by Anton Shilov
11/18/2009 | 07:21 AM
Nvidia Corp. on Wednesday displayed the first photo of its next-generation flagship graphics card based on the Fermi-GF100 (also known as NV60, G300, GT300) graphics processing unit (GPU). The graphics board seems to be complete antipode to the prototype/dummy board the company demonstrated two months ago.
The first image of the next-generation Nvidia GeForce graphics card shows a working board rendering the already well-known Unigine Heaven DirectX 11 benchmark. The photo was published in Nvidia’s blog on Facebook, thus, may be considered as the first official picture of the Fermi-GF100 graphics card. It is noteworthy that the first relatively public demonstration of Nvidia’s DirectX 11 GPU happens over five months later than the arch-rival ATI, graphics business unit of Advanced Micro Devices, showcased its Radeon HD 5800 chip in action at Computex Taipei 2009 in early June.
The new adapter itself is very long – over 27 centimeters (GeForce GTX 285 – 26.7cm, Radeon HD 4870 – 24.13cm, Radeon HD 5870 – 28.2cm) – and it features 8-pin + 6-pin PCIe power connectors, which implies that the GeForce Fermi-GF100 is rather power hungry, something not exactly surprising considering the fact that the GPU consists of over three billion of transistors. Earlier this year Nvidia demonstrated what it called “a prototype” of its Fermi graphics card, which seems to be short and featured only two 6-pin power connectors. Later on it was revealed that during the demo Nvidia only used a dummy non-working sample.
A good news is that Unigine Heaven benchmark does not seem to produce major artifacts with the current version of Nvidia’s drivers for the new breed of graphics cards. Unfortunately, the fact that a benchmark can be rendered on a graphics card does not give any clue regarding actual performance of the GPU that powers that card.
The flagship Fermi graphics processor will feature 512 stream processing engines (which are organized as 16 streaming multi-processors with 32 cores in each) that support a type of multi-threading technology to maximize utilization of cores. Each stream processor has a fully pipelined integer arithmetic logic unit (ALU) and floating point unit (FPU). The top-of-the-range chip contains 3 billion of transistors, features 384-bit memory GDDR5 memory controller with ECC and features rather unprecedented 768KB unified level-two cache as well as rather complex cache hierarchy in general. Naturally, the Fermi family is compatible with DirectX 11, OpenGL 3.x and OpenCL 1.x application programming interfaces (APIs). The new chips will be made using 40nm process technology at TSMC.
Chief executive of Nvidia said it would ramp up production of Fermi-GF100 graphics processor in Q1 of its fiscal 2011. Nvidia’s first quarter of fiscal year 2011 begins on the 26th of January and ends on the 26th of April, 2010.