Slashdot Mirror
Sony Says Nobody Will Ever Use All the Power of a PS3
Tighthead Prop writes "Sony executive Phil Harrison has made some brash comments about the Cell processor and the PlayStation 3. Harrison says that the current PS3 game lineup is using less than half of the machines power, adding that 'nobody will ever use 100 percent of its capacity.' Is he right? 'The major reason Harrison wants to hype up the "unlimited" potential of the PS3's architecture is to downplay comparisons between games running on Sony's console and Microsoft's Xbox 360. The two systems are not completely dissimilar: they both contain a PowerPC core running at 3.2 GHz, both have similarly-clocked GPUs, and both come with 512 MB of RAM.'"
Something about 640k of RAM...
It's natural law, all the available power of any machine will ultimately be used. The only real reason the PS3 full power may not be used will probably be linked to the 9 cores architecture and the inherent diffulty to use them all at once effciently.
Conclusion: they are trying to present a bad news as a good one, business as usual...
I'm not sure this is something I would want to brag about. If you made the system so complex that it was impossible to use to its fullest potential, then why did you make it so complex and/or powerful? Sounds like admitting to a lot of wasted effort.
Well, if it's not possible to use all the power in the PS3, there's no point in making a more powerful console in a few years time, right?
Should have made it a bit less powerful and consequently cheaper then I suppose. They'd have sold more and make more money that way.
http://twitter.com/onion2k
As long as they make available the Next-next generation PS4 available soon enough...
That is what had happened after the SNES-Genesis-etc days (From the N64 onwards), the "next gen" iteration life span has became shorter and shorter so developers just start to get familiar with the system when the Next-gen system gets out.
I will sound like the old-grandpa but I liked more when the game generations lasted longer, you could see really nice things done with the technology and the hardware had more "value" (see for example all the NES peripherals) as you *knew* the system will remain active for a long time and more games would likely come.
I won't buy the "eye toy" or the "maracas" or the "bongos" today for any system because I know that only 1 or two games would ever be available.
Ubuntu is an African word meaning 'I can't configure Debian'
Nobody uses 100% of the power of their desktop computer either - and nobody complains about it. It would take a very, very tricky program to simultaneously max out the processors, graphics, memory, and disk bandwidth.
Nobody every uses 100% of the power of their car, either. Sure, you LIKE to have the 250 HP engine, but you only use it for 3 seconds on the on-ramp. And hopefully nobody uses the full power of their 800 watt home theatre system. The excess power is there for the momentary condition - not to use all of the time.
Cell just isn't that suited to gaming.
With the GPU doing graphics, one core doing AI/Gameplay, another doing Physics, another doing Audio/Networking/Input you've pretty much got all the processing power you need. If you start spreading a game out across too many cores it's going to negatively effect the speed of the game due to the fact you're going to spend all your time trying to keep threads in sync. I'd argue that this is why Sony has it wrong and MS has it right. The GPU can handle graphics, then the 3 cores can be used as mentioned above - this seems the optimal division of work in a game engine. I'm convinced that 4 physical processing units at 4ghz would be better than 8 physical processing units at 3.2ghz so perhaps that would've been a better route for Sony if they really felt the need to beat the 360 on performance.
To me the Cell seems more suited to number crunching type applications, the sort where you can offload large amounts of data to each cell and let them go on their merry way processing these chunks without having to worry about whether every few bytes of data is in sync.
I honestly wonder if Sony management just assumed that the Playstation 3 would cell like the PS2 and PS1 and hence just insisted they use it as the tool to bring down the prices of Cell and BluRay regardless of whether they were fit for purpose or not.
I seriously hope you work for Sony if you're spending this much effort promoting the console online.
Unlike the fellow above, I actually hope Sony is not paying you, because I would hope they'd get more for their money than a list of "Untitled" exclusives...
Or perhaps the PS3 utilizes a stupid design considering that developers have repeatedly shunned the most complex console.
Even in the 16 bit days a lot of developers went with the genesis rather than the SNES even though the SNES had greater capabilities because the genesis had more CPU and it was simply easier to develop games for it.
In the 32 bit era the Saturn was basically ignored by everyone because it was a nightmare to code for, meanwhile the PS1 came out. It had roughly half the raw CPU power of the Saturn (The Saturn has two Hitachi SH-2 processors, while the PS1 has one MIPS R3000 at about the same clock rate - SH-2 and R3000 are both pretty pathetic 32-bit RISC designs, although many people will hate on me for saying that about the R3000.) The PS1 however made it EASIER to do a lot of things - you only had to focus on one CPU and they did transparency in hardware. You can't use their graphics chip for general purpose computation though. The Saturn is more powerful and you CAN do transparency but you have to do it in software using the second CPU (in order to do the rendering in a timely fashion) and that is hard. So again the more powerful platform is neglected.
Those who would use Dreamcast as a counterexample should note that it died more because of Sony marketing than because of anything else. Developers abandoned it and waited for PS2, which turned out to have specs about an order of magnitude less powerful than announced.
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
How can a game be "in development" if the developer is still TBD?
The correct answer is:
C) The Cell is a poor general purpose processor.
If you're at all familiar with the fundamentals of CPU design, it should be blindingly obvious that the Cell should be very good at handling streaming vector data, but relatively poor at more general purpose calculations.
We still haven't even used the full capability of a 300mhz processor and 32mb video card. The bottleneck is not hardware, it's software. Inefficient code, outdated methodologies, and improper application of libraries is a much greater bottleneck than the hardware in any system.
More cycles and more memory doesn't mean that developers are capable of using better graphics and logic, it means they can be lazier in their optimization. Games which take up 5gb of hard drive space do so because they can, not because they must. Developers know the user has 100+ gigabytes available on their hard drive, so no further optimization is necessary. They know that the video card has 256mb or more memory, so they don't optimize the game anymore than they need to. We only need 3ghz processors because developers can throw away as many cycles as they want. On a needs basis, the actual logic and graphics of the most powerful game available probably would require a 300mhz processor and 32mb of video memory. All the rest is a buffer for waste.
This isn't a sleight against coders, I'm a professional developer too. I've seen a lot of applications that could be optimized further but other tasks are much higher up the priority tree because even though the program could be more efficient, it doesn't need to be.