40nm graphics chip battle heats up

Chip foundry TSMC says it has fixed its new 40nm process line so that yield rates to deliver quality on a par with its 65nm node process.

Mark Liu, Senior VP of Operations at TSMC made the claim as TSMC opened up a new chip shop yesterday.

According to a report at Digitimes, Liu said yields had be adversely affected by chamber-matching problems which have now been resolved.

TSMC's Fab 12, Phase 5 scheme is preparing to operate a 28nm process.

Both AMD and Nvidia are using TSMC's 40nm line to manufacture their newest graphics chip offerings, so both firms will benefit from TSMC's announcement. AMD is already offering a full line-up of 40nm DirectX 11 chips in the shape of its "Evergreen" family of graphics chips the ATI Radeon HD 5000-series. Nvidia's upcoming Fermi chips are still in the pipeline and should be launched by March.

The transition to 40nm should mean both firms can crank up the volume by delivering more chips at less cost. AMD currently has an advantage, both in bragging rights and in tangible product offerings, as well as the fact that it can currently claim to offer the fastest graphics chip around.

The ball's back in Nvidia's court in the seemingly never-ending game of real-life Pong.