NVIDIA has just announced what is being touted as the world’s most advanced GPU for professionals – the Quadro GV100. This card is based on NVIDIA’s Volta architecture and packs in 5120 CUDA cores. The GV100 is made for many different workloads from machine learning to HPC, it is actually the only professional graphics card capable of processing workloads around the new RTX raytracing suite. That by itself should make the card sell quickly.
Along with the 5120 CUDA cores you have 32 GB of HBM2 memory and up to 7.4 TFLOPS of power for double-precision rendering, 14.8 TFLOPS for single-precision wordkloads, 29.6 TFLOPs half-precision, and 118.5 TFLOPs for deep learning through its Tensor Cores. The current Quadro GP100 only offered 10.3 TFLOPs for single-precision rendering.
So you have 5120 CUDA cores, 640 Tensor cores, 320 texture units, and 128 ROPs which run at 1450 MHz boost clocks. The HBM2 will run a 1.7 Gbps on a 4096-but bus, which does feature ECC (error correction for data-critical workloads). The TDP for this card is 250W, which is not bad compared to the 230W of the Quadro P100.
As far as connectors go you have four DisplayPort 1.4 connectors, so NVIDIA has dropped the DVI port that was on the Quadro GP100. The card features standard Quadro Sync and Stereo connectors for synchronized refresh and quad-buffered stereo. You also have support for dual NVIDIA NVLink operation.