This will be a bitch to find, but common sense will tell you going trough two memory controllers adds a lot of latency.
But funny you ask for me to source things while MikeBs claims go unsourced.
The only thing he needed to source but didn't in these past few posts, AFAIK, is the "developer quotes", and
for the record, MikeB, I would really like to hear your source for those before I believe them. I have asked MikeB for clarification of his arguments and points in the past, so don't pretend I'm his apologist. I've contradicted him in other threads, and will in future threads.
This will be a bitch to find, but common sense will tell you going trough two memory controllers adds a lot of latency.
If you are referring to the simple fact that the XDR memory is separated from the RSX by the Cell, then please provide your explanation of why this compares badly to the fact that the 360's CPU faces the same ordeal, only worse, and with all 512 MB of its RAM.
If you are referring to something else, perhaps the Cell's 'Ring Connector', and since you say that it is slower and has more latency than the Xenos' 'Cross type connector', you must have some idea -- which I invite you to convey to me -- as to how much slower the one is than the other. Unless, of course, you know only that it is slower but it could be by 1% for all you know. You have NEVER YET here explained why the RSX will have a disadvantage in accessing the XDR other than hand-waving about "latency" with no numbers, estimates, or even guesses as to the amounts involved.
Important!And you have also done absolutely nothing to answer the allegation that all this is a sideshow to the true strength of the PS3's memory architecture, namely that the Cell and RSX each have 256 MB of memory that they can access without having to go through the other -- which is certainly not true of the 360.
I am not an expert, but I think I am knowledgeable enough to understand any explanation you give -- and if you are right, maybe you'll convince MikeB.