PC game industry gearing up for multi-core generation

Ever since the PC hardware industry grew aware that the silicon chip was soon to hit the hardest dead-end in computer history, they shifted gears and went from single-core, chip-melting speedy processors to multi-core, lower-speed versions of today’s processors.

Soon, the software industry even followed suit, offering everything from multi-core, multi-processor capable operating systems to 64-bit ready kernels.

However, even as time dragged on, the game industry tucked tails and didn’t catch up. With true-blue hardcore gamers shifting from 3 GHz beasts to 2.4 GHz multi-cores, it finally dawned on the many game devs that the time to adopt their titles to multi-core processing was now. Better late than never.

Chris Taylor‘s Supreme Commander was one of the very (few) first to have been designed from the ground up with multi-core processing in mind. Released last February, it took advantage of the extra core to simulate more realistic AI, while providing enough processing power to handle epic numbers of units in a significantly large map.

But then many other PC games, including Blizzard Entertainment’s World of Warcraft, and id Software‘s Quake 4 and Doom 3 were patched to take advantage of an extra processing core. It wasn’t as optimal a solution as Supreme Commander, but according to Ted Pollak of Jon Peddie Research, “It won’t give the same kind of performance, but it’s going to help, and it’s better than nothing.”

But since the developers spend too much time and resources designing a game from the ground up (and that’s with middleware, pre-acquired graphics engine code and an army of devs, folks), the one reason why a game developer turned its back on Intel‘s drive for multi-core games is the more time needed to adopt a serial coded game to parallel processing platforms.

Thanks to RapidMind, the guys behind the development platform for IBM‘s CELL BE processor, a solution for the multi-core development is ready for use. Ray DePaul, CEO of RapidMind, said, “It’s a way to continue to use serial programming but achieve a parallel approach to data parallelism.”

Their solution makes use of a similar development platform used for the CELL 8-core processor: an x86 version of an application programming interface (API) that will accept serial programming code and decide how to distribute the processing load to the cores available to the software. PeakStream is one other multi-core solutions provider that has an API very similar to RapidMind’s.

So all those early adopters and hardware enthusiasts out there take note: soon you’ll be able to finally tout real scores from benchies on true-blue, multi-core designed games. But what would that mean for a good percentage of single-core users still out there?

Ever since the PC hardware industry grew aware that the silicon chip was soon to hit the hardest dead-end in computer history, they shifted gears and went from single-core, chip-melting speedy processors to multi-core, lower-speed versions of today’s processors.

Soon, the software industry even followed suit, offering everything from multi-core, multi-processor capable operating systems to 64-bit ready kernels.

However, even as time dragged on, the game industry tucked tails and didn’t catch up. With true-blue hardcore gamers shifting from 3 GHz beasts to 2.4 GHz multi-cores, it finally dawned on the many game devs that the time to adopt their titles to multi-core processing was now. Better late than never.

Chris Taylor‘s Supreme Commander was one of the very (few) first to have been designed from the ground up with multi-core processing in mind. Released last February, it took advantage of the extra core to simulate more realistic AI, while providing enough processing power to handle epic numbers of units in a significantly large map.

But then many other PC games, including Blizzard Entertainment’s World of Warcraft, and id Software‘s Quake 4 and Doom 3 were patched to take advantage of an extra processing core. It wasn’t as optimal a solution as Supreme Commander, but according to Ted Pollak of Jon Peddie Research, “It won’t give the same kind of performance, but it’s going to help, and it’s better than nothing.”

But since the developers spend too much time and resources designing a game from the ground up (and that’s with middleware, pre-acquired graphics engine code and an army of devs, folks), the one reason why a game developer turned its back on Intel‘s drive for multi-core games is the more time needed to adopt a serial coded game to parallel processing platforms.

Thanks to RapidMind, the guys behind the development platform for IBM‘s CELL BE processor, a solution for the multi-core development is ready for use. Ray DePaul, CEO of RapidMind, said, “It’s a way to continue to use serial programming but achieve a parallel approach to data parallelism.”

Their solution makes use of a similar development platform used for the CELL 8-core processor: an x86 version of an application programming interface (API) that will accept serial programming code and decide how to distribute the processing load to the cores available to the software. PeakStream is one other multi-core solutions provider that has an API very similar to RapidMind’s.

So all those early adopters and hardware enthusiasts out there take note: soon you’ll be able to finally tout real scores from benchies on true-blue, multi-core designed games. But what would that mean for a good percentage of single-core users still out there?