The book speaks in terms of the whole videogame division as a whole,so the "cuts" could have been in things like first party software development, or paying for fewer software exclusives. Just anywhere they could reduce budget. Might have been a little savings from various areas, combined with eating the remainder of the cost.
I'm not aware of any major hardware spec downgrades at any point, so that seems unlikely.
The idea of being scared to go over the opponents RAM total is interesting, and on it's face is basically a surrender to not having technical superiority at first blush. However, I wonder if the dollars are better spent in silicon, getting you better return? AKA match the RAM, try to spend more than the other guy on chips is a more cost effective strategy than trying to outdo in both areas?
Hypothetically I mean, not that microsoft did this.
I remember reading a story somewhere that I think it was the PGR team, threw a party with champagne when they got the 512 news, they were so happy.
Corwin B, I think your EDRAM idea is a better one for next generation if at all. According to Wavey's table 720P 4X requires ~28MB for the framebuffer. Round it off and you'd need 30MB instead of 10. I'm not sure that much EDRAM is even technically fab-able. AFAIK you'd be looking at 250-300 million transistor just for that, which would probably be as expensive as the GPU and CPU itself, and seems like overkill for the amount of shader power 360 has.
I do love the idea of not tiling though, it makes the system conceptually so much simpler, and tiling is a niggling performance drag. I would like to see it next gen. Next gen is where the EDRAM idea becomes very interesting if not a necessity for consoles, else they have to figure out another major way to tackle GPU bandwidth (and just imagine what kind of BW the next gen will need, with 8800 already needing 80+ GB/s to feed itself).
________
Live Sex
Last edited by a moderator:
