| CGI-Quality said: PCs are, in fact [and sometimes], bought for their power advantage over consoles. Other than that, I hear a lot of what you're saying (although you can benchmark the same machine with the same program and get different results, it just depends on the processes at the given time - hence the dynamics), but some of it is thinking too hard at best. The specs aren't "random", either. It's just a matter of one of them being ahead. The % amount may be pointless to discuss, but the advantage is not. |
Ah, I see what you mean with 'dynamic results'. You are right on this. But professional benchmarks are done with fixed environment, to avoid such confusion. Hence I mentioned SPEC, it is done very professional.
PC for gaming are sometimes bought because of advantages, but not the raw power, but more 'I can play this game I love in 60 fps in multiplayer even with much stuff going on.' Or higherresolution, better textures or such things. The power allows for such things, but it is not the power alone, also the games utilizing it.
With 'random' specs I only meant, we don't have a full spec-sheet, we only know a few highlights of the consoles. No, a teardown doesn't uncover all details.
You're right, that we can decide that one detail of the spec of a machine is ahead of another. But the consoles have a lot of technical details. Are you sure one thing isn't offset by others? For instance, one less discussed detail is the switching to PC-tech, meaning also the use of the x86/x64-instruction-set. This instruction-set has problems associated that make it less efficient than other instrution-sets of different architectures. Nobody discusses that decision. But the effects on branch-prediction, instruction-prefetch and instruction pipeline are there. Have the manufacturers done something to reduce the negative effects of this instruction-set, because they don't need the full PC-compatibility? I don't know, but such decision may affect the performance on a noticeable scale.
I see you agree on the discussion of %.
