Compute Updates (BTN: September 26th, 2023)
As part of keeping the community informed on important Render Network updates, we’d like to take the next few weeks to provide detail on key Render Network initiatives.
The Render Network has made progress in the last few weeks on the proposed expansion of the Render Network into new forms of AI and ML workloads, as proposed in RNP-004. Recently a key milestone was achieved with IO.net (formerly ANTBIT) demonstrating test compute jobs running on actual Render Nodes during the Ray Summit.
What was significant about the demo was the processing of real AI training and inference workloads on-demand across a decentralized pool of GPUs — ranging from 3090s to higher performance devices like 4090’s, A100s, and A6000s. This aggregation of decentralized supply, where decentralized hardware configurations and GPU models are simply configured and rented, was an essential feature of the preview.
The use of Render Network GPUs is one of the first demonstrations of decentralized computing to solve the compute shortage facing the AI industry, where supply can be successfully aggregated from a massive pool of underutilized GPUs around the world that have been leftover from proof of work mining, 3D rendering and cloud gaming applications.
While this is a significant milestone, this does not mean the first Compute client is yet live on the Render Network.
Up next, io.net will commence integrating their platform with the Render Network API, which will facilitate both rendering and new forms of AI and ML compute jobs being seamlessly orchestrated by the Render Network. The Render Network core team will also need to expand its API’s to facilitate io.net’s model of reserving compute.
In the near future, the Render Foundation will be announcing an opt-in process for nodes to be a part of the pool used for the Compute client.
Current and future Node operators should actively follow the Render Network social channels to stay updated.
Do not opt-in directly with io.net, as you will only be able to run both render and compute jobs if you opt-in directly via the Render Network.
More information regarding the Compute client, BME and other initiatives will be released in the upcoming presentation at Breakpoint 2023 in October.
Join us in the Rendering Revolution at:
Website: https://render.x.io
Twitter: https://twitter.com/rendernetwork
Knowledge Base: https://know.rendernetwork.com/
Discord: https://discord.gg/rendernetwork
Telegram: https://t.me/rendernetwork
Render Network Foundation: https://renderfoundation.com/
For developers looking to leverage the Render Network for GPU compute, the Request for Compute continues to be open to help the network expand support for a wider range of AI, ML, and 3D visualization applications.
