Atikali Havendoorr
Mannar Focused Warfare
Gallente Federation
48
   |
Posted - 2013.07.30 00:05:00 -
[1] - Quote
Mobius Wyvern wrote:Gloomy Cobra wrote:Billi Gene wrote:Gloomy Cobra wrote:So dust needs to be its own game before it can become something else, CCP knows how to dream big but empty words don't mean a thing. You need to start from the bottom in order to get to the roof. at what point do you realise there is a process going on? we're on the stairs to the first floor as we speak, though i have to say the basement was rough going at some points... Are we really? They showed jets back in 2009 and large maps in 2012, exactly how are we at the stairs? more like we took steps back. Where is the trophies that has been promised back in 09? They even promised other stuff this year at fanfest you think we will have it by the end of the year as they said? I have high hopes for this game but at the end of the day you cant really put it on the same page as other great games. Its unfair to compare dust to bf3 but sense dice took over StarWars lets see how long it takes for them to throw out a "full release game" that has everything in it. Again, that was all made in an engine unsuited to this platform, and they had to toss it all and start over because you can't just "port stuff" over to an entirely new engine.
So much self-pwnage.
If they just realized what miserable piece of cr*p the PS3 is. |
Atikali Havendoorr
Mannar Focused Warfare
Gallente Federation
49
|
Posted - 2013.07.30 20:49:00 -
[2] - Quote
Nosid Katona wrote:Atikali Havendoorr wrote:Mobius Wyvern wrote:Gloomy Cobra wrote:
Are we really? They showed jets back in 2009 and large maps in 2012, exactly how are we at the stairs? more like we took steps back. Where is the trophies that has been promised back in 09? They even promised other stuff this year at fanfest you think we will have it by the end of the year as they said? I have high hopes for this game but at the end of the day you cant really put it on the same page as other great games. Its unfair to compare dust to bf3 but sense dice took over StarWars lets see how long it takes for them to throw out a "full release game" that has everything in it.
Again, that was all made in an engine unsuited to this platform, and they had to toss it all and start over because you can't just "port stuff" over to an entirely new engine. So much self-pwnage. If they just realized what miserable piece of cr*p the PS3 is tech wise. I believe he is referring to the Unreal Engine; not the ps3 hardware. As they have been saying the ps3 has great hardware for a console but is a ***** to program for; so it isn't the system hardware cause we have seen some really amazing games on ps3. If you don't like the ps3 why own one anyway? Don't like it go get a PC or Xbox and go away. :p
I know. But they scrapped Carbon because it couldn't run very well on the PS3. It is made for PC, a generic architecture.
I bought a PS3 to play games with, not programing. Blu-ray and free online was the winning points for me. And not being M$.
The PS3 is sh*t hardware wise though, especially for a gaming platform. A good gaming platform is not foremost about being powerful, but being easy to use (for the programmer). Spending time on just learning how to make things work on a specific platform, instead of the game itself, is basically waste of time. And also money.
I'm not a programmer, but I have a good view of how things work. I have immersed myself in the matter, and I can tell you that the Cell is a totally new way of thinking. I think STI is way ahead of time, and also suffering from things going another direction than they predicted and hoped for. The whole point of the Cell is "to bridge the gap between the CPU and GPU". When the project started in 2000, CPUs and GPUs were very separate. Now they have come closer together by much easier and cheaper means than the Cell made it. A CPU can do GPU stuff, and vice versa. Not very good, but it is possible.
Originally Sony planned to have two Cells in the PS3. Only. The Cell is meant to do both generic calculations, and vector stuff. But they ran into manufacturing problems because of the difficult design, and in the end had to rethink it all, and scrap one Cell to be able to get enough working chips to actually release the PS3. They called Nvidia to put a GPU in it, and the architecture remake that followed is what totally f*cked up the PS3. Basically the Cell and the Nvidia chip don't "speak" to each other, so they had to come up with many quirky stuff and workarounds to make them play. That's why the memory is divided in 256MB for Cell and 256MB for the RSX. The RSX can't access the Cell memory directly, and the Cell can't access the RSX memory directly, only via detours. One way is good, but another is terrible. Don't remeber which, but one data route is limited to 16MB/s. Usual number here are same in gigabytes. This made programming very difficult. Originally the PS3 was meant to have 512MB unified memory which both Cells could access. And I haven't even mentioned Cell itself...
The Cell is something you can research for decades, and still find new ways to use it. It is truly groundbreaking.
But..! a big but here... We still get decent performance leaps with the usual methods of manufacturing process shrinking, microarcitecture tweaks and adding cores. But we are very soon at the top of what is achievable in relation to cost, and not far away from the outright physical border. And when that happens, I think the Cell will shine. But it will take some more years for that, maybe up to a decade.
They didn't put the Cell in the PS3 because it was the best choice. The put it there to get the Cell out on the market, and thus paving the way for a whole new processor era, rivaling basically everyone from Intel to Texas instruments, since it was meant to be scalable and fit in everything from supercomputers to TVs. That was the plan.
They had a very strong brand and platform to put it in: Playstation 3. Thus they would force propgrammers to use it, spread knowledge, and thus making it more viable in other ways.
This basically made gaming programmers real world test players for the Cell, which had to figure out themselves how to work with it. Everything was brand new, so compilers and other tools were in an infant phase.
And having many cores, especially weak ones, is generally a very bad thing for game programing. Since everything is run in real time, all threads have to be synced for every frame, which makes thread splitting very difficult, and mostly impossible. One calculation is awaiting the result from another, thus they cannot be done simultaneously, in parallel, but must be done serial. That makes one strong, or a few fairly strong cores a way better choice for gaming than many weaker cores. I would say that more than 4 cores for gaming is not very useful, even in the years to come. And it's not because of lazy progammers, like many think, but because of the nature in game programming.
This got way longer and than I thought. But there you go. |