Nvidia RTX3080 parameter exposure 3840 stream processor 20G memory

With 2020 approaching, expectations are growing for the Nvidia GeForce RTX next-generation flagship card. According tomedia TweakTown, the latest revelations about the RTX 3080 and 3070 configurations have been released.

Nvidia RTX3080 parameter exposure 3840 stream processor 20G memory

Nvidia’s latest GPUs: GA103 and GA104 are widely considered to correspond to the RTX 3080 and 3070. Early parameters indicate that the RTX 3080 will be equipped with 3840 stream processors, with a memory of 10 or 20GB of GDDR6 and a bit width of 320bit. The RTX is slightly inferior, with a 3072-stream processor with a memory of 8/16GB and a graphics card width of 256bit.

It is reported that the new flagship card architecture for ampere architecture, the use of 7nm process, but there are other reports that TSMC’s production capacity is insufficient, Nvidia may use Samsung contract.

Nvidia RTX3080 parameter exposure 3840 stream processor 20G memory

The most mysterious is the Ampere GA102, which is the corresponding RTX 3080Ti or the new TITAN-class graphics card. Only its bit width is given in the blast, which is 384bit wide, and the other parameters are still a mystery. Compared with the 20X0 series, the 3080 and 3070 have a huge improvement in the memory, and the performance is bound to take a leap forward. In addition, it has been reported that Nvidia may officially announce a new product line in March and launch its next-generation flagship graphics card in June 2020.