[warubozu]

Damn, that's a huge fine their facing:

http://news.bbc.co.uk/1/hi/business/4340388.stm

 

[jpr703]

That fine may sound like a lot of money, but I'd be shocked if the extra revenue they took in from the price fixing doesn't exceed it!

Fines in the US tend to be just a cost of doing business. Cheat until you get caught, accept the fine so that the politicians can look good and pocket the difference between what you could have made honestly and what you made cheating minus what you had to pay out in fines.

Huge corporations like Samsung aren't stupid and wouldn't jepordize their bottom line. They know how to work the system and they wouldn't do this stuff if those kind of penalities really hurt them in the end.

 

[cerbie]

Samsung was one of the smart ones. They played both sides. Micron and Infenion, FI, just plain got smacked down over it.

Since the article is crap, here's what happened (there's basically this and the bad caps for computer industry scandals in the last few years

):

Rambus came out with RDRAM, which, for non-notebooks, was superior to the upcoming DDR. It was cheaper to manufacture, ran very fast, and only had small latency problems, which were not important once it got up to speed. It had current leakage and bus issues, but those have been dealt with with duture revisions (Rambus RAM is popular for non-PC stuff, such as the Playstation series of game consoles, and many cell phones and embedded devices).

Intel offered to foot the bill to retool factories to make these chips. The big manufacturers, save for Samsung and Elpida, refused, because they had to pay a royalty (even though it would end up much cheaper later on) to Rambus.

So, they conspired to sell SDR SDRAM and DDR SDRAM at under cost, claiming the competition was doing so, and for awhile, ti looked genuine. This set them up for corporate welfare help, which got them by for several years, as they flooded the market with cheap SDRAM, keeping RDRAM at high prices, and keeping XDR (the follow-on to RDRAM, that is much better in terms of latency and power use) from getting a foothold in the personal computer market. Because of this, RDRAM is still extremely expensive.

Ultimately, it hurt consumers in several ways, not the least of which being a slower increase in performance, and no major decreases in the price of the technologies used. It took Intel over two years to get a chipset out that could beat their RDRAM-based i850E. Because of this whole stink, AMD never bothered to support RDRAM in any way (SiS tried, though).

But, Rambus, a company of mostly lawyers, got a loophole and tore it open. The current SDRAM, now DDR SDRAM, and possibly DDR-II SDRAM, used technology similar to what Rambus discussed a few years earlier with JEDEC (an open standards body for the DRAM companies). After a few court cases, they got the precedent they needed, which came down to the definition of a memory bus in terms of their accepted patents.

They started suing because of this, and in doing so, found what they needed to get the FTC and such to start looking into the various RAM companies' less-than-honest involvement with the technology. Several manufacturers have had to pay Rambus back-royalties, because of this, and some additional damages due to price-fixing (the price-fixing cases have gone on longer).

 

[cire]

Quote:

Originally Posted by cerbie Rambus came out with RDRAM, which, for non-notebooks, was superior to the upcoming DDR. It was cheaper to manufacture, ran very fast, and only had small latency problems, which were not important once it got up to speed. It had current leakage and bus issues, but those have been dealt with with future revisions

RIMM was fast on paper, but in real life performance it was slow. they used the "quad-pumped" FSB netburst tried to pull off and failed. it overclocked like crap, it had very high latency (too high to be used efficiently with AMD chips at the time), had a completely different archetecture that probably wouldnt go to well with AMD's memory system and was expensive. the cheapness of DDR1 and SDR couldnt have been all attributed to this unfair trade, as SDR was pretty mature and had already been produced for a few years and DDR1 was just a modification of SDR.

 

[marvin]

Yes, RDRAM was slow for AMD, but that's because AMD favors low latency over high bandwidth, which is the opposite of what RDRAM did well.

RDRAM was perfect for the P4 platform because the P4 need obscene amounts of memory bandwidth for it's quad-pumped bus. Without RDRAM, Intel performance dropped a lot until they brought dual channel DDR solutions out.

For Intel at least, RDRAM was the superior choice.

cire:

It doesn't matter if DDR is just an offshoot of SDR. RDRAM and DDR were produced on the same processes, and RDRAM did not have significantly larger die sizes, rejection rates, or packaging requirements. They should have been price competitive, but weren't due to collusion. Yes, there's the small matter of licencing fees, but that doesn't come close to the price differential between DDR and RDRAM.

 

[pbirkett]

Personally, I'd like to see Grado fined for their ridiculous european prices - i mean come on... nearly $1300 for RS1's, thats disgraceful, especially when some of you guys complain they are overpriced over there too...

In fact, theres quite a lot of companies i'd like to see fined for ripping us brits off - we seem to be the whipping boys of the world when it comes to being ripped off.

 

[RedLeader]

yea, what's up with the UK being so bloody expensive?

 

[cerbie]

Quote:

Originally Posted by cire RIMM was fast on paper, but in real life performance it was slow. they used the "quad-pumped" FSB netburst tried to pull off and failed. it overclocked like crap, it had very high latency (too high to be used efficiently with AMD chips at the time), had a completely different archetecture that probably wouldnt go to well with AMD's memory system and was expensive. the cheapness of DDR1 and SDR couldnt have been all attributed to this unfair trade, as SDR was pretty mature and had already been produced for a few years and DDR1 was just a modification of SDR.

RDRAM was very fast in real life. Even the old Willamettes feel speedy on 850 boards. They did have high latency, but the P4 was made for high bandwidth, and loved it. Once RDRAM hit 1333MHz, it was at a point where it was competitive in latency, as well, though still not great for mobiles.

Also, how was NetBurst a failure? AMD had a slight edge early on, then the P4s had the best performance around, right up until the A64s. It changes when you start counting in value (the P4s still suck for the money); but as the technology itself goes, it was a success until the Prescott stopped ramping.


The cheapness of SDRAM already has been attributed to unfair trade, in the US, by the FTC. Rambus got their wrists slapped, but they were in the (legal) right when it came down to it. Current prices are fine, but the prices from around 00 to 02, where they dropped like a rock, have been proven to be due to anticompetitive practices.


I think you guys over the pond pay so much because someone thinks everything is dollars, and if it isn't, it should have the same number beside it.

 

[cire]

Quote:

Originally Posted by cerbie Also, how was NetBurst a failure? AMD had a slight edge early on, then the P4s had the best performance around, right up until the A64s. It changes when you start counting in value (the P4s still suck for the money); but as the technology itself goes, it was a success until the Prescott stopped ramping.

netburst is pretty fast, but really its hellishly inefficient, therefore a failure in my mind. northwoods had 30-40% current leakage, which is pretty crappy and prescotts going up to almost 60% leakage. netburst just ended up being a intel's hackjob; lets keep sticking on more pipe stages until we can scale the chip even further, with no regard for power consumption! seriously though, if intel Isreal can alone manage to modify the P3 design into an amazingly powerful, but at the same efficient with clock cycles and power, then what was the netburst team doing?

 

[nysulli]

well if were gonna talk about p4 desgin flaws, read this article, and keep in mind that the pentium M is heavily p3 based in its design, and you'll really see how flawed the whole design of the p4 is

http://www.tomshardware.com/cpu/20050525/index.html

 

[cerbie]

Quote:

Originally Posted by cire netburst is pretty fast, but really its hellishly inefficient, therefore a failure in my mind. northwoods had 30-40% current leakage, which is pretty crappy and prescotts going up to almost 60% leakage. netburst just ended up being a intel's hackjob; lets keep sticking on more pipe stages until we can scale the chip even further, with no regard for power consumption! seriously though, if intel Isreal can alone manage to modify the P3 design into an amazingly powerful, but at the same efficient with clock cycles and power, then what was the netburst team doing?

The folks working on it were trying to go faster than they could. Prescott was supposed to get to near 5GHz, when the next gen would come in. It's clear they could have done better, too, given the dual-cores (similar heat output to single-core), but AMD beat them squarely there (though once again they are having issues with keeping up to demand). The real trouble with it is they based its success on everything going as planned (oops), and, as always, it did not.

Between all of that and the ICH4, Intel has not been in the good graces of many a techie for the last few years

, in addition to SiS and nVidia making very solid chipsets for the Dark Side.

 

[cire]

Quote:

Originally Posted by cerbie Between all of that and the ICH4, Intel has not been in the good graces of many a techie for the last few years , in addition to SiS and nVidia making very solid chipsets for the Dark Side.

dark side...says you...

kinda pathetic really...AMD can't come up with their own good chipsets. though the dothan really is one amazing piece of silicon, i'll give them that.

eww...tom's hardware...that dude is full of **** a lot of the time....

 

[cerbie]

Yes it's the dark side. The stuff on the dark side is cooler, and there are fewer on the dark side. So, it fits

.

AMD can, and has, come up with their own good chipsets...but they don't want to step on their partners' toes, so only grudgingly let them be used on workstation and server boards (my second Athlon board was a 760), and don't update them very often. That was a major problem early with the Athlons, and why I went from a KT133A to 760/686B (I didn't have the money for a full 760 board and RAM at the time, so put up w/ the VIA southbridge and left it free of controller cards and SoundBlasters

). I also wouldn't be surprised if AMD would have a lot of trouble keeping up with the demand, if they were the primary chipset for their CPUs.

Starting with the 735, though, SiS was basically head and shoulders above everyone else (they're server-worthy, without a doubt), with nVidia finally getting the nForce2 in there, as well (I wish someone like DFI would make SiS-based boards, or someone like Iwill, that does, would market them for retail). It's less of a concern with the A64s, of course.

 

[nysulli]

Quote:

Originally Posted by cire dark side...says you... kinda pathetic really...AMD can't come up with their own good chipsets. though the dothan really is one amazing piece of silicon, i'll give them that. eww...tom's hardware...that dude is full of **** a lot of the time....

i know i know, but once in a while they don't get to biased, and everything else i've ever read based on the pentium M leads to the same conclusion, also try to find some old school comparision charts between a 1gig p3 and a 1.5gig p4, its scary how closely matched they were when the p4 first came to light, well that and how badly a 1.4ghz athlon spanked the 1.7 p4 at 1/5th its cost

 

[jlo mein]

Quote:

Originally Posted by pbirkett Personally, I'd like to see Grado fined for their ridiculous european prices - i mean come on... nearly $1300 for RS1's, thats disgraceful, especially when some of you guys complain they are overpriced over there too... In fact, theres quite a lot of companies i'd like to see fined for ripping us brits off - we seem to be the whipping boys of the world when it comes to being ripped off.

Is $1300 the MSRP of Grado for europe? Or just what stores actually charge you.

It could be because of tariffs/duty/quotas. I know when I buy things from UK to Canada, the government slaps duty on me for not buying domestically.

I think in north america we take (esp canadians) NAFTA for granted. I don't think the USA has any sort of free trade agreements with the UK so that could be why grado's are so expensive (to encourage UK people to buy UK/european headphones).