NVIDIA at this time introduced NeuralVDB, which brings the facility of AI to OpenVDB, the industry-standard library for simulating and rendering sparse volumetric information, akin to water, fireplace, smoke and clouds.

Building on the previous decade’s improvement of OpenVDB, the introduction at SIGGRAPH of NeuralVDB is a game-changer for professionals working in areas like scientific computing and visualization, medical imaging, rocket science and visible results. By decreasing reminiscence footprint by as much as 100x, it permits creators, builders and researchers to work together with extraordinarily massive and sophisticated datasets in actual time.

Over the previous decade, OpenVDB has earned Academy Awards as a core expertise used all through the visual-effects {industry}. It has since grown past leisure to industrial and scientific use instances the place sparse volumetric information is prevalent, akin to industrial design and robotics.

Last 12 months, NVIDIA launched NanoVDB, which added GPU help to OpenVDB. This delivered an order-of-magnitude speedup, enabling quicker efficiency and simpler improvement — and opening the door to real-time simulation and rendering.

NeuralVDB builds on the GPU acceleration of NanoVDB by including machine studying to introduce compact neural representations that dramatically scale back its reminiscence footprint. This permits 3D information to be represented at even increased decision and at a a lot bigger scale than OpenVDB. The result’s that customers can simply deal with huge volumetric datasets on units like particular person workstations and even laptops.

NeuralVDB affords a major effectivity enchancment over OpenVDB by compressing a quantity’s reminiscence footprint as much as 100x in comparison with NanoVDB. This permits customers to transmit and share massive, advanced volumetric datasets way more effectively.

To speed up coaching as much as 2x, NeuralVDB permits the weights of a body for use for the next one. NeuralVDB additionally permits customers to realize temporal coherency, or easy encoding, through the use of the community outcomes from the earlier body.

Hitting this trifecta of dramatically decreasing reminiscence necessities, accelerating coaching and enabling temporal coherency permits NeuralVDB to unlock new potentialities for scientific and industrial use instances, together with huge, advanced quantity datasets for AI-enabled medical imaging, large-scale digital twin simulations and extra.

Learn extra about NeuralVDB.

Watch the NVIDIA particular deal with at SIGGRAPH on demand, and be part of NVIDIA on the convention by Thursday, Aug. 11, to see extra of the newest expertise breakthroughs in graphics, AI and digital worlds.