Re: New 360 Pic
Kuu, you're assuming the multi-core 360 and PS3 cpu is perfect, which I highly doubt.
I do not mean dual core in the sense of dual CPU. I mean two main processing units on a single CPU.
Branch prediction and cache problems are not 100% solved to this day. It is already known that AMD has a better branch prediction in the A64 versus the P4, and that IBM’s research efforts for branch prediction in other processors is far behind those two because they posses less experience and spend less money on development. (It also shows up in power-pc tests against the AMD and Intel counterparts.)
Cache is only there to increase bandwidth. The method for determining the amount of cache needed varies from algorithm, and it is not perfect. Too much cache, and you loose transistors for processing, not enough, and the processor can’t be fed continuous data to work as well as it can. Especially for a first time processor like the cell, I don’t think they have accounted all possible processing scenarios at full efficiency--there's simply no previous experience for a cpu of this type. It’s not an evolutionary refinement, but a new type of processor, and like all new things, there’s going to be problems.
You also completely misunderstand me in terms of high resolution. I never said that high rez in and of itself is bad. (that’s f***-ing obvious) I'm saying that at high resolution, when you do poor work, it looks much worse that it would be at lower rez. Why? Because if a lazy designer uses a low rez texture in a high rez game, it’s very apparent. You can’t do the same level of work that would be appropriate for the PS1 and expect it to fly on the PS3. The higher the rez goes, the more work is needed to fill it in--it is never less work to do, or just the same as you claim.
I work in the goddamn print industry, I know what the hell I'm talking about.
I’m not anti-future, I’m not anti-tech. (Why the hell would I be?) You want to think I am for the sake of your argument, but that’s not what I’m saying. I’m saying that the route that the engineers set about creating is high dubious in it’s efficacy with the given resources. IMO, the radical approach that they went with, is not route I would have taken because it creates as just as many problems than it solves.
BTW, I don’t what Kailea-D’s comments are about… It’s just stating the obvious.
Kuu, you're assuming the multi-core 360 and PS3 cpu is perfect, which I highly doubt.
I do not mean dual core in the sense of dual CPU. I mean two main processing units on a single CPU.
Branch prediction and cache problems are not 100% solved to this day. It is already known that AMD has a better branch prediction in the A64 versus the P4, and that IBM’s research efforts for branch prediction in other processors is far behind those two because they posses less experience and spend less money on development. (It also shows up in power-pc tests against the AMD and Intel counterparts.)
Cache is only there to increase bandwidth. The method for determining the amount of cache needed varies from algorithm, and it is not perfect. Too much cache, and you loose transistors for processing, not enough, and the processor can’t be fed continuous data to work as well as it can. Especially for a first time processor like the cell, I don’t think they have accounted all possible processing scenarios at full efficiency--there's simply no previous experience for a cpu of this type. It’s not an evolutionary refinement, but a new type of processor, and like all new things, there’s going to be problems.
You also completely misunderstand me in terms of high resolution. I never said that high rez in and of itself is bad. (that’s f***-ing obvious) I'm saying that at high resolution, when you do poor work, it looks much worse that it would be at lower rez. Why? Because if a lazy designer uses a low rez texture in a high rez game, it’s very apparent. You can’t do the same level of work that would be appropriate for the PS1 and expect it to fly on the PS3. The higher the rez goes, the more work is needed to fill it in--it is never less work to do, or just the same as you claim.
I work in the goddamn print industry, I know what the hell I'm talking about.
I’m not anti-future, I’m not anti-tech. (Why the hell would I be?) You want to think I am for the sake of your argument, but that’s not what I’m saying. I’m saying that the route that the engineers set about creating is high dubious in it’s efficacy with the given resources. IMO, the radical approach that they went with, is not route I would have taken because it creates as just as many problems than it solves.
BTW, I don’t what Kailea-D’s comments are about… It’s just stating the obvious.
Comment