Jim Jeffers Principal Engineer and Manager, HPC...

© 2016 Intel Corporation

Jim JeffersPrincipal Engineer and Manager, HPC Visualization

© 2016 Intel Corporation. All rights reserved. *Other names and brands may be claimed as the property of others.

Software Defined Visualization

Delivers Higher Visual Fidelity…Of Larger DataSETS

…On Existing HPC Infrastructure…Through Software

2


The Challenge with Traditional Large Scale Visualization

HPC cluster performs modeling and simulations

Dedicated VisualizationHW (GPUs) and SW

Client devices view the final images

HW Visualization

Using Dedicated hardware and specialized software


The Challenge with Traditional Large Scale Visualization

HPC cluster performs modeling and simulations

Dedicated VisualizationHW (GPUs) and SW

Client devices view the final images

Bottlenecks… I/O, Scheduling, Memory Size, Power,…


12

Bringing Your Data into FocusIntel-Supported Software Defined Visualization (SDVis)

Embree- CPU Optimized Ray Tracing Algorithms - ‘Tool kit’ for Building Ray Tracings Apps- Broadly Adopted by 3rd Party ISVs- Web Site: http://embree.github.io

OSPRay1

- Rendering Engine Based on Embree- Library and API Designed to Ease the

Creation of Visualization Software- Web Site: http://ospray.org

OpenSWR1

- Rasterization Visualization on CPUs- Good Enough to Replace HW GPU- Supports ParaView, Visit, VTK- Web Site: http://openswr.org

Standard OpenGL Image

Image Rendered by OSPRay

HW Visualization

Software Defined Visualization

1 Releasing June’16

http://embree.github.io/

http://ospray.org/

http://openswr.org/


Your Work. More Vivid.

Benefits of Software Defined Visualization

• Open-sourced technology delivering vivid visualization of complex, enormous data sets

• Innovative software library suite for visualizing results, by unlocking the parallelism already in your system

• High-Fidelity for gaining deeper insights in science and industry, faster

Magnetic Reconnection Model, Courtesy Bill Duaghton(LANL) and Berc Geveci(Kitware)

Gravational Waves : GR-Chombo AMR Data, Stephen Hawking CTC, UCambridge; Queens College, London; visualiazation, Carson Brownlee, Intel)

Data: Max-Planck Institute for Biophysical Chemistry


Unlocking Large Datasets

Software Defined Visualization can handle massive data sets, and use the most advanced imaging techniques for greater insight from compute and modelingRay tracing a 12-billion (450 GB) particle cosmology dataset at 4K resolution with ambient occlusion, using a zero-overhead balanced “p-k-d tree”. Performance is interactive (7 fps at 4 megapixels) on a quad-socket Intel® Xeon® processor E7 v3 based workstation, with no use of level-of-detail or simplification.

Complements of Ingo Wald, Aaron Knoll, Gregory P. Johnson, Will Usher, Valerio Pascucci and Michael E. Papka. “CPU Ray Tracing Large Particle Data with P-k-d Trees.” IEEE Visualization (SciVis) 2015 (conditionally accepted for publication).


SDVis ISC16 ANNOUCEMENTS

8


ParaView v5.1 / VTK 7.1 OSPRay Integration

OpenSWR MESA v12.0 Integration!

VMD v1.9.3 OSPRayIntegration (coming soon)

Chemosensory Bacteria model courtesy John Stone UIUCData Courtesy: CEI

See It Live in theSuperMicro Booth!


See It Live Right Here in the Intel Booth!


SDVis ISC16 ANNOUCEMENTS (cont.)

9

Intel® PCCs for Visualization


OSPRay 1.0 Features

Geometry Types (all float):

Triangle Meshes,Cylinders,Spheres,Streamlines,ImplicitIsosurfaces

Volume Types (double, float, short, uchar):

Structured Bricked Volumes, Structured App-shared Volumes, Data replicated scaling volumes,, Data-distributed Volumes [beta]

Linear Transfer Functions

2D Texture Mapping

Lights:

Ambient Light, Directional Light, HDRI (environment light), Point Light, Quad Light, Spot Lights

Framebuffer formats:

RGBA (4-uchar per pixel), SRGB (4-uchar per pixel in SRGB color space), RGB32F (4-float per pixel)

Cameras:

Pinhole Perspective, Orthographic, Panoramic

Renderers:

SciVis (scientific visualization)

– Ambient Occlusion, Direct Lighting (hard shadows, Volume Integration (one volume per scene)

Path Tracer (photo realistic rendering)

– Soft shadows, Environment lighting, BRDF based materials, …

Runtime environments:

Local rendering (multithreaded CPU),

COI Offload (N attached coprocessor cards in a workstation) [deprecated],

MPI Offload (N compute nodes as workers)

– Can be heterogeneous across Xeon and Xeon Phi (KNC + KNL) nodes

10


OSPRay Roadmap Features

Near to Mid Term:

OSPRay external extendability

Apps can write OSPRay extensions (new object types) within the app, making it easy to extend to application-specific problems

Adaptive volume sampling

SciVis renderer enhancements:

Soft shadows

AMR Volumes (preview in ParaView)

Multiple volumes in a single scene

Data parallel volume rendering (which will deprecate our direct volume renderer)

Mid to Long Term:

True “data parallel” rendering

Volumes AND surfaces

Data parallel API, In Situ API

Automatic device management

“Detect” hardware (KNC/KNL coprocessors, MPI nodes) and adapt at runtime

Improved load balancing across different hardware architectures and MPI topologies

11


OpenSWR Features and Roadmap

Linux OpenGL 3.x support (MESA v12.0)

ParaView, VisIt, VMD, CEI EnSight tested

SSE -> AVX2 Optimized

Tested Up to 1024 Nodes (Trinity)

Deployed at TACC (Stampede, LoneStar, Maverick, Stallion, Rattler)

Near to Mid

Windows Validation

AVX-512 Optimizations

12


Performance Leadership: Embree (Knights Landing) v. NVIDIA* OptiX* (Titan X)

Software and workloads used in performance tests may have been optimized for performance only on Intel microprocessors. Performance tests, such as SYSmark* and MobileMark*, are measured using specific computer systems, components, software, operations and functions. Any change to any of those factors may cause the results to vary. You should consult other information and performance tests to assist you in fully evaluating your contemplated purchases, including the performance of that product when combined with other products. For more information go to http://www.intel.com/performance.

Embree 2.9.0, Intel® SPMD Program Compiler (Intel® SPC) 1.9.0Intel C/C++ Compiler 16.0.1

NVIDIA* OptiX* 3.9.0, CUDA* 7.5Intel Xeon E5-2699 2x18 cores, 2.3 Ghz

Source: Intel

http://www.intel.com/performance


Performance: OSPRay (ray traced) vs GPU (OpenGL)

Software and workloads used in performance tests may have been optimized for performance only on Intel microprocessors. Performance tests, such as SYSmark* and MobileMark*, are measured using specific computer systems, components, software, operations and functions. Any change to any of those factors may cause the results to vary. You should consult other information and performance tests to assist you in fully evaluating your contemplated purchases, including the performance of that product when combined with other products. For more information go to http://www.intel.com/performance.



Summary

Software Defined Visualization:

• Intel® SSF software components designed for growing data analysis challenges

• OpenSWR, OSPRay and Embree rendering libraries using CPUs and main memory

• Scalable, Interactive Performance and Higher Fidelity particularly for Large Data

• Integrating into prominent Vis tools, ParaView*, VisIt, EnSight*, VMD, Brayns, ….

• All freely available (Open Source), developed and maintained by Intel

Remove bottlenecks

Enable greater visual insight


THANK YOU!…and a shameless plug!

Intel® Xeon Phi™ Processor High Performance Programming,Knights Landing Edition, Jim Jeffers, James Reinders, Avinash Sodani (c) 2016, publisher: Morgan Kaufmann


17


19

Image Attributions

Conference scene. Model courtesy of Anat Grynberg and Greg Ward.

Crown. Model provided by Martin Lubich under the Creative Commons Attribution License http://creativecommons.org

CSAFE Heptane Gas Dataset. Model courtesy of the Center for the Simulation of Accidental Fires and Explosions (CSAFE) at the Scientific Computing and Imaging Institute (SCI), University of Utah. http://dx.doi.org/10.1103/PhysRevLett.113.155005

Karst Ground Water Simulation Dataset. Model courtesy Texas Advanced Computing Center (TACC) and Florida International University.

Magnetic Reconnection. Model courtesy Bill Daughton (LANL) and Berk Geveci (KitWare). Also, see Guo et al. 2014, Physical Review Letters, 113, 155005.

Mazda, Art Deco, and Villa. Models copyright® Evermotion used under Evermotion Commercial License agreement.

Richtmyer-Meshkov Iso-Surface. Model courtesy Lawrence-Livermore National Labs (LLNL).

XYZ-Dragon. Model courtesy of Stanford Computer Graphics Laboratory.

http://creativecommons.org/

http://dx.doi.org/10.1103/PhysRevLett.113.155005


2

Legal Disclaimers

Software and workloads used in performance tests may have been optimized for performance only on Intel microprocessors. Performance tests, such as

SYSmark and MobileMark, are measured using specific computer systems, components, software, operations and functions. Any change to any of those factors

may cause the results to vary. You should consult other information and performance tests to assist you in fully evaluating your contemplated purchases,

including the performance of that product when combined with other products.

All information provided here is subject to change without notice. Contact your Intel representative to obtain the latest Intel product specifications and

roadmaps.

Results have been estimated or simulated using internal Intel analysis or architecture simulation or modeling, and provided to you for informational purposes.

Any differences in your system hardware, software or configuration may affect your actual performance.

Intel technologies’ features and benefits depend on system configuration and may require enabled hardware, software or service activation. Performance

varies depending on system configuration. No computer system can be absolutely secure. Check with your system manufacturer or retailer or learn more at

https://www-ssl.intel.com/content/www/us/en/high-performance-computing/path-to-aurora.html.

Tests document performance of components on a particular test, in specific systems. Differences in hardware, software, or configuration will affect actual

performance. Consult other sources of information to evaluate performance as you consider your purchase. For more complete information about

performance and benchmark results, visit http://www.intel.com/performance.

Intel, the Intel logo, Xeon, Xeon Phi, Intel Optane and 3D XPoint are trademarks or registered trademarks of Intel Corporation or its subsidiaries in the United

States or other countries.

*Other names and brands may be claimed as the property of others.

© 2016 Intel Corporation. All rights reserved.

https://www-ssl.intel.com/content/www/us/en/high-performance-computing/path-to-aurora.html



23

Optimization Notice

Intel's compilers may or may not optimize to the same degree for non-Intel microprocessors for optimizations that are not unique to Intel microprocessors. These optimizations include SSE2, SSE3, and SSE3 instruction sets and other optimizations. Intel does not guarantee the availability, functionality, or effectiveness of any optimization on microprocessors not manufactured by Intel.

Microprocessor-dependent optimizations in this product are intended for use with Intel microprocessors. Certain optimizations not specific to Intel microarchitecture are reserved for Intel microprocessors. Please refer to the applicable product User and Reference Guides for more information regarding the specific instruction sets covered by this notice.

Notice revision #20110804

Jim Jeffers Principal Engineer and Manager, HPC...

Documents

Transcript of Jim Jeffers Principal Engineer and Manager, HPC...