CAA: Christian Anime Alliance

I figure that some of the gamers here might be interested in this article. Then again, anyone interested in computing might also find in pretty neat too.

http://spectrum.ieee.org/dec06/4745

Well, like I thought, no one has tapped the cell yet... The sony machines have allways had hardware that took a long time to get used to... This gen will be even moreso. They may never get the chance with all the bad PR, though... despite fewer hardware problems than the first 2 machines and all problems being software in nature that can be updated... Unlike the PS2...

I miss it every time I don't have firefox with the built in spellcheck... It makes life so much easier.

I also have to admit I'm looking forward to finding out how well the processor can also be used in scientific applications
http://www.lbl.gov/Science-Articles/Archive/sabl/2006/Jul/CellProcessorPotential.pdf

http://arrakis.ncsa.uiuc.edu/ps2/

I remember the PS2 server... Made by stringing a bunch of them together. If the gov't ends up buying 100 PS3s or so, that would be a big chunk of change taken from Sony since they build them at loss and the government isn't likely to be picking up Blu-ray movies or PS3 games... Of course, all they really want is the cell, and they can probably get that on a more standard server setup rather than actually using PS3s like they did with the PS2s before.

hmmm that seems a little stupid, wouldnt it just be better to purchace the chipsets instead of buying ps3's becuase theres no grantee that the ps3 hardware even makes the best use of the chip and i doubt theres much room for upgradeability off of the ps3 motherboard. maybe scientist arent that smart after all. or perhaps they want to atack thier enemies week spot for great damage!

eventually Cell type processors will start showing up in other types of machines like servers and dedicated processing units (for distributed computing and rendering). With further hardware revision the maximum memory amount (currently 4GB) will probably increase, giving it an increase in performance after the decrease created by modifying it's purpose to doing personal computing.

Well, if you are building an application for multiple systems - IE, if you are looking to do distributed processing, you've probably already got a good candidate for a cell-enabled application. I would think it would be more about the parser and compiler than the system. There are benefits to this technology, too. Most 2nd hand PS3s will be dirt cheap - and most will just have BR defects. I can DEFINATELY see uses for this technology.

Shinja wrote:hmmm that seems a little stupid, wouldnt it just be better to purchace the chipsets instead of buying ps3's becuase theres no grantee that the ps3 hardware even makes the best use of the chip and i doubt theres much room for upgradeability off of the ps3 motherboard. maybe scientist arent that smart after all. or perhaps they want to atack thier enemies week spot for great damage!

To be fair, the people who are looking into this aren't planning on buying up loads of PS3's. This did happen with the original guys who put together the PS2 cluster, but they needed the capacity to be built quickly and cheaply; they didn't have the time or resources needed to make dedicated machines. However, now that people are more aware of the computational possibilities of gaming platforms, there are plans in the works for more specialized units to be sold to the scientific market.

Mithrandir wrote:Well, if you are building an application for multiple systems - IE, if you are looking to do distributed processing, you've probably already got a good candidate for a cell-enabled application.

So, would a string of PS3s be a good candidate to build render farms out of? Of course, I know very little about that high end of computing, but I guess the program would have to be built specifically for the multiple cells and would have to be made to run on Linux or something... Only Maya runs on Linux, I believe... Out of the major modeling and animation programs...

I imagine that the software to let a cell platform group do such tasks is already in progress, Linux distros have had Cell support for about 6 months actually its just no systems were available outside the development groups within Sony, IBM and Nvidia.

That is my thinking as well.