Official announcement: Toshiba starts PS3 memory producton.

DaveBaumann said:
Seems to me that Democoder already said exactly what we were talking about:

DemoCoder said:
IBM does basic research before applications are ready, and money spent on that research ends up in multiple applications. Even money being spent on CELL at Austin is not solely intended for PS3, since IBM is reusing that work for HPF projects at Livermore.
Click to expand...
Click to expand...

Perhaps next time you'll read the entire thread before commenting. For example you would have seen this following it:

DemoCoder said:
What you say may be correct, the only information I have is about Watson/Yorktown, since that is where I worked, and I still know people there.
Click to expand...

And even if we assume that he's correct in what he stated (as I later did) - it's still considered a positive externality or 'spill-over' benefit, NOT a direct investment of the STI funding in non-PS3 fields as per the origional argument. If it's too much to understand, as you don't seem to have grasped it the first time I explicitly stated this, then I'm sorry. This is basic macroeconomics Dave.
 
Perhaps next time you'll read the entire thread before commenting. For example you would have seen this following it:
Click to expand...

Ummm, Funny that Vince, since the rest of the thread, and DC's further comments weren't written when you responded that what he talks about didn't relate to the conversation when clearly they did, but you just decided to selectively quote which you deride so many others for when you believe people are doing it to you. But, there you go - double standards don't apply to you do they Vince?
 
Enthusiasts, critics, skeptics, zealots, whatever ... the level of interest for itself is a success. Every tiny detail dissected, analysed, revalued ... IMHO even this situation is noteworthy.
 
This is blatently wrong. I do appologize as I've furthered this incorrect statement by responding to it as I did, which was my mistake.

In 2000, they projected the lithography nodes they'd utilize as 100nm, 70nm, 50nm.

In early 2002 they amended this to 90nm, 65nm, 45nm. And announced the worlds smallest eDRAM cell @ 65nm, with a follow on their dual-loading 65nm process with a fully SOI 45nm one.

If anything, Toshiba and Sony are ahead of the game - not behind which your stating. Although, the differential between the 2001 projections and the 2003 actual are very small as 65nm & 70nm are basically equivalent in practicle terms.
Click to expand...

I am not arguing about that.

Let me word it the another way, In 2000, they overestimate the capability on 70nm with regard to eDRAM density. So once again, I wasn't talking about the process, but its capability.

Again, we've already made the distinction between Cell the architecture which may well have 90nm parts for some niches and the Broadband Engine which is 95% probable to be based on the 65nm process. You can speculate all you want, but it flies in the face of what's known.
Click to expand...

What's known ? Its known that Broadband Engine is 4 Processor Elements group together. What they were designing was this Processor Element. And they start this design for 100nm process. As better process available they move on. That's the basis of this scalable architecture, design the basis and repeat.

You're avoiding the statement. He never stated that the eDRAM is the only and main system memory - which is what you're implying.
Click to expand...

He said the memory for the main processing element. As opposed to the memory for graphic processor or other sub elements.

All he said is that it's possible for use within a PE at sub-100nm scales -- something which you're extrapolating to mean it's the "main" memory.
Click to expand...

Yes, and the patent cover this "main" memory in more detail. Even the title of the patent cover it. What we know now is that this memory was not eDRAM at 100nm process. Though at 70nm or 50nm its possible to use eDRAM.

What I am saying is his comment maybe base on what he thought 70nm was capable, this capability was done through projection. Now they have hard data on the capability of 65nm. And from that data the eDRAM for 65nm is alot less dense than what they thought what 70nm process was capable.

So given the hard data in 2003, he may or may not change his assesment on the subject how practical it is to use eDRAM as the memory for main processor element at 65nm process.

So at this point the memory that the patent describe is not certain to be eDRAM.
 
Re: ...

Deadmeat said:
Demo, was this the man at the meetings that would have crazy visions of Teraflop devices in homes?(Million of homes)
Click to expand...
No, Kutaragi probably stole the idea from the crazy guy.
Click to expand...

Now I understand... DGMA's doesn't just hate Kutaragi because he came up with the "programmer's hell" that is CELL, but he also hates him even if he didn't come up with the idea. He simply hates Kutaragi, CELL or no CELL.

Life makes sense now...
 
I honestly don't rememeber what the crazy pervasive guy looked like except that he was caucasian. But no one really deserves credit. People at Sun and Xerox were also thinking up the same ideas around the same time.

Cell was an outgrowth of work at Yorktown on "healable software". The idea that you can run software on faulty hardware, by dynamically modify the SW so as to route around problems. Similar to the way you mark bad blocks on HD sectors. First IBM applied to this to anti-virus technology, but sticking "probes" on your HD and waiting for them to get infected. After they got infected, the EXEs were compared against trusted versions on IBM's servers and an "antidote" was generated automatically just for your PC. The "probe/fix" method was extended to "fix" software running in RAM whose bits got flipped by bad CPU/memory. Cause if you have an architecture like Blue Gene with 1 million processors, on a given day, some of them will probably fail.

When I say outgrowth, I mean, people had always considered mega-CPU projects, but they always had significant operational maintainence problems because as the size grew, so did the probability of failure. It wasn't until IBM decided to say "ok, we'll let some percentage of CPUs fail, but we're correct errors caused" that a project like Blue Gene became plausible. That is, the idea of self-healing machines was the enabling architecture to scale past a few thousand CPUs.
 
...

but he also hates him even if he didn't come up with the idea. He simply hates Kutaragi, CELL or no CELL.
Click to expand...
One of my hypothesis is that Kutaragi Ken got the idea of CELL during his visit to IBM circa 1999 and decided to be the first to commercialize the technology in the consumer market. In other word, Kutaragi didn't come up with the idea of CELL, like Steve Jobs before him 20 years ago, Kutaragi decided to become the first commercializer of somebody else's idea/technology. Only difference is that Kutaragi is actually paying a royalty to IBM for CELL architecture, whereas I don't recall Steve paying off Xerox for Mac GUI.
 
Re: ...

Deadmeat said:
Only difference is that Kutaragi is actually paying a royalty to IBM for CELL architecture, whereas I don't recall Steve paying off Xerox for Mac GUI.
Click to expand...

Well this part is a troll, if I ever saw one! (Specifically, the part about ole Stevo) ...not that I am expecting any action to take place... :?
 
:rolleyes: Yes, that must be what happened...

Anyways, I just wanted to throw up a flag for whom it should concern, not even dignify the remark with an actual corrective reply (I mean, what's the point, really? This forum is certainly not about talking about the history of Apple, whether or not fallacious statements have been posted to the record.).
 
Only difference is that Kutaragi is actually paying a royalty to IBM for CELL architecture
Click to expand...

Only SONY AND TOSHIBA DO NOT HAVE TO PAY ROYALTY'S TO IBM FOR USING CELL

Seriously, this is getting out of hand, we have told you this TONS OF TIMES already and you just ignore it. You blow us off EVERY TIME, your a joke, a troll, someone who should be removed.
 
randycat99 said:
not even dignify the remark with an actual corrective reply
Click to expand...

It's not. But I'll toss him a bone or two on the off-chance he actually wants to look at facts now and then, as opposed to regurgitating whatever part of popular tech mythology he likes. It's not like the information is hard to find or anything...

Sorry for the distraction folks, but that particular topic just sticks in my craw. It's so dumb and yet it... never... dies...! :oops: When one starts trying to base an arguement or make analogies off it, it's just time to make with the smackdown.

Well, Kutaragi is at least giving the credit where it is due, whereas the Xerox researchers didn't see a dime from Mac and Windows...
Click to expand...

Because researchers in big companies see direct profit? Xerox management rather liked the deal being made, and if one wants to complain about poor management decisions, the line forms WAYYYYYYYYYYYYYYYYY back there. Meanwhile, Mac would have still shipped their GUI, and Xerox would have paid just as little attention to PARC as ever. (And funny thing is, without Jobs' visit, PARC's contributions would likely have remained just as unknown and unseen as all the others to have fallen by the wayside [Atari and Amiga being the biggest--and in my opinion best at the times--and even they are little but curiosities if remembered at all, or sighed at longingly by former owners] )

We now return you to your regularly scheduled post. :p
 
Jeez.. talk about sensetive. This is just hardware speculation people.. nobody is making fun of your dog.
 
Jeez.. talk about sensetive. This is just hardware speculation people.. nobody is making fun of your dog.
Click to expand...

You tell that to Deadmeat, Ken Kutaragi's limo ran over his dog; I would be pissed off too.
 
Enough. Or i will lock this thread. Behave people. Behave. Now if you don't mind its time for dinner and then more poker.
 
DemoCoder said:
Cell was an outgrowth of work at Yorktown on "healable software". The idea that you can run software on faulty hardware, by dynamically modify the SW so as to route around problems. Similar to the way you mark bad blocks on HD sectors. First IBM applied to this to anti-virus technology, but sticking "probes" on your HD and waiting for them to get infected. After they got infected, the EXEs were compared against trusted versions on IBM's servers and an "antidote" was generated automatically just for your PC. The "probe/fix" method was extended to "fix" software running in RAM whose bits got flipped by bad CPU/memory. Cause if you have an architecture like Blue Gene with 1 million processors, on a given day, some of them will probably fail.
Click to expand...

Ahh, now see, if you stated this to begin with I would have agreed. Before I comment on this, it's still not relevant to my argument with Dave as it's an externality that's "spilled over" into the Cell project - as opposed to the "spill out" which we were discussing that would support his BS stance.

Anyways, IBM's Autonomic Computing Initiatives (like self-healing, self-optimizing, self-configuring) has definitly influenced projects on the scale of the Blue Gene's, I think this goes without saying really. What I question is it's place [it being the self-healing you mentioned, not the other facets of ACI] within a single IC on the scale of any given Cell manifestation, the Broadband Engine being one of them. Personally, I feel the concept of self-healing within a set-piece IC in a console as ridiculous. If there is a fault, you buy a new one or send it back to Sony - not compromise the development system in a truely futile attempt to mask a fault within a closed-box enviroment.

Yet, you're not wrong. Hell, I agree with you for the most part but as I mentioned before on a diffent scale. I stated:

[url=http://www.beyond3d.com/forum/viewtopic.php?p=206838#206838 said:
Vince. Pg 4[/url]]The architectural differences are obvious on a per-IC level. If you wish to theorize about how the concept of non-local cellular processing is analogous to how the IBM projects work on a macroscopic scale, then I'd agree based on what Kutaragi has stated. The way in which interconnection and processing tasks look to be distributed is very much akin to the BG projects, which arguably, are biologically inspired. Yet, this isn't an architectural feature - and as you very well know, a analogous tasks don't imply similar constructs underlying it them
Click to expand...

I see the Cell project borrowing heavily from the research projects you mentioned (such as the Blue Gene permutations) not in an sub-IC level but rather in their inter-IC communitation and distributed processing ability - if it is realized as the form of Cell Computing that's been discussed by Kutaragi and in the Suzuoki Patent. It's only here that you start to see where the benefits of IBM's ACI technology when you consider a future in which a Cell network/processing fabric exists. One in which digital content (be it media or data) is streamed between Cell's and you can imagine the loss of singular IC/Cell's from the fabric to be a common occurance. Only the technology you talked about can allow this non-local/Cell computing paradigm to be realized feasible.

But, the leap you didn't make is that BlueGene is analogous to a STI Cell Network, not a single IC. Just as a BlueGene 'system' can exist in many manifestations, invarient to the actual constructs underlying a single node (be that Power4's, or the reductionist SMT cores @ Yorktown, or the lastly more powerful STI Broadband Engine's from Austin), that same ideology applied here.

So, when you talk about STI Cell - you need to realize that the basic building blocks that scale up for use in the PS3 which Austin are designing may never be actually used in a Cell network thats analogous to a BG. Rather, they may just be another Power on steroids.

PS. I asked if they had an RANN research because of your stance on consciousness and your blatent bashing of Roger Penrose in that one thread which I forgot about when I went away for the weekend. While I don't agree with everything he says, at some level of abstraction he is vindicated if it pans out.
 
Yes, well we're used to that. 'twil be The Year of the Demos, tho! Hehe... And the "more released specs" to shift all the rabid supporters/naysayers of their consoles of choice into high gear. Myself, I just want more stuff to drool over. ^_^
 
You must log in or register to reply here.
Back
Top