Copyright © 2017 Your Company Name 1
Your Company Logo goes here.
See Slide 8 for instructions
Neil Trevett, Khronos President
VP Developer Ecosystem, NVIDIA
May2017
Vision Acceleration API Landscape:
Options and Trade-offs

© Copyright Khronos Group 2017 - Page 2© Copyright Khronos Group 2017 - Page 2© Copyright Khronos Group 2017 - Page 2© Copyright Khronos Group 2017 - Page 2
The Search for Vision Performance and Portability
Leads to a layered Ecosystem – and Graph Programming
Explicit
Kernels
Diverse Hardware
(GPUs, DSPs, FPGAs)
GPU FPGA
DSP
Dedicated
Hardware
Single Source C++
Languages
Graph
Programming
TensorRT
Code Intermediate
Representations PTX HSAIL

© Copyright Khronos Group 2017 - Page 3
© Copyright Khronos Group 2017 - Page 3
OpenVX – Performance Portable Vision
• OpenVX developers express a graph of image operations (‘Nodes’)
- Using a C API
• Nodes can be executed on any hardware or processor coded in any language
- Implementers can optimize under the high-level graph abstraction
• Graphs are the key to run-time optimization opportunities…
Array of
Keypoints
YUV
Frame
Gray
Frame
Camera
Input
Rendering
Output
Pyrt
Color
Conversion
Channel
Extract
Optical
Flow
Harris
Track
Image
Pyramid
RGB
Frame
Array of
Features
Ftrt-1OpenVX Graph
OpenVX Nodes
Feature Extraction Example Graph

© Copyright Khronos Group 2017 - Page 4© Copyright Khronos Group 2017 - Page 4© Copyright Khronos Group 2017 - Page 4© Copyright Khronos Group 2017 - Page 4
OpenVX Efficiency through Graphs..
Reuse
pre-allocated
memory for
multiple
intermediate data
Memory
Management
Less allocation overhead,
more memory for
other applications
Replace a sub-
graph with a
single faster node
Kernel
Fusion
Better memory
locality, less kernel
launch overhead
Split the graph
execution across
the whole
system: CPU /
GPU / dedicated
HW
Graph
Scheduling
Faster execution
or lower power
consumption
Execute a sub-
graph at tile
granularity
instead of image
granularity
Data
Tiling
Better use of
data cache and
local memory

© Copyright Khronos Group 2017 - Page 5© Copyright Khronos Group 2017 - Page 5© Copyright Khronos Group 2017 - Page 5© Copyright Khronos Group 2017 - Page 5
OpenVX Evolution
OpenVX 1.0
Spec released October 2014
Conformant
Implementations
OpenVX 1.1
Spec released May 2016
Conformant
Implementations
AMD OpenVX Tools
- Open source, highly
optimized for x86 CPU and
OpenCL for GPU
- “Graph Optimizer” looks at
entire processing pipeline and
removes, replaces, merges
functions to improve
performance and bandwidth
- Scripting for rapid
prototyping, without re-
compiling, at production
performance levels
http://gpuopen.com/compute-product/amd-openvx/
New Functionality
Expanded Nodes Functionality
Enhanced Graph Framework
OpenVX 1.2
Spec released 1st May 2017
New Functionality
Conditional node execution
Feature detection
Classification operators
Expanded imaging operations
Extensions
Neural Network Acceleration
Graph Save and Restore
16-bit image operation
OpenVX
Roadmap Discussions
New Functionality
Under Discussion
NNEF Import
Accelerated
Programmable user
nodes
Come to the next
presentation for more
details!

© Copyright Khronos Group 2017 - Page 6© Copyright Khronos Group 2017 - Page 6© Copyright Khronos Group 2017 - Page 6© Copyright Khronos Group 2017 - Page 6
Khronos NNEF (Neural Net Exchange Format
• Range of Neural Network tools and inferencing architectures is rapidly increasing
• NNEF encapsulates neural network formal semantics
- Structure, Data formats
- Commonly used operations (such as convolution, pooling, normalization, etc.)
• Cross-vendor Neural Net file format removes industry friction
- Simple exchange between tools and inferencing engines
- Unified format for network optimizations
NN Authoring Framework 1
NN Authoring Framework 2
NN Authoring Framework 3
Inference Engine 1
Inference Engine 2
Inference Engine 3
NN Authoring Framework 1
NN Authoring Framework 2
NN Authoring Framework 3
Inference Engine 1
Inference Engine 2
Inference Engine 3
Every Tool Needs an Exporter to
Every Accelerator

© Copyright Khronos Group 2017 - Page 7
© Copyright Khronos Group 2017 - Page 7
OpenVX 1.2 and Neural Net Extension
• Convolution Neural Network topologies can be represented as OpenVX graphs
- Layers are represented as OpenVX nodes
- Layers connected by multi-dimensional tensors objects
- Layer types include convolution, activation, pooling, fully-connected, soft-max
- CNN nodes can be mixed with traditional vision nodes
• Import/Export Extension
- Efficient handling of network Weights/Biases or complete networks
• Will be able to import NNEF files into OpenVX Neural Nets
Vision
Node
Vision
Node
Vision
Node
Downstream
Application
Processing
Native
Camera
Control CNN Nodes
An OpenVX graph mixing CNN nodes
with traditional vision nodes

© Copyright Khronos Group 2017 - Page 8© Copyright Khronos Group 2017 - Page 8© Copyright Khronos Group 2017 - Page 8© Copyright Khronos Group 2017 - Page 8
Building Graphs in C++ Rather than an API
Programmatic approaches to Graph Programming
Explicit
Kernels
Diverse Hardware
(GPUs, DSPs, FPGAs)
GPU FPGA
DSP
Dedicated
Hardware
Single Source C++
Languages
Graph
Programming
TensorRT
Code Intermediate
Representations PTX HSAIL

© Copyright Khronos Group 2017 - Page 9
© Copyright Khronos Group 2017 - Page 9
Halide – C++ Based Image Processing
• Halide is a programming language embedded in C++
- C++ code builds an in-memory representation of a Halide graph
- Compiled to an object file, to be run in the same process
• Separate descriptions of ‘algorithm’ per output pixel and execution ‘schedule’
- Can hand or auto-optimize schedule without affecting the algorithm correctness
• Compiler based on LLVM/Clang
- Compiler targets: X86, ARM, CUDA, OpenCL, and OpenGL
Halide
Program
(Image processing
Algorithm)
Human
Expert
Autoscheduler
Halide
Schedule
(mapping of
algorithm to
hardware)
Halide Code
Generator
(LLVM/Clang)
Platform-
optimized binary

© Copyright Khronos Group 2017 - Page 10© Copyright Khronos Group 2017 - Page 10© Copyright Khronos Group 2017 - Page 10© Copyright Khronos Group 2017 - Page 10
Khronos SYCL - Single Source C++
• Single-source heterogeneous programming using STANDARD C++
- Use C++ templates and lambda functions for host & device code
• Kernel Fusion in C++ is a widely used compiler technique - proven to work
- Halide, Eigen, Boost.Compute, …
- Optimization at the C++, not assembly, level
- Achieves better performance on complex software than hand-coding
• Rapid optimization of multiple libraries - more information at http://sycl.tech
- SYCLBLAS
- SYCL Eigen
- SYCL TensorFlow
- SYCL GTX
- triSYCL
- ComputeCpp
- VisionCpp
- ComputeCpp SDK

© Copyright Khronos Group 2017 - Page 11© Copyright Khronos Group 2017 - Page 11© Copyright Khronos Group 2017 - Page 11© Copyright Khronos Group 2017 - Page 11
Graph Programming - Fusion Results
• C++ Kernel fusion provides optimization benefits
- Tiled operations in local memory
- Reduced bandwidth to off-chip memory
Courtesy Codeplay:
https://www.slideshare.net/AndrewRichards28/open-standards-for-adas-andrew-richards-codeplay-at-autosens-2016-66476890

© Copyright Khronos Group 2017 - Page 12© Copyright Khronos Group 2017 - Page 12© Copyright Khronos Group 2017 - Page 12© Copyright Khronos Group 2017 - Page 12
Convergence with Standard ISO C++
• SYCL aligns the hardware acceleration of OpenCL with direction of the C++ standard
- Use of C++14 with open source C++17 Parallel STL hosted by Khronos
• Khronos working with others on bringing proposals to ISO C++ for:
- Executors – for scheduling work
- “Managed pointers” or “channels” – for sharing data
• Hoping to target C++ 20
- But timescales are tight
• The eventual aim is for standard C++ to
support advanced acceleration
offload

© Copyright Khronos Group 2017 - Page 13© Copyright Khronos Group 2017 - Page 13© Copyright Khronos Group 2017 - Page 13© Copyright Khronos Group 2017 - Page 13
Explicit Programming and Run Times
Run-time support for accelerator offload
Explicit
Kernels
Diverse Hardware
(GPUs, DSPs, FPGAs)
GPU FPGA
DSP
Dedicated
Hardware
Single Source C++
Languages
Graph
Programming
TensorRT
Code Intermediate
Representations PTX HSAIL

© Copyright Khronos Group 2017 - Page 14© Copyright Khronos Group 2017 - Page 14© Copyright Khronos Group 2017 - Page 14© Copyright Khronos Group 2017 - Page 14
OpenCL as Parallel Language/Library Backend
C++ based
Neural
network
framework
MulticoreWare
open source
project on
Bitbucket
Compiler
directives for
Fortran,
C and C++
Java language
extensions
for
parallelism
Language for
image
processing and
computational
photography
Single
Source C++
Programming
for OpenCL
Approaching 200 languages, frameworks
and projects using OpenCL as a compiler
target to access vendor-optimized,
heterogeneous compute runtimes
Low Level Explicit APIs
Vision
processing
open source
project
Open source
software library
for machine
learning

© Copyright Khronos Group 2017 - Page 15© Copyright Khronos Group 2017 - Page 15© Copyright Khronos Group 2017 - Page 15© Copyright Khronos Group 2017 - Page 15
OpenCL 2.2 - Top to Bottom C++
OpenCL 1.0
Specification
Dec08 Jun10
OpenCL 1.1
Specification
Nov11
OpenCL 1.2
Specification
OpenCL 2.0
Specification
Nov13
Device partitioning
Separate compilation and linking
Enhanced image support
Built-in kernels / custom devices
Enhanced DX and OpenGL Interop
Shared Virtual Memory
On-device dispatch
Generic Address Space
Enhanced Image Support
C11 Atomics
Pipes
Android ICD
3-component vectors
Additional image formats
Multiple hosts and devices
Buffer region operations
Enhanced event-driven execution
Additional OpenCL C built-ins
Improved OpenGL data/event interop
18 months 18 months 24 months
OpenCL 2.1
Specification
Nov1524 months
SPIR-V in Core
Subgroups into core
Subgroup query operations
clCloneKernel
Low-latency device
timer queries
OpenCL 2.2
PROVISIONAL
May167months
Single Source C++ Programming
Full support for features in C++14-based Kernel Language
API and Language Specs
Brings C++14-based Kernel Language into core specification
Portable Kernel Intermediate Language
Support for C++14-based kernel language e.g.
constructors/destructors
OpenCL C++ Kernel Language
SPIR-V 1.1 with C++ support
SYCL 2.2 for single source C++

© Copyright Khronos Group 2017 - Page 16© Copyright Khronos Group 2017 - Page 16© Copyright Khronos Group 2017 - Page 16© Copyright Khronos Group 2017 - Page 16
OpenCL Conformant Implementations
OpenCL 1.0
Specification
Dec08 Jun10
OpenCL 1.1
Specification
Nov11
OpenCL 1.2
Specification
OpenCL 2.0
Specification
Nov13
1.0|Jul13
1.0|Aug09
1.0|May09
1.0|May10
1.0 | Feb11
1.0|May09
1.0|Jan10
1.1|Aug10
1.1|Jul11
1.2|May12
1.2|Jun12
1.1|Feb11
1.1|Mar11
1.1|Jun10
1.1|Aug12
1.1|Nov12
1.1|May13
1.1|Apr12
1.2|Apr14
1.2|Sep13
1.2|Dec12
Desktop
Mobile
FPGA
2.0|Jul14
OpenCL 2.1
Specification
Nov15
1.2|May15
2.0|Dec14
1.0|Dec14
1.2|Dec14
1.2|Sep14
Vendor timelines are
first implementation of
each spec generation
1.2|May15
Embedded
1.2|Aug15
1.2|Mar16
2.0|Nov15
2.0|Apr17

© Copyright Khronos Group 2017 - Page 17© Copyright Khronos Group 2017 - Page 17© Copyright Khronos Group 2017 - Page 17© Copyright Khronos Group 2017 - Page 17
Strong Vulkan Momentum
Games Studios publicly
confirming that work is
ongoing on Vulkan Titles
All Major GPU Companies shipping Vulkan Drivers – for Desktop and Mobile Platforms
Mobile, Embedded and Console Platforms Supporting Vulkan
Embedded Linux
In first 12months :
#Vulkan Games on PC = 11
In first 18 months
#DX12 Games on PC = 19
https://en.wikipedia.org/wiki/List_of_games_with_Vulkan_support
Android TVAndroid 7.0 Nintendo Switch

© Copyright Khronos Group 2017 - Page 18© Copyright Khronos Group 2017 - Page 18© Copyright Khronos Group 2017 - Page 18© Copyright Khronos Group 2017 - Page 18
Vulkan Explicit GPU Control
GPU
High-level Driver
Abstraction
Context management
Memory allocation
Full GLSL compiler
Error detection
Layered GPU Control
Application
Single thread per context
GPU
Thin Driver
Explicit GPU Control
Application
Memory allocation
Thread management
Synchronization
Multi-threaded generation
of command buffers
Language Front-end
Compilers
Initially GLSL
Loadable debug and
validation layers
Vulkan 1.0 provides access to
OpenGL ES 3.1 / OpenGL 4.X-class GPU functionality
but with increased performance and flexibility
Loadable Layers
No error handling overhead in
production code
SPIR-V Pre-compiled Shaders:
No front-end compiler in driver
Future shading language flexibility
Simpler drivers:
Improved efficiency/performance
Reduced CPU bottlenecks
Lower latency
Increased portability
Graphics, compute and DMA queues:
Work dispatch flexibility
Command Buffers:
Command creation can be multi-threaded
Multiple CPU cores increase performance
Resource management in app code:
Less hitches and surprises
Vulkan Benefits
SPIR-V pre-compiled
shaders

© Copyright Khronos Group 2017 - Page 19© Copyright Khronos Group 2017 - Page 19© Copyright Khronos Group 2017 - Page 19© Copyright Khronos Group 2017 - Page 19
Vulkan Influence on OpenCL Roadmap
• OpenCL converging with Vulkan – expanding beyond graphics + more processor types
- Thin, powerful, explicit run-time for control and predictability
- Feature sets and dial-able precision for target market agility
- Installable tools and three layer ecosystem for flexibility
- Vulkan renderpasses are already a way to enabled tiled processing
• OpenCL also considering optional reduced precision for vision and inferencing
- Will enable significant number of DSP implementations to become conformant
Thin, explicit run-time with rigorous
memory/execution model.
Low-latency, fine-grain pre-emption
and synchronization
Dial-able types
and precision
Features that can be enabled for particular target markets
Real-time Pre-
emption and
QoS scheduling
Explicit
Asynch
DMA
Self-synchronized,
self-scheduled
graphs
Stream
Processing …
Math
Libraries
Vendor-supplied and open
source middleware
Language
Front-ends
Tool
Layers
Installable tool &
validation layers
Applications
API
Definitions

© Copyright Khronos Group 2017 - Page 20© Copyright Khronos Group 2017 - Page 20© Copyright Khronos Group 2017 - Page 20© Copyright Khronos Group 2017 - Page 20
SPIR-V Ecosystem
LLVM
Third party kernel and
shader Languages
SPIR-V
• Khronos defined and controlled
cross-API intermediate language
• Native support for graphics
and parallel constructs
• 32-bit Word Stream
• Extensible and easily parsed
• Retains data object and control
flow information for effective
code generation and translation
OpenCL C++OpenCL C
GLSL
Khronos has open sourced
these tools and translators
IHV Driver
Runtimes
Other
Intermediate
Forms
SPIR-V Validator
SPIR-V (Dis)Assembler LLVM to SPIR-V
Bi-directional
Translator
Khronos plans to open
source these tools soon
HLSL
https://github.com/KhronosGroup/SPIRV-Tools
‘glslang’ GLSL to
SPIR-V compiler
Front-end compilers leverage clang

© Copyright Khronos Group 2017 - Page 21© Copyright Khronos Group 2017 - Page 21© Copyright Khronos Group 2017 - Page 21© Copyright Khronos Group 2017 - Page 21
Possible Convergence of Graph Technologies
• API-created Graphs such as OpenVX benefit from flexibility of user-programmed nodes
- OpenVX Tiling extension lets them participate in tiled/fused optimizations
- But currently user nodes can run only on the CPU
• Perhaps use a C++ language, such such as a Halide algorithm, to program user nodes
- That can be offloaded and scheduled within the OpenVX graph
- Perhaps use SPIR-V to define node capabilities and store portable Node programs
Vision
Node
Vision
Node
Vision
Node
Downstream
Application
Processing
Native
Camera
Control
CNN Nodes
Discussion – should an OpenVX user
programmed node in a C++ domain
specific language may have its
executable stored as SPIR-V?
Programmed
User Node

© Copyright Khronos Group 2017 - Page 22© Copyright Khronos Group 2017 - Page 22© Copyright Khronos Group 2017 - Page 22© Copyright Khronos Group 2017 - Page 22
Safety Critical APIs
New Generation APIs for safety
certifiable vision, graphics and
compute
e.g. ISO 26262 and DO-178B/C
OpenGL ES 1.0 - 2003
Fixed function graphics
OpenGL ES 2.0 - 2007
Shader programmable pipeline
OpenGL SC 1.0 - 2005
Fixed function graphics subset
OpenGL SC 2.0 - April 2016
Shader programmable pipeline subset
Experience and Guidelines
Vulkan SC being discussed
Small driver size
Advanced functionality
Graphics and compute
OpenVX SC 1.1 Released Today
Restricted “deployment” implementation
executes on the target hardware by reading
the binary format and executing the pre-
compiled graphs
Khronos SCAP
‘Safety Critical Advisory Panel’
Guidelines for designing APIs that
ease system certification.
Open to Khronos member AND
industry experts. If interested to
join contact ntrevett@nvidia.com

© Copyright Khronos Group 2017 - Page 23
© Copyright Khronos Group 2017 - Page 23
Key Takeaways and What’s Next?
• Vision Tools and APIs are becoming increasingly sophisticated
- Ecosystem is layering libraries, language and run-times
• Compiler technologies becoming increasingly central – especially C++
- To enable graph-based language-based solutions with kernel fusion
- LLVM and Clang open source projects being used by many
• Safety-critical APIs becoming increasingly essential
- Many vision applications need system certification
• Still no cross-vendor camera APIs?
- Is the time yet right for this to be a target for standardization?
• Please join if your company interested in Khronos open standards
- Neil Trevett | ntrevett@nvidia.com | @neilt3d