NVIDIA DESIGNWORKS Ankit Patel - ankitp@nvidia.com Prerna Dogra - - - PowerPoint PPT Presentation

nvidia designworks
SMART_READER_LITE
LIVE PREVIEW

NVIDIA DESIGNWORKS Ankit Patel - ankitp@nvidia.com Prerna Dogra - - - PowerPoint PPT Presentation

NVIDIA DESIGNWORKS Ankit Patel - ankitp@nvidia.com Prerna Dogra - pdogra@nvidia.com 1 Autonomous Driving Deep Learning Visual Effects Virtual Desktops Gaming Product Design Visual Computing is our singular mission 2 RENDERING PHYSICS


slide-1
SLIDE 1

1

Ankit Patel - ankitp@nvidia.com Prerna Dogra - pdogra@nvidia.com

NVIDIA DESIGNWORKS

slide-2
SLIDE 2

2

Autonomous Driving Deep Learning Visual Effects Virtual Desktops Gaming Product Design

Visual Computing is our singular mission

slide-3
SLIDE 3

3

https://developer.nvidia.com/designworks

RENDERING

Multi-Display Capture SDK Warp and Blend GPUDirect for Video Video Codec SDK Iray SDK OptiX SDK MDL SDK vMaterials NV Pro Pipeline GVDB Voxels VXGI GRID SW MGMT SDK NVAPI/NVWMI PhysX

PHYSICS VOXELS VIDEO MANAGEMENT DISPLAY

slide-4
SLIDE 4

4

NVIDIA GPU GRAPHICS DRIVER

Iray OptiX PhysX NvPro Pipeline VXGI GVDB Video Codec GPU Direct for Video MDL Mosaic Warp and Blend Capture GRID Management NVAPI

vMaterials

Application Partners End-Users (Designers, Artists, Scientists)

Tools and technologies for Professional Visualization Application Developers

CUDA DRIVER

NVWMI

slide-5
SLIDE 5

5

RENDERING

slide-6
SLIDE 6

6

IRAY SDK

developer.nvidia.com/iray-sdk

  • Use Case: physically based rendering for product

design

  • Unprecedented visual quality and fidelity;

enabling fluid and interactive product design flow

Rendering by simulating the physical behavior

  • f lights and materials

Iray 2017

slide-7
SLIDE 7

7

MDL SDK

Material Definition Language for seamless and quick integration of physically based materials into renderers

developer.nvidia.com/mdl-sdk

MDL support in Iray for Maya Physically based Materials

MDL SDK 2017

  • Use Case: physically based rendering for product

design

  • Enables designers and artists to understand how

materials impact product design

slide-8
SLIDE 8

8

vMaterials

Library with hundreds of ready to use real world materials

Collection of MDL materials vMaterials for Product Design

  • Use Case: physically based rendering for product

design

  • Library of verified materials for customers to use

with MDL enabled renderers

developer.nvidia.com/vmaterials

slide-9
SLIDE 9

9

OPTIX RAY-TRACING ENGINE

  • Use Case: Ray tracing for design, defense, and

audio synthesis applications

  • OptiX does the “heavy lifting” of ray tracing:

traversal, intersection, acceleration, allowing you to concentrate on rendering technique

developer.nvidia.com/optix

Rendered with Iray powered by OptiX

CUDA accelerated framework for high performance programmable ray-tracing

OptiX in Pixar’s Lighting tool OptiX 4.1

slide-10
SLIDE 10

10

NVIDIA Pro Pipeline

Matryoshka IRay Realtime Scene Rendering Crash Simulation Scene - ESI group

Example high performance rendering pipeline

  • Use Case: OpenGL or Vulkan viewport rendering
  • Modularized architecture for reuse in custom

rendering pipelines

  • Open-Source project lead by NVIDIA

developer.nvidia.com/nvidia-pro-pipeline

slide-11
SLIDE 11

11

VOXELS

slide-12
SLIDE 12

12

VXGI

An implementation of a global illumination algorithm known as Voxel Cone Tracing

Scene View and Voxel Opacity Model View Scene with Global Illumination - realism of rendered images

  • Use Case: Professional Rendering applications

with heavy geometry

  • Adding Global illumination to your scene improves

the realism of rendered images

  • Reduces Content creation time
slide-13
SLIDE 13

13

GVDB VOXELS SDK

Framework for simulation, compute and rendering of sparse voxels on the GPU

Motion Pictures Scientific Simulation

3D Printing

Process Engineering

NEW

  • Use Case: Process Engineering, 3D Printing,

Motion Pictures and Scientific simulations

  • Inspired by Academy Award-winning OpenVDB

library, GPU accelerated Voxel computing.

  • The finished part has a specific shape, while the

internal voronoi can vary in cell density

slide-14
SLIDE 14

14

3.2 lbs 5.3 lbs

slide-15
SLIDE 15

15

slide-16
SLIDE 16

16

PHYSICS

slide-17
SLIDE 17

17

PHYSX SDK

developer.nvidia.com/physx-sdk

Model demonstrating PhysX Clothing

Engine to generate realistic Physical effects with HW accelerated physics simulation

  • Use Case: Real-time product design investigation
  • Scalable, multi-platform solution supporting wide

range of devices

  • Enabled Haptic Feedback for OPTIS’S Virtual build

process simulator, Dental training simulation systems – understand difference between collision with a tooh or the gum.

slide-18
SLIDE 18

18

VIDEO

slide-19
SLIDE 19

19

VIDEO CODEC SDK

GAME STREAMING (GEFORCE Now) – ULTRA LOW LATENCY

Interactive, single frame latency, 4K, HDR

GPU VIRTUALIZATION – QUALITY & RELIABILITY

Capture + encode, low-latency, multiple display, lossless

VIDEO TRANSCODING – PERFORMANCE/WATT

Codecs, resolution, quality, bitrate, Live, broadcast

VIRTUAL REALITY – ACCURACY

Video frame interpolation(ASW), Camera stitching, Computer Vision applications

API for GPU-accelerated Video Encode and Decode

developer.nvidia.com/video-codec-sdk

NVIDIA Video Codec SDK as used in VMware Horizon running on NVIDIA GRID

Video Codec SDK 8.0

slide-20
SLIDE 20

20

CPU

NVDEC NVENC CUDA Cores

Buffer

Decode HW* Encode HW*

Formats:

  • H.264
  • H.265

Bit depth:

  • 8 bit
  • 10 bit

Color**

  • YUV 4:4:4
  • YUV 4:2:0
  • Lossless

Resolution

  • Up to 8K***

Formats:

  • MPEG-2
  • VC1
  • VP8
  • VP9
  • H.264
  • H.265

Bit depth:

  • 8 bit
  • 10 bit
  • 12 bit

Color**

  • YUV 4:2:0
  • YUV 4:4:4
  • Lossless

Resolution

  • Up to 8K***

* See support diagram for previous NVIDIA HW generations ** 4:2:2 is not natively supported on HW *** Support is codec dependent

VIDEO CODEC SDK

slide-21
SLIDE 21

21

GPUDIRECT FOR VIDEO

  • For IO board manufacturers who write device drivers
  • Stream video in and out of GPU memory at sub-frame

transfer times

  • OpenGL, DirectX and CUDA on Windows or Linux

Transfer video in and out of GPU memory

developer.nvidia.com/gpudirectforvideo

Comparison of systems with and without GPUDirect for Video

slide-22
SLIDE 22

22

DISPLAY

slide-23
SLIDE 23

23

MOSAIC

  • Use Case: Create a stereoscopic 3D CAVE, build an

expansive digital signage wall

  • Connect up to 16 displays from a single workstation

Multi-display scaling and synchronization

developer.nvidia.com/capture-sdk

Bezel Correction Seamless Image Projector Overlap Desktop Management

slide-24
SLIDE 24

24

CAPTURE SDK

  • Use Cases: Streaming, Remoting and broadcasting

applications

  • Enables developers to easily and efficiently capture and
  • ptionally encode display content or application’s

graphics context

  • Capture up to 8K resolution

High quality low latency capture

VMware Horizon Blast uses NVIDIA Capture SDK to increase number of users per server through improved latency, reduced bandwidth and lower CPU utilization. developer.nvidia.com/capture-sdk

Image Courtesy Accuweather.

slide-25
SLIDE 25

25

WARP AND BLEND

Framework that allows display applications to combine multiple projectors to one single display.

developer.nvidia.com/warp-and-blend

Simulator environment with large curved front displays Image courtesy of Joachim Tesch. Max Planck Institute for Biological Cybernetics.

  • Use Case: Immersive VR, multi-projection display walls,

simulator environment for interactive presentations, projection mapping onto scale physical models

  • Adjust geometry and intensity for multi-projector

systems

  • Minimal performance delay to display pipeline

compared to external boxes using FPGAs

slide-26
SLIDE 26

26

MANAGEMENT

slide-27
SLIDE 27

27

NVIDIA WMI SDK

Remotely control NVIDIA Quadro, GRID and NVS GPUs.

developer.nvidia.com/nvwmi-sdk

Remote administration and configuration of NVIDIA professional GPUs in Windows machines (virtual or physical) on corporate networks

  • For Windows IT admins to remotely perform

administrative tasks. ○ Configure GPU settings ○ Retrieving GPU information ○ Perform automated tasks

slide-28
SLIDE 28

28

GRID SW MANAGEMENT SDK

Build monitoring and management solutions

  • ptimized for NVIDIA GRID

developer.nvidia.com/grid-software-management-sdk

Available on compatible NVIDIA GRID-certified servers with Tesla M6, M10, or M60

Platforms:

  • Real-time vGPU utilization data
  • Manage the lifecycle of a VDI deployment
  • Monitor the performance of GPU-accelerated VDI

environments

slide-29
SLIDE 29

29

ADDITIONAL SDKs

developer.nvidia.com/vrworks developer.nvidia.com/gameworks

Cutting-edge technology for game development Amazing virtual reality experiences

slide-30
SLIDE 30

30

DESIGNWORKS - GTC 2017

S7454 - NVIDIA ADVANCED RENDERING S7236 - RAYTRACING INTERIORS FOR STILLS AND VR WITHIN MINUTES WITH IRAY INTERACTIVE S7357 - WARPING & BLENDING FOR MULTI-DISPLAY SYSTEM USING NVIDIA DESIGNWORKS S7352 - SEE THE BIG PICTURE: HOW TO BUILD LARGE DISPLAY WALLS USING NVIDIA DESIGNWORKS APIS/TOOLS S7455 - INTRODUCTION TO THE NVIDIA OPTIX RAY TRACING ENGINE S7583 - HOMEBYME: HOW IRAY, VCA, DEEP LEARNING AND VR HELPS YOU EXPERIENCE YOUR NEW APARTMENT BEFORE IT IS BUILT S7453 - NVIDIA ADVANCED RENDERING PRODUCTS FOR END USERS S7194 - LIGHT BAKING WITH IRAY S7440 - CREATE HIGH-QUALITY MATERIALS FROM SCANS WITH MDL AND SUBSTANCE S7810 - ACCELERATION OF MULTI-OBJECT DETECTION AND CLASSIFICATION TRAINING PROCESS WITH NVIDIA IRAY SDK (PRESENTED BY SAP) S7328 - THE NVIDIA IRAY LIGHT TRANSPORT SIMULATION AND RENDERING SYSTEM L7109 - NVIDIA GRID DEPLOYMENT S7756 - INDUSTRIAL-GRADE HAPTICS WITH HAPTX AND PHYSX S7424 - INTRODUCTION AND TECHNIQUES WITH NVIDIA GVDB VOXELS S7425 - 3D PRINTING WITH NVIDIA GVDB VOXELS S7111 - NVIDIA VIDEO TECHNOLOGIES AND SDK: OVERVIEW AND LATEST UPDATES

slide-31
SLIDE 31

31

QUESTIONS?