Nvidia Launches Quadro RTX 4000 GPU With 2,304 CUDA Cores

Nvidia's Quadro RTX series continued growing today with the addition of the Quadro RTX 4000 professional graphics card powered by the company's most recent Turing architecture.

The Quadro RTX 4000 serves as the midrange model in Nvidia's Quadro RTX family, which is currently comprised of the Quadro RTX 8000, RTX 6000 and RTX 5000. Similar to the GeForce GTX 2080 and the Quadro RTX 5000, the new Quadro RTX 4000 is built around Nvidia's Turing TU104 silicon that's produced under TSMC's 12nm FinFET manufacturing process. However, Nvidia has disabled a couple of Streaming Multiprocessors to meet the target shader count.

The Quadro RTX 4000 comes equipped with 2,304 CUDA cores, 288 dedicated Tensor cores for AI and 26 RT cores for real-time ray tracing applications. Unfortunately, Nvidia hasn't yet revealed the graphics card's operating frequency. The chipmaker also outfitted the Quadro RTX 4000 with 8GB of GDDR6 memory on a 256-bit memory bus. Nvidia says the Quadro RTX 4000 can deliver up to 7.1 TFLOPS of single precision (FP32) performance, 6 GigaRays/sec of ray tracing performance and 43 Tera RTX-OPS. 

Measuring 4.4 x 9.5 inches (11.176 x 24.13 cm), the Quadro RTX 4000 features a slim, single-slot blower design and should fit without problems in the majority of workstation cases. With a moderate 160W TDP rating, the Quadro RTX 4000 only depends on the PCIe 3.0 x16 interface and a single 8-pin PCIe power connector for power. Display outputs on the graphics card include three DisplayPort 1.4 ports and a VirtualLink USB-C port.

The Quadro RTX 4000 has an estimated price tag of $900 (~£694.18). The graphics card will be available on Nvidia's website in December from the chipmaker's workstation partners in the likes of Dell, HPE and Lenovo, as well as its authorized distribution partners in North America, Europe and Asia-Pacific.