<strong>It has been reported that Playstation 3 due out in 2004 will have some form of "distributed computing"
If IBM and Sony are doing R&D could Apple be doing the same?</strong><hr></blockquote>
Last I checked only 1 in 125 console game systems had access to any kind of newtork. This means that for every ONE dreamcast plugged in there are 124 others that are NOT. Would parents really put in another phone line JUST for the kids xbox? Not likly. Just as rediculous, do you think the developers have some other form of networking up their sleeve? Is DSL equiped PS3s on the horizon or wireless xbox2s? This is baloney. Even the most trying games for the computer don't use distributive computing yet. How much can you improve gameplay by spreading the task across a giant dialup network of xboxs full of latency that you could not ALSO accomplish with simply putting a better processor/more ram/more disk space/faster optical drive in the unit itself. It's not like mario is decoding genetics here. It's a GAME and very few games require a bunch of slow 'crunching'. I bid that this is a much a dead end as the wolf topic in the first place. Wishful thinking from a generation of 'net bunnies' who think EVERYONE has broadband and EVERYONE buys the newest console every year. Simply not the fact jack. Mr David Becker of CNET can take his article and put it back where it came from. (his anus) Nice fantasy article though. I particularly like the bit with the interview with Okamoto when he said "Maybe the PlayStation 6 or 7 will be based on biotechnology". LOL. Great sci-fi wonder of a technical genius master jedi that he is, I have no doubt that sony will continue to push the envelope on console design and interface options but this is not a reality in the next couple years. Wait until standardized universal access broadband penetrates a solid 25% of US households and THEN you can talk about having your Nintendo be a slave for other gamers when you aren't on it. Speaking of which, my tamagochi is crying...I have to go feed him. Later!
The author is not writing wishful thinking, if anyone it is Sony and IBM. They have stated that they are spending time and money on exactly this.
"Okamoto said Sony is working with IBM to apply Big Blue's research in "grid computing," a variation of distributed computing, to the next PlayStation."
You aren't thinking in general enough terms, I suspect. In this case I believe they are talking about distributing computation across many processors as opposed to across a traditional network. The "network" in this case is likely to be in one box, or even on one piece of silicon.
You know, I can run Linux on my PS2 already... so let's not keep this discussion strictly to games.
It's a very solid idea... it doesn't suprise me to hear it from anyone anymore. Have read all the naysaying on the Wolf thread, but it seems most have the application of the technology in reverse.
Think about distributed computing... a-la SETI. Very popular... and everyone who's involved is donating their processor to the cause. Now in an office environment, that can get nasty - not in the respect that you can't or shouldn't get the cycles, but in that the bean counters simply don't like to give things away. So if you have a large enough group of computers, you can rent/lease out mass cycles on the open market. This isn't just happening today, this was happening yesterday for Jevus' sake.
Couple of angles, then: 1) Cost of ownsership can begin to be re-evaluated if the machine itself, with little interaction from the user/owner, can broker processor cycles on a net-based market and produce income. You don't need to look much farter into the future than last week to know that you can already sell your excess power (they HAVE to buy it, actually) to power companies if you roll your own at home and can create an excess beyond usage. My kids leave the damn PS2 running day and night - being able to sell cycles would certainly help to offset the cost of ownership and usage. 2) Grassroots movements, relying on the donation of cycles, are suddenly going to have access to more types of equipment/processors/platforms at lower and lower costs. Standards will begin to emerge. Anything that brings the sheer power of tens of thousands of G4s, Athlons, PS3 - whatever - to bear on a computationally deep or intensive task and puts that power in the hands of anyone who needs it is a good thing - it the kind of thing that means I may actualy get a damned flying car sometime before I die, and that's an excellent thing, because I want my damned flying car.
EDIT:
I believe Wired had a sneak peek at a testing unit from Sony a while back - a "concept car" of playstations if you will. It had something like 16 Processors... insane framerates, etc. If I can find it I'll post some fo the info... may be an indicator.
I don't think that the network is in one box. They aren't talking about multiprocessor computers. They call it Distributed computing. There is a link where they talk about things like SETI. It seems to me that IBM may be working on the technoligy that makes it possible to share CPU power over broadband for everday processing. It doesn't sound so far out that the speed of your computer depends on how much community processing power you are willing to buy.
It was said in the wolf thread that computer in urban areas, or places with more Macs, would have more processing power. It would be a self perpetuating market the more Macs that sell the fast they become. This article makes me think it isn't so far off.
Project "Wolf," as described, sounded like clustering. Clustering is not distributed computing. Distributed computing breaks a problem up into lots of little chunks that can be solved discretely; clustering unifies two or more pieces of hardware so that they can tackle a single, large problem (such as running a database). Clustered machines are all in the same room - usually on the same rack - and hooked up to each other with the fastest, lowest latency connections available (e.g., 50-1000 times faster than a cable connection). Distributed computers on the other hand can be flung far and wide, and they don't always (or even usually) require a fast, always-on connection to the other machines.
Thus, the Playstation model and "Wolf" (as described in the relevant thread) are two completely different things. If Apple adopted both, they would impact different product lines, and solve different problems: A distributed computing service would impact their desktop and laptop lines, while clustering would primarily impact their server line (meaning XServe and the products to follow it).
As to the truthfulness of the "Project Wolf" rumor, I can only point out that the XServe is just begging to be clustered, so I would not be in the least bit surprised to hear that Apple is working on this.
I dunno. You still can't sell me on the idea that every house is now going to have a console gameing system which is leasing out (or donating) it's processor etc to some stranger over a traditional dialup line or dedicated line to help render some other little kid's Final Fantasy 12. It just doesn't make sense. To have networked games is fine. To have games that 'keep playing' after you walk away is fine. But SETI and this are like apples and oranges. Seti sends you a chunk, you decode, and then send back. How can that help the 95% of games which are instant response 'point click bang' type games?
As for the this is in 1 box... fine and dandy. Make the sucker collaborate with itself. Stick a dual processor in that sucker. Amoph is sure right about the clustering thing vs. the colaborative computing thing. This topic shouldn't really involve wolf.
In summary. PS3 having newtork support? Of course. Having games which swap info? Sure. Having a couple hundred thousand PS3s sharing resources to make a mega processor? Not likely. It sounds too much like.. "By your powers combined, I am Captian xBox!"
Amorph your answer doesn't help. I'm guessing that IBM and Sony aren't lieing about this. They are even claiming to impliment it in 2 years. The question than is how?
IBM invented FUD, which is the tactic of making early claims of incredible technology in order to preempt other companies' announcements and introduce uncertainty about their products (Microsoft and Intel are also masters of this tactic - Win XP still hasn't lived up to some of the early hype for Windows NT 3!). It works: A lot of people on here bashed Moto for failing to come anywhere near IBM's 1GHz G3 almost a year ago - but IBM had merely touted it in a press release. They still haven't actually released it, and in the meantime Mot has cleared 1GHz with their more complex G4. But IBM still has the better rep as a PPC provider.
So, again: Don't believe the hype. There will probably be an improvement for those games that are amenable to being distributed, such as massively multiplayer games. Unless IBM and Sony are thinking of rolling Internet 2 into the world's living rooms in record time, though, that's probably all it will be, and it probably won't offer a thousandfold improvement even in the best case. It's not a matter of the technology in the console, it's a matter of the narrow, fragile, high-latency bandwidth connecting the consoles together across the internet - and that's outside of Sony's and IBM's ability to change.
We'll see in two years. I won't be in the least bit surprised if their claims become more modest as the project's deadline gets pushed farther and farther back, until the shipping product - impressive, undoubtedly, but not mindblowing - finally appears.
[QB: A lot of people on here bashed Moto for failing to come anywhere near IBM's 1GHz G3 almost a year ago - but IBM had merely touted it in a press release. They still haven't actually released it, and in the meantime Mot has cleared 1GHz with their more complex G4. But IBM still has the better rep as a PPC provider.
You are right , but you must remember also that IBM create the power4, one of the most effecient chip everbuilt, and more the more extensible architecture. So the better reputation of IBM has much more to do with the powerchip, than the early announcement of a 1 ghz chip.
No kiddin', Amorph. Does anyone remember the last time a relatively complex tech product did NOT ship later than promised by its company? You'd think they would have learned by now. <img src="graemlins/oyvey.gif" border="0" alt="[No]" />
Comments
<a href="http://news.com.com/2100-1040-866288.html"; target="_blank">http://news.com.com/2100-1040-866288.html</a>;
<strong>The article is at:
<a href="http://news.com.com/2100-1040-866288.html"; target="_blank">http://news.com.com/2100-1040-866288.html</a></strong><hr></blockquote>;
This is one of the reasons I keep going on and on and on about the need for a "better way" to program vector units.
<strong>It has been reported that Playstation 3 due out in 2004 will have some form of "distributed computing"
If IBM and Sony are doing R&D could Apple be doing the same?</strong><hr></blockquote>
Last I checked only 1 in 125 console game systems had access to any kind of newtork. This means that for every ONE dreamcast plugged in there are 124 others that are NOT. Would parents really put in another phone line JUST for the kids xbox? Not likly. Just as rediculous, do you think the developers have some other form of networking up their sleeve? Is DSL equiped PS3s on the horizon or wireless xbox2s? This is baloney. Even the most trying games for the computer don't use distributive computing yet. How much can you improve gameplay by spreading the task across a giant dialup network of xboxs full of latency that you could not ALSO accomplish with simply putting a better processor/more ram/more disk space/faster optical drive in the unit itself. It's not like mario is decoding genetics here. It's a GAME and very few games require a bunch of slow 'crunching'. I bid that this is a much a dead end as the wolf topic in the first place. Wishful thinking from a generation of 'net bunnies' who think EVERYONE has broadband and EVERYONE buys the newest console every year. Simply not the fact jack. Mr David Becker of CNET can take his article and put it back where it came from. (his anus) Nice fantasy article though. I particularly like the bit with the interview with Okamoto when he said "Maybe the PlayStation 6 or 7 will be based on biotechnology". LOL. Great sci-fi wonder of a technical genius master jedi that he is, I have no doubt that sony will continue to push the envelope on console design and interface options but this is not a reality in the next couple years. Wait until standardized universal access broadband penetrates a solid 25% of US households and THEN you can talk about having your Nintendo be a slave for other gamers when you aren't on it. Speaking of which, my tamagochi is crying...I have to go feed him. Later!
"Okamoto said Sony is working with IBM to apply Big Blue's research in "grid computing," a variation of distributed computing, to the next PlayStation."
It's a very solid idea... it doesn't suprise me to hear it from anyone anymore. Have read all the naysaying on the Wolf thread, but it seems most have the application of the technology in reverse.
Think about distributed computing... a-la SETI. Very popular... and everyone who's involved is donating their processor to the cause. Now in an office environment, that can get nasty - not in the respect that you can't or shouldn't get the cycles, but in that the bean counters simply don't like to give things away. So if you have a large enough group of computers, you can rent/lease out mass cycles on the open market. This isn't just happening today, this was happening yesterday for Jevus' sake.
Couple of angles, then: 1) Cost of ownsership can begin to be re-evaluated if the machine itself, with little interaction from the user/owner, can broker processor cycles on a net-based market and produce income. You don't need to look much farter into the future than last week to know that you can already sell your excess power (they HAVE to buy it, actually) to power companies if you roll your own at home and can create an excess beyond usage. My kids leave the damn PS2 running day and night - being able to sell cycles would certainly help to offset the cost of ownership and usage. 2) Grassroots movements, relying on the donation of cycles, are suddenly going to have access to more types of equipment/processors/platforms at lower and lower costs. Standards will begin to emerge. Anything that brings the sheer power of tens of thousands of G4s, Athlons, PS3 - whatever - to bear on a computationally deep or intensive task and puts that power in the hands of anyone who needs it is a good thing - it the kind of thing that means I may actualy get a damned flying car sometime before I die, and that's an excellent thing, because I want my damned flying car.
EDIT:
I believe Wired had a sneak peek at a testing unit from Sony a while back - a "concept car" of playstations if you will. It had something like 16 Processors... insane framerates, etc. If I can find it I'll post some fo the info... may be an indicator.
[ 07-02-2002: Message edited by: walrusjb ]</p>
It was said in the wolf thread that computer in urban areas, or places with more Macs, would have more processing power. It would be a self perpetuating market the more Macs that sell the fast they become. This article makes me think it isn't so far off.
In fact, there's been work to make distributed MUDs now for at least five years.
<strong>It certainly doesn't make sense for every game, but think about MMORPGs. Those practically beg for a distributed computing solution.
In fact, there's been work to make distributed MUDs now for at least five years.</strong><hr></blockquote>
<a href="http://www.matrixfans.net/videogame/index.html"; target="_blank">The Matrix Video Game, MMORPGs</a>
I like, I like
<img src="graemlins/smokin.gif" border="0" alt="[Chilling]" />
Thus, the Playstation model and "Wolf" (as described in the relevant thread) are two completely different things. If Apple adopted both, they would impact different product lines, and solve different problems: A distributed computing service would impact their desktop and laptop lines, while clustering would primarily impact their server line (meaning XServe and the products to follow it).
As to the truthfulness of the "Project Wolf" rumor, I can only point out that the XServe is just begging to be clustered, so I would not be in the least bit surprised to hear that Apple is working on this.
[ 07-03-2002: Message edited by: Amorph ]</p>
As for the this is in 1 box... fine and dandy. Make the sucker collaborate with itself. Stick a dual processor in that sucker. Amoph is sure right about the clustering thing vs. the colaborative computing thing. This topic shouldn't really involve wolf.
In summary. PS3 having newtork support? Of course. Having games which swap info? Sure. Having a couple hundred thousand PS3s sharing resources to make a mega processor? Not likely. It sounds too much like.. "By your powers combined, I am Captian xBox!"
So, again: Don't believe the hype. There will probably be an improvement for those games that are amenable to being distributed, such as massively multiplayer games. Unless IBM and Sony are thinking of rolling Internet 2 into the world's living rooms in record time, though, that's probably all it will be, and it probably won't offer a thousandfold improvement even in the best case. It's not a matter of the technology in the console, it's a matter of the narrow, fragile, high-latency bandwidth connecting the consoles together across the internet - and that's outside of Sony's and IBM's ability to change.
We'll see in two years. I won't be in the least bit surprised if their claims become more modest as the project's deadline gets pushed farther and farther back, until the shipping product - impressive, undoubtedly, but not mindblowing - finally appears.
[ 07-03-2002: Message edited by: Amorph ]
[ 07-03-2002: Message edited by: Amorph ]</p>
[QB: A lot of people on here bashed Moto for failing to come anywhere near IBM's 1GHz G3 almost a year ago - but IBM had merely touted it in a press release. They still haven't actually released it, and in the meantime Mot has cleared 1GHz with their more complex G4. But IBM still has the better rep as a PPC provider.
[ 07-03-2002: Message edited by: Amorph ]
[ 07-03-2002: Message edited by: Amorph ][/QB]<hr></blockquote>
You are right , but you must remember also that IBM create the power4, one of the most effecient chip everbuilt, and more the more extensible architecture. So the better reputation of IBM has much more to do with the powerchip, than the early announcement of a 1 ghz chip.