At Siggraph 2024, the preeminent conference in the graphics industry, Meta and Nvidia stood out with a series of impressive announcements that underscored their dominance in shaping the future of graphics technology. As the event draws researchers, architects, and engineers from across the industry, both companies showcased pivotal advancements in their respective fields.
Nvidia Expands OpenUSD Ecosystem with New NIMs
Nvidia has long been a driving force behind the OpenUSD standard, a pivotal format initially developed by Pixar and later open-sourced to become OpenUSD. This format is integral to Nvidia’s Omniverse cloud platform, which facilitates cross-platform XR applications. At Siggraph 2024, Nvidia introduced a range of new NIM (Nvidia Interface for Microservices) microservices designed to enhance USD workflows.
The newly unveiled NIMs include the USD Code NIM, USD Search NIM, and USD Validate NIM. These services aim to streamline generative AI model deployment, simplifying tasks such as code generation, asset search, and validation. The USD Search NIM, in particular, is anticipated to be a game-changer for developers by simplifying the process of locating 3-D assets through natural language or image queries—a significant enhancement for creative and engineering applications.
In addition to these previewed NIMs, Nvidia revealed upcoming tools like the USD Layout NIM and USD SmartMaterial NIM. USD Layout allows users to create OpenUSD-based scenes from text prompts, while SmartMaterial applies realistic materials to CAD objects. These tools not only simplify 3-D workflows but also deepen the integration of developers and users into Nvidia’s ecosystem, reinforcing the reliance on NIMs alongside Omniverse.
Nvidia also introduced three new fVDBs (framework volumetric data blocks), built on the OpenVDB library. These fVDBs represent a leap in volumetric data simulation and rendering, offering four times the spatial scale and processing speed of previous frameworks. The fVDB Mesh Generation NIM creates OpenUSD-based meshes from point cloud data, while the fVDB Physics Super-Res NIM provides high-resolution physics simulations. The fVDB NeRF-XL NIM enables the creation of large-scale neural radiance fields in OpenUSD, facilitating rapid development of 3-D environments from 2-D image captures.
Meta Unveils Cutting-Edge Research Projects
While Nvidia made headlines with its technical advancements, Meta also presented significant research initiatives that capture the imagination of the graphics community. Although detailed announcements from Meta were not covered in this update, their contributions to the field continue to drive innovation and push the boundaries of what’s possible in graphics technology.
Overall, Siggraph 2024 served as a showcase of how Meta and Nvidia are leading the way with pioneering technologies and tools, setting new standards for the industry and offering a glimpse into the future of graphics and AI.
Related topics:
Can Sora Revolutionize Educational Practices?