As always, there’s a flood of news coming out of NVIDIA’s annual GPU Technology Conference, being held this week in San Jose. In addition to the much-anticipated Pascal debut, which we’ll be reporting on later today, CEO Jen-Hsun Huang revealed an expansive update to NVIDIA’s SDK. Enhanced deep learning capabilities and support for the just announced Tesla P100 GPU are among the new features.
CUDA, of course, is the centerpiece. NVIDIA says more than a million developers have now downloaded the CUDA Toolkit and that there are more than 400 GPU-accelerated applications benefiting from CUDA software libraries. Here’s a quick tour of seven software updates in the SDK that NVIDIA says matter most:
1) Deep Learning – cuDNN 5, NVIDIA’s GPU-accelerated library of primitives for deep neural networks, now includes Pascal GPU support; acceleration of recurrent neural networks, which are used for video and other sequential data; and additional enhancements used in medical, oil & gas and other industries.
NVIDIA says these improvements will allow deep learning developers to rely on cuDNN’s optimized routines for designing and training neural network models, and reduce the need for low-level performance tuning. cuDNN accelerates frameworks such as Google TensorFlow, UC Berkeley’s Caffe, University of Montreal’s Theano and NYU’s Torch.
2) Accelerated Computing – CUDA8, the latest version of NVIDIA’s parallel computing platform, gives developers direct access to new Pascal features such as unified memory and NVLink. Also included is a new graph analytics library, nvGRAPH, which the company says can be used for robotic path planning, cyber security and logistics analysis, expanding the application of GPU acceleration in the realm of big data analytics.
“One new feature developers will appreciate is critical path analysis, which automatically identifies latent bottlenecks in code for CPUs and GPUs. And for visualizing volume and surface datasets, NVIDIA IndeX 1.4 is now available as a plug-in for Kitware ParaView, bringing interactive visualization of large volumes with high-quality rendering to ParaView users,” according to NVIDIA.
3) Self-Driving Cars – NVIDIA announced an end-to-end HD mapping solution for self-driving cars; it’s is built on top NVIDIA’s DriveWorks software development kit, the company’s deep learning platform for the automotive industry. DriveWorks provides libraries, tools and reference applications for automakers, tier 1 suppliers and startups developing autonomous vehicle computing pipelines.
4) Design Visualization – Iray, NVIDIA’s photorealistic rendering solution, has been adapted for virtual reality (VR) with the introduction of new cameras within Iray that let users create VR panoramas and view their creations with greater accuracy reports NVIDIA. Also announced was Adobe support of NVIDIA’s Materials Definition Language, bringing the possibility of physically based materials to a wide range of creative professionals.
Iray is used in a wide array of industries to create photorealistic models. Two prominent licensees are Dassault Systèmes and Siemens PLM. Iray is also available as a plug-in for popular software like Autodesk 3ds Max and Maya.
5) Autonomous Machines – The new cuDNN version 5 improves deep learning inference performance for common deep neural networks, allowing embedded devices to make decisions faster and work with higher resolution sensors. NVIDIA GPU Inference Engine (GIE) is the neural network inference solution for application deployment; developers can use GIE to generate optimized implementations of trained neural network models on NVIDIA GPUs, according to the company.
Robots, drones, submersibles and other intelligent devices require autonomous capabilities. The Jetpack SDK, which powers the Jetson TX1 Developer Kit, includes libraries and APIs for advanced computer vision and deep learning, enabling developers to build extraordinarily capable autonomous machines that can see, understand and even interact with their environments.
6) Gaming – Recently three technologies were added to NVIDIA GameWorks, a combination of development tools, sample code and advanced libraries for real-time graphics and simulation for games. The additions include Volumetric Lighting, Voxel-based Ambient Occlusion, and Hybrid Frustum Traced Shadows.
7) Virtual Reality – Several features have been added to VRWorks, NVIDIA’s suite of APIs, sample code and libraries for VR developers. One enhancement is Multi-res Shading, which accelerates performance by up to 50 percent by rendering each part of an image at a resolution that better matches the pixel density of the warped VR image. Also, VRWorks Direct Mode treats VR headsets as head-mounted displays accessible only to VR applications, rather than a normal Windows monitor in desktop mode.
NVIDIA says VRWorks helps headset and application developers achieve the highest performance, lowest latency and plug-and-play compatibility. Several partners – Sólfar Studios (Everest VR), Fusion Studios (Mars 2030), Oculus and HTC – are demonstrating these new technologies at GTC.