Ageia se retracta...

Ageia Withdraws Xbox 360 Comparisons
By Mark Hachman
September 8, 2005
Executives at physics-chip Ageia backed off recent claims that the upcoming Xbox 360 console would be unable to perform advanced physics processing.

At the European Game Developers' Conference in London last week, Ageia executives claimed that Microsoft's game console lacked the horsepower to run a fluidic physics simulation, which would model water flow or other liquids.

advertisement
Wednesday evening, Ageia vice president of marketing Andy Keane acknowledged that including those comparisons in its presentation was a mistake, and that they were based on speculation from public specifications released on Microsoft. Although the company has tested its software on Xbox 360 development kits, Keane said, the tests were of extremely limited scope, and weren't designed to test the limitations of the hardware.

"The summary of the information below is that AGEIA would like to go on record that we do not have data to support performance comparisons for the PS3 or Xbox360 that would impact any of our physics features," Keane said in an email, which was also sent to other news organizations. "Specifically, statements that the Xbox360 cannot run fluid simulations are not correct. In addition, conclusions about relative performance should not have been stated or implied in our presentations."

In an interview, Keane acknowledged that the original claims regarding the Xbox 360, as reported by ExtremeTech, were a part of the presentation. However, the Ageia employees that presented at the European GDC did not develop the presentation, and did not provide the information in its proper context, Keane said. The slide has since been removed, he added.

"We really have no clue," Keane said in an interview, regarding the capabilities of the Xbox 360 and its ability to process the Ageia physics SDK, known as Novodex.

However, the company has tested Novodex on single-core and dual-core PC processors, as well as the Ageia chip, so those comparisons remain valid, Keane said. The benchmarks run on the PlayStation 3 and Xbox 360 simply tested to see if the SDK would run, and do not provide performance data, unlike the performance tests run on the PC systems.

In videos shown in the presentation, the physics computations of the boulder demonstration were running in software in the first case, typically between 4-6 frames per second on high-end PC processors, consuming almost all of the CPU cycles, Keane said. The second demo, performed using the PhysX processor, consumed about 20 percent of one processor and generated between 40 and 50 frames per second.

Keane, who was a former vice president of marketing at graphics-chip pioneer 3Dfx, said evaluating physics processing would prompt its own set of debates, in much the same way a graphics chip's performance is heavily scrutinized. "That's why we think that benchmarking physics is going to be as big a war as [benchmarking] a GPU," Keane said