14 DAY TRIAL // The race to the next-generation 40nm-based graphics processing unit continues to be one of the main issues concerning speculators in the industry. Both NVIDIA and ATI, the world's leading vendors of GPUs, have so far announced their first desktop and mobile solutions designed using said manufacturing process, but both companies are yet to unveil their next-generation chip architectures. On that note, according to recent details in the industry, the Santa Clara, California-based AMD is said to miss its one-year refresh cycle for GPUs by several additional months.
For those who have been keeping a close eye on what's going on in the industry, the situation is more or less surprising. According to a recent news-article on Hot Hardware, the chip maker appears to be affected by the same reported issues with 40nm yields that TSMC has been having for a while now. These problems are affecting both AMD and NVIDIA, as both companies use TSMC's facilities and process technology, which could prove as a relatively good thing for Advanced Micro Devices. However, it's clear that the world's leading graphics chip makers are struggling to get their new GPUs out in time for the launch of Microsoft's Windows 7 operating system.
There have been previous reports indicating that TSMC's 40nm yields were below 30 percent, something that might explain the situation with AMD's first 40nm-based card, the Radeon 4770. Since its debut, late this April, the card has been scarce to find, which was an indication that TSMC's problems are still affecting its customers.
AMD could turn to its recent spinoff, Globalfoundries, but the chip may need to wait until the latter has the necessary facility to develop bulk silicon, expanding its offering from the current IBM SOI (Sillicon-on-Insulator) technology.