VR-Zone

Ever before the release of GF100, a dual-GF100 product had been rumoured. The outrageous power consumption figures of the original GF100 meant a dual-GPU GF100 was unlikely, yet the rumour persisted, fueled by Galaxy and Mars exhibiting prototypes with exotic power solutions. However, no dual-GF100 made it to the market. Following the successful GF104 launch, a dual-GF104 was then rumoured, and unlike dual-GF100 feasible under 300W TDP. However, into 2011, and GTX 295 from 2009 remains the last NVIDIA dual-GPU product. At CES 2011, EVGA is exhibiting a dual-GPU card, which appears to be in its final design stages.



The first thing worth noticing is that the new dual-GPU card is a monster. Cooled by a hefty triple-fan cooler, the card is powered by 2x8-pin power, suggesting a TDP in the 375W range. This also breaks the PCI-SIG compliance standards. Furthermore, considering only EVGA has shown such a product, it may be a non-reference product.



On the rear, 8 memory chips are present; 4 per GPU. With no visible footprints for extra memory chips, it does appear likely that each GPU has 8 chips, or 1GB memory. This could suggest that the card is either GF114-based (also hinted at by the slightly rectangular GPU footprint) or a heavily crippled GF110 (less likely).

Three DVI outputs are present, with one SLI connector, enabling Quad-SLI.

Needless to say, this is not the first time we have seen a Fermi generation dual-GPU, but none have made it to the market thus far. However, the EVGA dual-GPU does seem most likely to hit retail at some point.

Source: TechReport