So we are going from having mostly none of the specs to having all of the specs in the matter of days....funny how it works.
I think most people would think it a waste of time to drop individual specification values over an extended period. For developers not in the know, it would just leave them waiting until any bottlenecks or quirks show up before they can plan, and for other observers there's not really much value in a single number robbed of much of its context. It would seem like a product spec table is an example of something that is greater than the sum of its parts.
What other approach would it take, anyway?
PS5 advent calendar: for each day, reveal one number or letter that goes to an architectural value?
Mark Cerny goes on the livestream wearing an outfit of balloons, and every one he pops drops a bit of technical trivia?
"We'll see how much you want to know how much compute power there is," he says as he spirals in towards a critical remainder of his covering.
I hope there will be at least some hardware novelties to present outside of rdna2. Otherwise we're going to have a simple repetition of what we already know.
I hope to hear more about things that make the architecture more approachable to developers or provide infrastructure for the platform, much like the APIs and in-built hardware Microsoft disclosed for storage, data residency, latency, and variable refresh rate.
There are signs that there was investment into backwards compatibility and integration between the hardware and OS for that purpose, and there could be other similar mechanisms since there's supposedly a custom storage subsystem that can give developers more control, and various disclosures or patents from AMD or Sony that might provide some services like the Xbox reveal touched on.
I think those other engineering efforts outside of headline specs could be very important, since Sony went with the more approachable platform in the current gen and the platform now needs to handle a complex transition for hardware and storage. I would say the risk is higher, depending on how in-house some of these efforts are. Sony was not entirely successful with its firmware and platform engineering with the PS4, and I think things could have been somewhat different if its competition's efforts weren't divided into so many speculative non-core directions or implementing emergency changes to the whole architecture ahead of launch.
I'd like to see how broad and deep the engineering effort goes for both of the upcoming platforms.
All of a sudden, teraflops be damned. While i do agree tf arent all that important though
9.2 PF confirmed...