Quadro GV100

The Quadro GV100 is NVIDIA's Volta-generation professional workstation GPU, launched in early 2018 and the predecessor to the Turing-based Quadro RTX series. Built on the full GV100 die — the same silicon used in the Tesla V100 data centre accelerator — it occupies a unique position in the workstation lineup, pairing exceptionally high memory bandwidth and strong double-precision FP64 compute with HBM2 memory rather than the GDDR6 used in successor Quadro cards.

The card carries 32GB of HBM2 memory with ECC support across a 4096-bit bus, delivering 870 GB/s of bandwidth. It features 5120 CUDA cores for 14.8 TFLOPS of FP32 throughput and class-leading double-precision compute. In a dual-slot form factor on PCIe 3.0 x16 rated at 250W, two cards can be bridged via NVLink to pool 64GB of combined ECC framebuffer. The Quadro GV100 predates dedicated RT cores and runs NVIDIA's ISV-certified Quadro professional drivers.

The Quadro GV100 was designed for scientific simulation, computational fluid dynamics, AI/ML training and inference, medical imaging, and high-precision engineering workloads that benefit from its exceptional memory bandwidth and FP64 throughput. It remains relevant in research and simulation environments where HBM2 bandwidth and double-precision compute capability are the primary requirements.