PSX5Central
Playstation/Gaming Discussions => PS3 Discussion => Topic started by: Nplayer-2 on March 15, 2001, 02:29:58 PM
-
I am sure most if not all of you already know about the deal Sony IBM and toshiba have made...they will create the most powerfull CPU ever..code name "cell" it will reach REAL speeds of 1 teraflop...have atleast 32MB of on-board ram..and all of this on one chip...meaning VERY cheap....but it will be HUGE;)
With this chip running at 3Ghz can have 32 VU\'s...i already hear "WTF"....yes this is like the PS2 nightmare all over again...except sony will have vast knowledge of this technology(thanks to PS2) and there for developers will have all the source code they could want...
The chip will have HUGE HUGE pipes BILLIONS of polygons will be able to get processed.....
this chip is the HOLY GRAIL...UNLIMTED effects will be possible(yes real-time ray-tracing...real-time renderman..real-time everything).....:D
I predict in the next 10 years sony will be KING of graphics...nvidia will be no more because they just don\'t have the vast resources sony and IBM have...
point is...PS3 is going to be the be all end all console:D
-
Than Microsoft will duplicate it using PC parts which will make it look superior and have stuff ported to it
-
I doubt it.......the chip i\'m talking about is the low-end;)
what people don\'t understand is in the future GPU\'s will NOT be needed any more...CPU\'s will be so damn powerfull they will be able to handle everything(IBM and SONY seem to be the only ones who know this).....
-
You have a pic under your name for a reason...I wonder if anyone will take your post seriously.
Good job. It\'s your first post this year WITHOUT insulting the Xbox. Kudos.
-
this pic has nothing to do with consoles.
-
can IBM actually deliver a teraflop part? My answer is, yes. Based
on 0.10 micron technology, EE3 will be able to clock at 3 Ghz. This means EE3 has to
deliver 333 flop per cycle to reach teraflop. Presuming that the fundamental data size of
128 bit vector does not change(there is no reason), each VU with 4 FMACs and 2
dividers will deliver 10 flop per cycle and you will need 32 VUs + CPU FPU to reach
teraflop. This is technically feasible since you have 100~200 million transistors to play
with at 0.10 micron level.
But how would programmers be able to manage 32 VUs, when they were unable to
cope with only 2 VUs of EE1? Recall that EE1 programming headache arose because of
direct VU visibility and the problem goes away if the multiple units are properly
shadowed, much like how programmers only see one pixel shader even though four are
actually present in GF3. Likewise, IBM could arrange 32 VUs in bank and keep only
one input to make 32 VUs appear as one. Under this programming mode, a programmer
would simply dump his data packet into VU bank input port; the VU manager read the
packet from input port and send it to an idle VU along with tagged script code. The
results are then sent to destination indicated by the script, be it CPU, memory, or
rasterizer.
We will have to wait until EE3 unveiling to find out about exact details, but I have my confidence in IBM and Sony.
Speaking of EE3, it is quite evident than GS3 will not have T&L since the so much
computational power is focused on EE3. From Sony\'s ISSCC2001 presentation, the GS3
will have 32~64 MB of eDRAM and clock at 714 Mhz. Sony\'s continued focus on
eDRAM makes GS3 T&L unlikely due to low clockspeed and die-size limitation. ***3 will have a very good CPU but the GPU could stay underwhelming.----DM
i think this is to complicated for most of you so this will be my last post in this thread.
-
:p
-
interesting, but its so far off, I\'ll wait till the articles and the chips are shown off to dissect it
Its going to be a heluva beast thats for sure.
-
I care about games, not planned chips by Intel....
-
ddaryl:
i agree with you....damn sony had to let us know about this:D..
i expect PS3 to handle episode1 CG with no problem in-game....if they come through with the teraflop claim;)...which IS possible.
-
so this will be my last post in this thread.
You lied, so you really are a troll...
-
It looks like some pretty good news. But I think it would be best to wait until it\'s further in developments. Who knows what kind of obstacles they\'ll have to overcome to make this chip. :cool:
-
This product in 10 years right? Look what the GeForce 3 is doing right now. God knows what nVidia has in stock for its 4th edition. No way is Sony going to be ruler of graphics when your speaking about nVidia. Nobody thought they would beat 3dfx, now look today. 3dfx bought by nVidia, nVidia controlling the graphics market for the most part. Plus lets not forget ATI, another strong companion.
Will this PS3 be able to make my bed? Make coffee? Brush my teeth? If so, count me in on buying one.
jm
-
Having huge pipelines isn\'t always an advantige.Alot of processing power can be wasted if the pipeline is to deep.The P4 has a 20 stage pipeline which is one of the reasons it get\'s to such high clock speeds but it does have disadvantiges.A G4 processor has a 7 stage pipeline and it can process a sertein amount of information on a few clock cycles,the P4 with it\'s deep pipeline will take more cycles to process the same amount of information.
-
What is geforce3 doing?..simple, BARELY beats the Ultra, and the KYRO2 for god sakes beat the ultra in some benchmarks.....
their method for 3D chips has reached it\'s limit..tilers are the future not brute force....
it\'s good you brought up 3dfx because THAT CAN HAPPEN TO NVIDIA TOO...oh it can\'t?(dejavue;))
What sony is trying to do is FAR beyond what nvidia can even think of today.....
just look at what the "cell" might do...Sony will be king of graphics in 5 years...mark my words.
NO-one:
i think the reson why it seems to be useless is because well no one has made any apps which need the P4...it is crap...but the "cell" is going to be built from the "ground up" 100%....it\'s going to be BRAND NEW TECHNOLOGY...scary eh?;)
we\'ll see where nvidia and sony are in 5 years i guess.
-
Originally posted by jm
This product in 10 years right? Look what the GeForce 3 is doing right now. jm
jm
1st this product will be out in 5 years not ten
2nd this is just the processor , this is not the the graphics Synthesiser
It will be interesting to see where Nvidia and others do graphically in the future, but your missing the whole point of this development, but for raw multifunctional processign power Sony\'s/Toshiba\'s EE is already king, and add this to IBM power knowledge and the fact they are designing this chip specifically for multimedia broadband applications its going to be very hard to compete with
I\'m also sure Sony will have huge advancements in their Graphic Sythensiser. We will see a few effects on chip as well as large amounts og memory and bandwidth.
My moneys on the Sony / Toshiba / and IBM team over anyone else
-
nO-One: PowerPC and Intelx86 processors are quite different, one is RISC the other CISC as such this would be reflected in their architecture- if that takes effect on the number of pipelines that are in the processor, I dont know for sure, but it makes sense to me. (FYI- \'R\'educted/\'C\'omplex Instruction Set Chip)
N-Player: I think its optimistic to think that Sony/Toshiba/IBM will muscle in on nVidias domain anytime soon, it took nVidia years to replace 3dfx as the \'market leaders\' at the moment there isnt even a sign of a EE based graphics board on the market, shame.
The fact that the geForce 3 isnt doing much more than current days cards at this moment isnt down to the design, but rather the software - they need to code support for the features on the card, its like playing pacman on your PC with a 3d card and complaining when things arent running faster. Until the new features are supported in games you wont see the potential of the card.
Go watch the DooM3 video a few times, thats a pretty good example of the evolutionary step from gf2 to gf3.
-bd
-
Go watch the DooM3 video a few times, thats a pretty good example of the evolutionary step from gf2 to gf3
Yes a nice demo that iD provided. When the graphics look like that in the actual game.. then I\'ll be amazed. While iD said it was real-time, and I don\'t doubt them, it doesn\'t mean it\'s actual gameplay. No AI/physics/etc.. No way in hell will the PC be able to handle a bunch of those high-polygon models on the screen @ once.
The GeForce 2 can theoretically do 20mpps, no game even does a third of that. While the GeForce 3 is powerful, the CPUs in todays PC\'s don\'t match.
- dm
- what\'s in the f\'cking box?
-
Think about it for a second......we are now only 5 years away from REAL-LIFE graphics(ie...no more BS..this is going to be hollywood movie graphics)...and sony looks to be the first to reach that..quite a milestone in my opinion:D
last gen the BIT wars were finally over..this gen the polygon wars will finally be over..next gen the POWER war is going to be over..for sony that is;)
Damn sony had to get me all excited about this:D
at the moment there isnt even a sign of a EE based graphics board on the market, shame.
i think they want to enter with a monster chip that will destroy everything in it\'s path...in comes "cell";)
-
datamage: From one of the plan files either during the last few months of Q3 development or around the time Paul Steed left/a Doom game was announced in the works Carmack said that they could use dynamic LOD on the models, that is the number of polys in use on the monsters players items would be determined by how far they are, what the detail level is set at and perhaps even how much strain the PC is under. Of course that can only save so many fps, your probably close to the truth although in 18 months time CPUs will have progressed somewhat. Heh, Im almost dreading the release of a game under the DooM banner for the simple reason that its unlikely it will live upto its name - Doom was all about horde upon horde of demonic beasties just asking to be mowed down, I think what might be coming may be closer to Quake 1/2 where by instead of having hordes of beasties you have sparsely populated maps with monsters which take a real pounding before they go down.
Nplayer: You can never have enough polys :P
-bd
-
well, i think we will reach a point where polygons just won\'t make much of a difference....maybe;)
i wonder why dr yassam is not here?;)
-
True we will reach a point where adding any more polys wont make anywhere near as big a difference as say throwing even a few thousand extra per frame would nowadays, but I feel that will be some time yet -anyway Im sure developers will find ways to eat up all the power new technology will offer them. As for Yassam, hes from the UK, yes? If so my clock tells a story... Its 6.26am at the moment :)
-bd
-
Originally posted by nO-One
Having huge pipelines isn\'t always an advantige.Alot of processing power can be wasted if the pipeline is to deep.The P4 has a 20 stage pipeline which is one of the reasons it get\'s to such high clock speeds but it does have disadvantiges.A G4 processor has a 7 stage pipeline and it can process a sertein amount of information on a few clock cycles,the P4 with it\'s deep pipeline will take more cycles to process the same amount of information.
I seem to be having a brain fart right now. Is a pipeline the same as a bus? If so I think you are wrong. A large bus is never a problem.