Nvidia CEO showed Quadro GV100 graphics card

CEO of NVIDIA Jensen Huang has just introduced the latest graphics card from the Quadro series. The name Quadro GV100 is based on the Volta architecture of NVIDIA. Graphic card is carrying 5120 CUDA core, almost double the previous series that carries the Pascal architecture.

GV100 comes to work on a wide range of workloads, from machine learning to High Perfomance Computing (HPC). NVIDIA's latest graphics card brings together 5120 CUDA cores with 32 GB of HBM2 memory, 7.4 TFLOPs for double-precision rendering, 14.8 TFLOPs for single-precision workload, 29.6 TFLOPs half-precision and 118.5 TFLOPs for deep learning through its Tensor Cores. In comparison, Quadro GP 100 is only able to offer the ability of 10.3 TFLOPs for single-precision rendering. Quadro GV100 is also the only professional graphics card capable of running RTX raytracing suite.

Specifications are carried by the graphic card is indeed remarkable. Equipped with 5120 CUDA cores, 640 Tensor cores, 320 texture units and 128 ROPs. The GPU itself runs at speeds up to 1450 MHz boost clock. GV100's HMB 2 memory runs with bandwidth of 1.7 Gbps at 4096 bit bus, comes with ECC error correction. Everything comes with a 250 watt TDP, which is bigger than a GP100 of 20 watts.

Quadro GV100 brings four DisplayPort 1.4 to its display output. Graphic card also comes with standard Quadro features, such as Quadro Sync and Stereo connectors. This graphics card also supports NVIDIA NVLink dual.