Other Nvidia GPUs
L40S, RTX and earlier Volta and Pascal cards. 113 providers on EveryNeocloud so far.
Beyond its flagship data centre chips, Nvidia sells many other GPUs that clouds rent out: the L40S and L4 of the Ada Lovelace generation, RTX workstation and gaming cards, the T4 for inference, and older data centre parts such as the Volta V100 of 2017 and the Pascal P100 of 2016. These cards fill the long tail of AI work, where a single cheaper GPU is enough.
The older generations
The P100 of 2016 brought high-bandwidth memory to Nvidia GPUs, and the V100 of 2017 introduced Tensor Cores, units built for the matrix maths of neural networks. The V100 powered Summit and Sierra in the US, Taiwania 2 and Japan's first ABCI. Earlier still, Titan at Oak Ridge used Kepler GPUs in 2012, one of the first large bets on GPUs for science.
Cards for inference and graphics
The L40S, launched in 2023, has 48 GB of memory and handles both AI inference and 3D graphics, and the small, low-power L4 suits video and light inference. Developer clouds such as Vultr and DigitalOcean rent these by the hour. Marketplaces also list many consumer RTX cards, such as the RTX 4090, which are cheap but lack the memory and fast links of data centre GPUs.
Famous other Nvidia GPUs providers
Still to be written: Vultr.
Further reading
Photo credits on this page
The small pictures above come from Wikimedia Commons under free licences. Some are cropped to fit.
- Summit: Carlos Jones/ORNL, CC BY 2.0, source
- Amazon Web Services: Raysonho @ Open Grid Scheduler / Grid Engine, CC0, source
- Titan: OLCF at ORNL, CC BY 2.0, source
- Tsubame: Kestrel, CC BY-SA 4.0, source
- Sierra: U.S. Department of Energy, Public domain, source