From X-bit Labs: Nvidia Corp., a leading supplier of graphics processing units (GPUs) and multimedia chips, has reportedly received the first engineering samples of its upcoming flagship GPU from the fab, a report claims. However, another report says that manufacturing yield is so dramatically low that it is uncertain whether the company will be in position to start revenue shipments of the chip this December. Nvidia’s code-named GT300 (G300, NV60) graphics processor is rumoured to have rather tremendous die area of 452mm², which means that only about 104 - 105 of such dies can fit onto one 300mm wafer. The chip is made using 40nm fabrication process at Taiwan Semiconductor Manufacturing Company, which still cannot boast truly high yields, just about one and a half months ago TSMC reported about 60% yield on 40nm process tech. Since Nvidia GT300 is very large, its yields are naturally lower than those of mainstream chips (with smaller die sizes) at 40nm, hence, it will not be a surprise if GT300’s yields are in the range between 30% (the yield of first 40nm chips at the start of commercial production earlier this year) and 50% at the moment. According to information from a Japanese hardware news web-site, Nvidia received the first chips back from the fab in early September and in the best case scenario it will be in position to begin revenue shipments of the chip sometimes in December. However, SemiAccurate web-site claims that out of four 300mm wafers that Nvidia got from TSMC, only seven chips worked, which means that the yield is about 2%. At present it is unclear from the reports whether Nvidia has received the first experimental chips for internal testing and experiments or has got engineering samples, which it provides to partners, game developers, etc. Two things are more or less clear about GT300 at the moment: Nvidia does have samples of the chip, yields of the chip are rather low. View: Article @ Source Site |