Originally posted by: Ice9
They don't care about innovation. They're willing to accept incremental changes. They are simply not prepared for revolutionary ideas and breakneck speeds. It's just too much for 'em.
You're so full of it, I'm surprised that you aren't choking on the brown stuff by now.
How can you attempt to tar anyone that believes that Rambus isn't somehow the "second coming of technological innovation" as you seem to paint them as, with a huge brush that those people are somehow "anti-innovation". Excuse me, but some of us would prefer to patronize
ethical tecnology companies. It's kind of like SCO claiming that Linux users/supporters, are "communists", and outright "thieves". Well, you just put yourself soundly in the same camp as SCO's PR people, congratulations. I hope you're happy with that.
Originally posted by: Ice9
I think if you asked the average PC-centric person in here, they'll say that "Cell" is just a waste of time and will never get off the ground. Just like what mainframers said in the early 80's as their jobs slowly disappeared.
For the record, I think Cell is a great idea, decentralizing and localizing processing needs, and allowing a connected network of nodes with processing capability, to act more or less like a coherent whole. In fact, I have some operating-system/storage-model designs that I've worked on privately in the past that would support such a model.
Originally posted by: Ice9
I know what I want to see in the future. And it doesn't involve Micron, Infineon, Hynix or anything that has "JEDEC" support unless they are willing to embrace the better ideas that are out there.
The technology world has grown faster than any standards setting body could ever keep up with. It takes a lean and mean company like Rambus to innovate. You can either accept that they have the better idea and make it happen, or you can be like the scores of people here who are willing to accept that minimal incremental change.
That's a major factor here that you seem to be ignoring - mfg costs. You yourself said that Rambus isn't a mfg - they likely don't tend to therefore see the costs of radical change vs. incremental change that a mfg would encounter. Most of the recent incremental "innovations" in the DRAM technology sector, have given us nearly the same level of performance increases as "radical" innovations would have, but at a lower cost to mfg. That's actually rather important, for a mass-market commodity part like DRAM.
I don't believe that JEDEC would outright reject any sort of radical design changes,
as long as the mfg costs could be kept low enough. That's the primary key, because that organization serves those mfgs, and that's a bit consideration for them.
Originally posted by: Ice9
Rambus RDRAM ran at blistering speeds for its time and had EVERYTHING going for it. But a single word was able to poison the well. "Latency". Everyone went on and on about how "latent" RDRAM was in theory - yet in practice it blew the doors off everything in its timeframe.
No, "in practice", it had horrible
real-world latency numbers, that actually got
worse with larger DRAM arrays.
You're living in a dream world, if you truely think that it was somehow faster for real-world applications. As a matter of fact, Rambus was architected original, to
cut costs, by using a low-pin count, but ultra-high-speed interface, in order to cut mfg (packaging/testing) costs. And so it did find some design wins, in console systems like the N64 and PS2. I don't have any problem with that. That's RDRAM's niche market. Not mainstream x86-based PC memory, because access patterns from running x86 code simply don't mesh well with RDRAM's technical attributes.
Not to mention the thermal issues - SDR/DDR tends to spread access patterns around nearly evenly across the chips on a DIMM, RDRAM did not, and thus
required heat-spreaders. Nevermind the requirement for C-RIMMs, being inconvenient for upgraders, and more costly for mobo makers.
Originally posted by: Ice9
And like you pointed out, these companies outright ADMITTED that they were trying to kill off RDRAM and subsequently Rambus. But the well was so effectively poisoned by this smear campaign, people made up new reasons every day for why they should hate Rambus as a company.
How is poor technical performance in one area (while excelling in another), a "smear campaign"? (And I thought that you claimed that the other memory mfgs colluded to kill Rambus, through market manipulation of prices, now it's a "smear campaign"? Huh?)
Originally posted by: Ice9
They criticized their royalty rates - when in practice RDRAM royalty rates were the lowest of nearly all technologies in the memory industry.
But
any additional royalty payments, are going to be of greater cost to the mfg than zero royalty payments.
Originally posted by: Ice9
They criticized the cost of manufacturing - and even Samsung admitted it was cheaper to manufacture RDRAM than it was to manufacture SDRAM (not even DDR!).
Were they speaking of individual chip costs, module costs, or overall system costs (including such things as C-RIMMs, etc.)? As far as *raw*, per-chip costs, excluding royalties, chips with less pins generally cost less in terms of testing/packaging. That's well-known in the industry. I'm guess that's what they were speaking of, and that you chose to take that quote out of context. I can't imagine it being the case any other way.
Originally posted by: Ice9
They branded Rambus as a company that sues for profit - when they never received a single award from any lawsuit (yet).
But if the reason that they sued, was to force the target company of the lawsuit, to capitulate and pay licensing fees, how does not recieving an award from the lawsuit itself, in an way invalidate the premise that the purpose of the lawsuit, was for Rambus to eventually profit from that action? It doesn't. Again, check your logic, please.
Originally posted by: Ice9
They claim that Rambus "stole" from JEDEC, when in fact it was the companies that make up the JEDEC consortium who stole from Rambus.
Are you basing that statement, solely upon the fact that Rambus, apparently, indeed was awarded the patent(s) in question covering the technology?
As far as Rambus "stealing", please refer to my prior empty-lot/car-lot analogy.
Originally posted by: Ice9
Oh yes, this shoe has been on the other foot the entire time, but many people on this forum cannot be persuaded by the truth, even when it's there in plain text. They simply MUST hate Rambus. Yet many of these companies regularly root for an underdog in the industry named AMD. It's a double standard I will never understand.
Uhm, because AMD doesn't pull unethical crap like Rambus (allegedly) did? AMD didn't patent something, already being designed into competitors products, and then go around and attempt to sue them for it? AFAIK, AMD even put HT under control of another organization, and they freely license it, just like PCI.
Originally posted by: Ice9
People HAVE to hate Rambus as a company because these little trade rags told them to.
Or perhaps we're not just all a bunch of blind fanboys, and shareholders, and that we don't believe that just because something is technically legal, that it automatically makes it right or not unethical.
Originally posted by: Ice9
They never saw the fire in the theater, but since several people yelled "Fire", they all flocked out of the Rambus theater as fast as they could assuming they were about to be burned.
Speaking as a consumer only, at this point - when Rambus-based systems came on the market, and Intel was pushing for a Rambus-only world, and yet, prices were higher, but performance was no better, and in some cases worse - that's when I decided that an RDRAM-based system wasn't in my future. It's much the same with Intel's early BTX/LGA775/915-925/DDR2 transition. No real performance increase, only an incompatibility and cost increase.