competitors had various unique innovations. The built-in crypto engine in Via, the power utlization of Transmeta, the 64 bit mode of AMD.
There are now two. While Via officially still makes x86 compatible chips you cannot find their latest ones in the developed world. The history has spoken on the effect of Intel's marketing practices.
I say that 4% is not enough. Should be 10.
It is however a bit odd to specify they are domination one type of chip instruction set, there were still a few RISC processors being made on the server and network device side and by then the consumer market was starting shifting to phones and on the edge of accepting tablets. Sure there is good things about compatibility for build targets but sticking with x86 legacy has drawbacks too.
All history shows is that those companies failed to compete. The question is whether the failure was due to monopoly abuse by Intel. ISTR that Cyrix chips were somewhat unreliable, I have never heard of Via, Transmeta had a bold vision but no market, and AMD competed effectively for awhile but has somewhat lost the plot in terms of price:performance. It's possible that other chips could have competed if not for Intel's rebates, but given the almost total failure of the other companies to establish mindshare, it's likely that AMD was the only company substantially hurt by Intel's rebate program. OTOH, perhaps I'm wrong, and that's a question for the courts to decide, which is what they're doing. The simple fact that these companies failed, however, is not prima facie indication of wrongdoing by Intel.
Cyrix was the first one to leave the market. The reason was lack of interest from major OEMs and Intel's shenanigans, locking up Cyrix in court over supposed patent infringement (which was proven to be baseless, but at that time, Cyrix ran out of funds).
You're right that their example is not indication of Intel's wrongdoing. However, it's a very visible symptom. Via offered nice, clean and power-efficient x86 CPUs and pushed interesting new form factors. Left to their own devices, they could have carved a niche in HTPCs, low-power desktops and laptops as well as home servers.
No OEM picked them up, Via effectively disappeared years ago. While planning my first NAS box seven years ago, I was looking for them, and they were only available imported, in very limited quantities and with implementations that left a lot to be desired.
Transmeta -- I've seen only one laptop offered with an Efficeon CPU. It was interesting, but laptops were out of my financial reach at the time, so I never got one. Still, it got several reviews which rated it pretty well, especially for the price. One area where it stood out was power efficiency assuming light-average use. Trying to force it into heavy workloads and efficiency dropped off a cliff -- performance tanked and power usage shot up. They would have been excellent CPUs for low-power devices assuming they could compete on a level field.
As for AMD, they have the edge in price:performance and always had. Even when they offered $1000 CPUs, they were significantly faster than Intel's equivalents, and afterwards, they continued to offer higher performance at lower price points than Intel (where Intel continued and continues to offer a high-end consumer CPU for $1000, AMD had to give up that price spot for years now).
Even four years ago, which was AMD's worst year in terms of performance, it was still competitive with Intel:
http://news.softpedia.com/news/Blind-Test-Shows-AMD-Machines-Run-Better-than-Intel-s-248202.shtml
In the high-end comparison, it was AMD's first generation Bulldozer vs. Core i7-2700K.
In the mainstream, it was AMD's Llano APU.
Both of them were widely considered inferior to Intel's offerings in "professional reviews" based on benchmarks, and the first generation Bulldozer required several iterations for the highest spec version to be considered competitive with Core i3, maybe i5 if the reviewer felt generous.
<hr />
If anybody considered Intel to be an upstanding paragon of virtue in terms of fair competition, I have a bridge to sell you in London.
Cyrix = Compatibility problems and lack of determination saw them quit the game early of their own accord / on their own failures. They might have done well for themselves ... if they had not given up so easily.
Transmeta = Runs their proprietary instruction set, that no one uses. So emulates any other instruction sets, such as x86. Results always performed horribly on the first iteration of any code as it needed translating. This made their performance "unreliable" at best and highly unpredictable at worst. They would have had great potential had they ever done straight x86. Or had they improved their translation speeds.
Via = Cheap crap that was often buggy. Did the IT industry a big favor by giving up as it cut tech support overhead significantly to no longer have them ending up in supply chains. Many companies started blacklisting them simply because of the support costs of running Via chips, mobos, or graphics.
AMD = (Usually) good tech, but lousy management. Numerous times they held potential ... but their success always inspires them to shoot themselves in the foot. For example, when they held a significant performance crown over Intel (poor Prescott) they ____ on their best mobo chipset manufacturer (nVidia) by buying AMD and turning their epic success into a major nosedive by alienating a significant gaming segment. It's the most extreme example, but is typical of their inability to understand their own place in the market and build on that instead of _____ on it.
Intel = Big, bulky, Chipzilla. Often taking risks and making mistakes, has the business strength to survive them. By default the defacto standard of x86 simply by merit of surviving the market and adapting when needed. (Even if that adaptation is sometimes rather slow to happen.)
The market was not a result of Intel giving discounts to bulk buyers. (Seriously, who does not expect to pay less per part when buying in bulk, be it in the form of initial price or in rebates?)
The market was simply a result of the various natures of the players involved. Intel won simply because they did not give up, and because they knew when to grab onto something, and when to let it go.
As Kenny Rogers sang,
"Every gambler knows
That the secret to survivin'
Is knowin' what to throw away
And knowin' what to keep
'Cause every hand's a winner
And every hand's a loser
And the best that you can hope for is to die
in your sleep."
You won't see me crying for Cyrix, Transmeta or Via. Nevertheless, your assertions about AMD and Intel are simply untrue.
AMD didn't shit at Nvidia. Nvidia made crappy and buggy chipsets that a lot of people got burned on. Seriously, nForce? Go and die. And Nvidia never treated its partners equally. If AMD tried to continue to accommodate Nvidia's temper and willingness to screw everyone over, it would have brought down AMD ten years ago.
ATi was up for grabs and AMD needed to coordinate efforts on new chipsets. ATi's chipsets maybe were behind Nvidia's in performance when they worked, but were also stable and generally not buggy.
Oh, and could you enlighten me how this alienated a significant gaming segment? Yes, I know that AMD bought ATi after their lackluster HD 2000 series while Nvidia was in its glory days, but it didn't alienate anyone. Intel's advantage in CPU cores at that time played a much bigger role.
--
As for Intel, I disagree with the notion that they won through technological superiority. Bulk discounts are perfectly acceptable and justified.
However, Intel was caught paying bribes to OEMs to demand that they do not carry AMD CPUs in their lineups and threatened to void OEMs' negotiated discounts if they carried AMD.
That is neither acceptable, nor justified and Intel paid a billion dollars to AMD in an out-of-court settlement.
Oh, and this part?
Intel won simply because they did not give up, and because they knew when to grab onto something, and when to let it go.
Hillarious.
- StrongARM/XScale/PXA. Intel sold it off to Marvell. Stupid move.
- Intel740. A failed attempt to compete in the GPU market. A classic example of Intel turning everything it touches into turd.
- Itanium. They should have seriously taken it behind the shed years ago.
- Mobile x86 attempts. Another example of completely missing the boat.
No, Intel has absolutely no notion of managing anything that's not x86, whether it's giving it up at the right time or holding on to it despite setbacks.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
