High-performance GPUs give hint of what to expect from next-gen workstation equivalents
At CES, Nvidia CEO Jensen Huang unveiled the hotly anticipated consumer-focused RTX 50-Series line up of Blackwell GPUs.
There was no mention of professional Nvidia RTX workstation GPUs built on the same GPU architecture. We expect these will follow in the next few months and the Blackwell GeForce boards should give a good indication of what to expect.
The flagship GPU – the GeForce RTX 5090 – comes with 32 GB of GDDR7 memory, which would suggest that professional Blackwell Nvidia RTX boards could go above the current max of 48 GB built into the Nvidia RTX 6000 Ada Generation.
The GeForce RTX 5090 is a dual slot PCIe Gen5 board, slimmer than the triple-slot GeForce RTX 4090, but draws more power (575 W compared to 450 W). If the workstation boards follow this upward power trend, this could give the workstation OEMs some thermal challenges, especially in multi-GPU workstations.
The GeForce RTX 5090 features 5th Generation Tensor cores for AI, boasting 3,352 TOPS, almost three times more than the Ada generation RTX 4090 (1,321 TOPS).
The GPU has fourth generation RT Cores for ray tracing with 2X the ray triangle intersection rate of the previous generation, and enhanced compression designed to reduce memory footprint. According to Nvidia, this allows Blackwell GPUs to ray trace levels of geometry that were never before possible.
Nvidia also introduced GeForce RTX 50 Series laptop GPUs, which will launch in Spring 2025. Features include new Max-Q technologies such as Advanced Power Gating, Low Latency Sleep, and Accelerated Frequency Switching which, according to Nvidia, increases battery life by up to 40%, compared to the previous generation.
Nvidia also announced details of DLSS 4, the latest release of its neural rendering technology, which uses AI to boost 3D performance.
DLSS 4, which includes DLSS Multi Frame Generation, will make its way into select design viz tools next month, including D5 Render and Chaos Vantage.
