Could PS3 and X360 manage the Crytec 2 engine ?

[pjbliverpool]

Heavenly Sword is more impressive in visuals(polys) and physics. The texture and world memory may need to scale down but that says a lot for your $3000 duo cores.

Last time I checked a dual core CPU didn't cost $3000, nor did a PC containing one have to cost that much. In fact by late 2006, a PC with the equivilent power of a PS3 may be possible for under $1000. If some PS3 rumors are correct then thats only double for a whole PC as opposed to something thats a gaming machine only.

 

[!eVo!-X Ant UK]

If some of the people are to be belived its that Crytec engine 2 is indeed a PURE dx9 engine ( IT'll have a shader model 4.0 patch simular to farcrys SM 3.0 patch ) then there is nothing stoping PS3 from running it the SAME as the PC. And will Cell having its own path i cant see why the PS3 version should'nt improve on the others version physics and A.I

Could Cell even emulate some of the features that DX10 patch might have in Software mode?, simular to the way PS2 has been software rendering effects that it has'nt got hardwired in ?

 

[Shifty Geezer]

They can add physics based eye candy into the game which would scale back if you don't have a PPU or your playing on a console.

Which is kinda my point. In the PC space complex physics can only be used for supplementary eyecandy if you can't be sure it'll be present. You can't create gameplay that's dependant on advanced physics. If the Crytek engine is targetting broad PC users, support for PhysX (even if not in the engine, in the titles written using the engine) will be confined to arbitary components. In the console space the devs know exactly how much resource they have.

Just because a game takes advantage of a PPU doesn't mean it needs one to run. Besides, everyone seems to be assuming that Cell is automatically going to be better than a dual core at Physics (along with all the other CPU work). I think thats a big assumption to make at this stage. John Carmack doesn't seem to hold them in such high regard.

There's been lots of debate on this. Aegia seemed very pleased with Cell saying it was more akin to their PhysX PPU. There's no one opinion.

But this discussion is drifting towards 'top-end PC versus PS3 power' which is way off topic, and done to death.

 

[!eVo!-X Ant UK]

Which is kinda my point. In the PC space complex physics can only be used for supplementary eyecandy if you can't be sure it'll be present. You can't create gameplay that's dependant on advanced physics. If the Crytek engine is targetting broad PC users, support for PhysX (even if not in the engine, in the titles written using the engine) will be confined to arbitary components. In the console space the devs know exactly how much resource they have.

There's been lots of debate on this. Aegia seemed very pleased with Cell saying it was more akin to their PhysX PPU. There's no one opinion.

But this discussion is drifting towards 'top-end PC versus PS3 power' which is way off topic, and done to death.

Indeed

 

[Shifty Geezer]

Actually in retrospect it's not that off topic, seeing as to run a PC engine a comparison between PS3/XB360 and the target PC platform needs be made. But it has been done to death!

 

[Titanio]

They can add physics based eye candy into the game which would scale back if you don't have a PPU or your playing on a console.

That's the difference between "effects physics" and "gameplay physics". On the PC, you can have a lot of the first and scale depending on available power, but you have to make sure that the second runs on all systems you're targetting. With the consoles, there's more scope for more ambitious "gameplay physics" because everyone has that power.

Just because a game takes advantage of a PPU doesn't mean it needs one to run. Besides, everyone seems to be assuming that Cell is automatically going to be better than a dual core at Physics (along with all the other CPU work)

Well, you might consider it an assumption, but AGEIA has classed PS3 has being in the same league as a PhysX chip in terms of physics performance.

I think thats a big assumption to make at this stage. John Carmack doesn't seem to hold them in such high regard.

John Carmack has made very very little comment on Cell performance characteristics - only the programming model and challenges that it presents.

 

[Griffith]

Heavenly Sword is more impressive in visuals(polys) and physics. The texture and world memory may need to scale down but that says a lot for your $3000 duo cores.

for the screenshots of the battle between the woman and some enemy, Heavenly Sword ran at 5 fps in realtime, and It don't impress me so much
the screenshots of hundreds of enemy, I don't know, but I bellieve that this is CG

Dx 10 GPU's and quad cpu, 2 GB machine will put in the dust ps3, I think you can't fight this point, if you are a fair person

 

[Griffith]

Well, you might consider it an assumption, but AGEIA has classed PS3 has being in the same league as a PhysX chip in terms of physics performance.
.

yes
but cell is not a physic processing unit but the central processing unit of the machine, it have to perform vertex+general+audio+compress/decompress+IA+physics processing

understand my point?

 

[Titanio]

yes
but cell is not a physic processing unit but the central processing unit of the machine, it have to perform vertex+general+audio+compress/decompress+IA+physics processing

understand my point?

I agree, in that I assume AGEIA's characterisation of PS3 is the amount of power a developer could have at its disposal for physics if it chose to prioritise physics. Obviously it's up to devs to make the balance between physics and other tasks (although I'd say in PS3 games, physics may enjoy a high priority). My point is that the amount of power available in PS3, for physics, even when you take out the rest the CPU needs to take care of, will be more than what's available in minimum spec PCs for some time to come (and that defines how much power a PC game could use for "gameplay-affecting" physics).

 

[!eVo!-X Ant UK]

for the screenshots of the battle between the woman and some enemy, Heavenly Sword ran at 5 fps in realtime, and It don't impress me so much
the screenshots of hundreds of enemy, I don't know, but I bellieve that this is CG

Dx 10 GPU's and quad cpu, 2 GB machine will put in the dust ps3, I think you can't fight this point, if you are a fair person

Not it did'nt, the indoor sections were 60fps @ 1080p, and out door sections slightly less

And ALL realtime

 

[tema]

for the screenshots of the battle between the woman and some enemy, Heavenly Sword ran at 5 fps in realtime, and It don't impress me so much
the screenshots of hundreds of enemy, I don't know, but I bellieve that this is CG

Dx 10 GPU's and quad cpu, 2 GB machine will put in the dust ps3, I think you can't fight this point, if you are a fair person

You believe it was CG? The finished product will look better than the demo!

 

[Griffith]

Not it did'nt, the indoor sections were 60fps @ 1080p, and out door sections slightly less

And ALL realtime

I read of 5 fps, and remastered in video editing to 60

can you give some evidence of what are you saying?

 

[Griffith]

You believe it was CG? The finished product will look better than the demo!

sure?
you play the finished product, or you work for sony, to say this?
this discussion is going bad...

 

[tema]

yes
but cell is not a physic processing unit but the central processing unit of the machine, it have to perform vertex+general+audio+compress/decompress+IA+physics processing

understand my point?

PPU 128gflops. Cell 200gflops. Cell is here. Where is the PPU?

 

[!eVo!-X Ant UK]

I read of 5 fps, and remastered in video editing to 60

can you give some evidence of what are you saying?

You are thinking of KZ, and evidence?? Why DeanoC, a poster here is one of the DEVELOPERS who are making the game and he's stated the FACTS

 

[Titanio]

You're both wrong. The "indoor" scenes were reasonable framerates, realtime (but not 60fps). The outdoor scenes with the armies weren't CG, they were in-engine, but running at 5fps.

I'd say both are very possibly running a decent clip now, 12 months later. Just a guess.

 

[!eVo!-X Ant UK]

You're both wrong. The "indoor" scenes were reasonable framerates, realtime (but not 60fps). The outdoor scenes with the armies weren't CG, they were in-engine, but running at 5fps.

I'd say both are very possibly running a decent clip now, 12 months later. Just a guess.

My mistake

 

[Griffith]

You're both wrong. The "indoor" scenes were reasonable framerates, realtime (but not 60fps). The outdoor scenes with the armies weren't CG, they were in-engine, but running at 5fps.

I'd say both are very possibly running a decent clip now, 12 months later. Just a guess.

I was refering to outdoor scenes, with "battle between woman and enemy at 5 fps"...

then this is the outdoor gameplay scene (dated 25 october):

this is not impressive for me


this is CG in my opionion (it's very different in quality), this is not gameplay (dated 28 jun)

http://ps3media.ign.com/ps3/image/article/614/614802/heavenly-sword-20060117011353953.jpg

I explain my point now?

I take casual pic, not the worse

 

[pjbliverpool]

John Carmack has made very very little comment on Cell performance characteristics - only the programming model and challenges that it presents.

Carmack specifically states that both console CPU's are way slower than the specs suggest and even with high optimisation, they only compare to regular PC CPU power:

http://forum.teamxbox.com/showpost.php?p=6694749&postcount=1

The difference between theoretical performance and real world performance on the CPU level is growing fast. On say, a regular Xbox, you can get very large fractions of theoretical performance with not a whole lot of effort. The Playstation 2 was always a mess with the multiple processors on there, but the new generations, with Cell or the Xbox 360 make it much, much worse. They can quote these incredibly high numbers of giga-flop or tera-flops or whatever, but in reality, when you do a straighforward development process on them, they’re significantly slower than a modern high end PC. It’s only by doing significant architectural work that you even have a chance of finding speed-ups to what the PC can do, let alone it’s theoretical performance. It’s only through trivial, toy or contrived applications that you can deliver the performance numbers they claim.

 

[scooby_dooby]

That's not at all what he is saying. He said with a STRAIGHTFORWARD development (i.e. from a PC developers standpoint) they are slower than PC.

He says you need signifigant architectural changes to make them realize their true potential (hello captain obvious)

Nowhere in that statement does he say the potential power for CELL or XeCPU is less than a regular PC CPU.