CES 2024 is an exciting time for a plethora of new product announcements, and expected to be chief among them this year are new graphics card options. While gamers are patiently awaiting the details of NVIDIA's upcoming GeForce RTX Super GPUs, there is some commotion in the data center market as well.
NVIDIA has announced the RTX 5880 Ada generation graphics card, packing a whopping 48GB of VRAM. While gamers may salivate at the thought of so much VRAM, it's nearly a requirement for today's modern workstation workloads and AI implementations. There are some potential caveats to be aware of with this GPU, however. Compared to the NVIDIA RTX 6000 Ada GPU, this RTX 5880 has less CUDA cores.
This restriction is likely to help NVIDIA navigate the stringent requirements and export bans of recent powerful GPUs, in order to be in compliance while still releasing a capable product. It also gives NVIDIA another price point to hit, for professionals who don't necessarily want to splurge on NVIDIA's flagship workstation card.
These are workloads that NVIDIA expects users to utilize this GPU for
With AI becoming such a heavy-hitting industry force, the demand for literally any capable GPU has outpaced the ability for NVIDIA to produce them in a timely fashion. Whilst a good problem to have, it has come with its own logistical headaches showcased with the announcement of this RTX 5880 GPU. While this GPU can be useful in a myriad of scenarios such as 3D modeling and video content creation, the Ai data center market is equally as hungry to gobble it up.
This is not the first cut-down version of its popular GPUs that NVIDIA has produced, with a GeForce RTX 4090D having a similar fate.
NVIDIA's listing of some key specifications for the RTX 5880
Worry not, the RTX 5880 still has some impressive specifications. Using the AD102 GPU WITH 14,080 CUDA cores and 440 tensor cores, it also packs 48GB of VRAM with a 384-bit bus. While it won't perform as well as the RTX 6000, it also has a lower TDP of 285 watts to make up for some of that decreased capability.
While pricing and availability is still to be seen, it appears this may be a GPU that is available globally and not just specifically for the Chinese market. With worldwide demand of AI data center GPUs at a high frequency currently, this is likely a solid decision to keep the market satisfied.
