Mark Cerny from Sony shared why the move to abandon proprietary processors for an x86 chip with unified architecture is the right move for the PS4.
When the Playstation 3 came out much was hyped about the sheer potential of PowerPC-based CELL processor designed by Sony, Toshiba and IBM. With eight-cores, the chip promised near photorealistic graphics in glorious high definition in an era where dual-core processors had just reached the PC market (2006).
The only problem with the CELL processor was most developers couldn’t figure out how to write multi-threaded code that would push it to its fullest potential. Studios that pushed out yearly ports (like EA) have cross-platform development tools which put the economy of developing for multiple platforms over deep architectural optimizations.
For the Playstation 4 Mark Cerny, the lead system architect, wanted to make sure that developers coding for the PS4 aren’t stuck in the CELL-quandary they faced in the last generation.
"The biggest thing is we didn't want the hardware to be a puzzle that programmers would be needing to solve in order to make quality titles," he said in an interview with Gamasutra. “[With the CELL] There was huge performance there, but in order to unlock that performance, you really needed to study it and learn unique ways of using the hardware."
Considering that the next-generation Xbox will also very likely use an x86 processor, Cerny says a philosophy of familiarity was a guiding force when picking the x86 architecture to develop the PS4.
"We want to make sure that the hardware is easy to use. And so having the familiar CPU and the familiar GPU definitely makes it easier to use," he said.
Cerny calls the PS4’s system architecture "supercharged PC architecture," pointing to the 8GB of unified GDDR5 memory – which is something he said developers wanted – available to the CPU and GPU. He also points to this unified architecture, eschewing PCIe, as a way that the PS4 might trump gaming PC in hertz-for-hertz performance.
"[If a PC] had 8 gigabytes of memory on it, the CPU or GPU could only share about 1 percent of that memory on any given frame. That's simply a limit imposed by the speed of the PCIe,” he said. “So, yes, there is substantial benefit to having a unified architecture on PS4, and it’s a very straightforward benefit that you get even on your first day of coding with the system.”
Both the CPU and GPU are on a “very large” single custom chip created by AMD for Sony that has many characteristics similar to an AMD APU. It has eight Jaguar cores, all on the same package as the GPU. While the memory is not on the chip, it is attached via a 256-bit bus giving it access to a shared pool of RAM at an impressive 176 GB/s.
With this kind of architecture, bottlenecks will be a thing of the past.
"With graphics, the first bottleneck you’re likely to run into is memory bandwidth. Given that 10 or more textures per object will be standard in this generation, it’s very easy to run into that bottleneck," he said. "Quite a few phases of rendering become memory bound, and beyond shifting to lower bit-per-texel textures, there’s not a whole lot you can do. Our strategy has been simply to make sure that we were using GDDR5 for the system memory and therefore have a lot of bandwidth."
Cerney didn’t share any details about the PS4’s launch lineup, but said it was going to be “stronger” than any previous Playstation launch.
The PS4 is expected this holiday season.
Check out the full interview at Gamasutra