14 DAY TRIAL // NVIDIA's next-generation graphics processing unit will be known as the GT300 and, according to recently revealed information, will provide users and application developers with a new and different architecture, compared to any solution that is currently available on the market. The GT300 will be launched as the world's first chip to take advantage of MIMD (Multiple-Instruction Multiple Data) architecture, as opposed to current GPUs from both NVIDIA and ATI, which are designed on SIMD (Single-Instruction Multiple Data) units. This change will reportedly enable NVIDIA's next-generation GPU to provide the consumer with a considerable boost in processing power.
The chip, expected to be released in a couple of months, is still wrapped under a veil of mystery, as specific details on the frequencies and memory type are yet to be unveiled. However, according to some recent articles on German-language Hardware-Infos and English-language Brightsideofnews, NVIDIA's upcoming GT300 GPU will be using an impressive 512 shader units, organized in sets of 32 cores. However, the real deal about GT300 is that it will be built using cores that rely on MIMD functions, capable of complex shader and computing operations on-the-go.
This feature alone is expected to provide NVIDIA's upcoming GPU with double the computing power of the company's current GT200 architecture. In theory, despite no indication on the speeds of the upcoming GT300-based graphics cards, a 2GHz clock speed for the 512MIMD will enable NVIDIA's card to deliver an impressive 3TFLOPS with Single Precision.
According to recent details, the new GT300 chip has the potential to impress everybody in the industry with the level of performance it can achieve. However, we still have to wait until we see the first samples of NVIDIA's next-gen GT300, which will most likely be available in the form of the GeForce GTX 380.