The Radeon RX 6900 XT appeared in Geekbench and of course, the question arises as to whether it can overtake the Geforce RTX 3080 or not. In terms of synthetic values, it is not quite enough at the moment, but it is narrow enough to perhaps make a difference.
The first Radeon RX 6900 XT appeared in the Geekbench OpenCL benchmark database. The fully equipped 21-Hertz Navi thumps on the map as an XTX version with 40 computes units, which are specified in Geekbench with a maximum frequency of 2 GHz. The card leads 5,120 shaders into the field, together with 16 gigabytes of memory – and this should at least beat the Geforce RTX 3080, for which the already tested Navi 21 cards are a little too slow.
It remains to be seen whether this will succeed in total. In any case, the values from the synthetic Geekbench suggest that the Radeon RX 6900 XT with 169,779 OpenCL points is 12 percent faster than a Radeon RX 6800 XT and 35 percent faster than a Radeon RX 6800. In comparison: the RX 6900 XT has 11 Percent more shader ALUs than the RX 6800 XT and 33 percent more than the RX 6800.
In Geekbench, that wouldn’t be enough against the Geforce RTX 3080 or even the RTX 3090. The Geforce RTX 3080 would be within range with 4 percent, but still a touch faster; the RTX 3090, meanwhile, is 19 percent ahead. When it comes to raster performance, it doesn’t have to be a problem. If you don’t count the frames individually, you get enough power for most situations. Meanwhile, the 16 GiByte memory speak for the Radeon; For the Geforce, the better performance in RT titles is likely, if one uses previous experience.
In a mixed index value, however, the Radeon RX 6900 XT is possibly a little closer, as the RX 6800 XT is only 10 percent missing from the RTX 3080. The prognosis that it could be tight would not be completely absurd. The customer would then have to decide whether he would like to spend 1,000 euros (RRP) for the Radeon or 700 euros (RRP) for the Geforce RTX 3080, which has 6 GiByte less memory.