Nvidia introduces the RTXA4500. The workstation GPU is designed for rendering simulations, intensive video content and implementing AI.
Its specifications aren’t groundbreaking. The idea behind the RTX A4500 is a middle ground between two previously introduced Nvidia GPUs: the A4000 and A5000. The amount of CUDA cores (7186), Tensor cores (244), RT cores (56), and GDDR6 working memory (20GB) is higher than the A4000 and lower than the A5000.
The computing power of 23.7 teraflops lies between the A4000 and A5000 as well. Despite the latter making it very likely that its price will fall in the middle of the A4000 (~€1,300) and A5000 (~€3,000), Nvidia has yet to announce an official price. In addition, the marketing page is ready, but the exact time of availability remains unknown.
The RTX A4500 supports NV-link. The technology allows multiple NV-link supported GPUs to be connected and deployed as a whole. A welcome addition, as it makes older GPUs reusable in new assemblies, and keeps the A4500 deployable after possible future upgrades.