Вы находитесь на странице: 1из 21
Stan Posey, CAE Industry Development NVIDIA, Santa Clara, CA, USA
Stan Posey, CAE Industry Development NVIDIA, Santa Clara, CA, USA
Stan Posey, CAE Industry Development NVIDIA, Santa Clara, CA, USA

Stan Posey, CAE Industry Development NVIDIA, Santa Clara, CA, USA

Stan Posey, CAE Industry Development NVIDIA, Santa Clara, CA, USA
Stan Posey, CAE Industry Development NVIDIA, Santa Clara, CA, USA

VIDIA and HPC Evolution of GPUs

Public, based in Santa Clara, CA | ~$4B revenue | ~5,500 employees based in Santa Clara, CA | ~$4B revenue | ~5,500 employees

Founded in 1999 with primary business in semiconductor industry in 1999 with primary business in semiconductor industry

Products for graphics in workstations, notebooks, mobile devices, etc. for graphics in workstations, notebooks, mobile devices, etc.

Began R&D of GPUs for HPC in 2004, released first Tesla and CUDA in 2007 R&D of GPUs for HPC in 2004, released first Tesla and CUDA in 2007

Development of GPUs as a co ‐ processing accelerator for x86 CPUs of GPUs as a co processing accelerator for x86 CPUs

PC Evolution of GPUs

2004: Began strategic investments in GPU as HPC co-processor Began strategic investments in GPU as HPC co-processor

Began strategic investments in GPU as HPC co-processor 2006: G80 first GPU with built-in compute features,

2006: G80 first GPU with built-in compute features, 128 cores; CUDA SDK Beta G80 first GPU with built-in compute features, 128 cores; CUDA SDK Beta

GPU with built-in compute features, 128 cores; CUDA SDK Beta 2007: Tesla 8-series based on G80,

2007: Tesla 8-series based on G80, 128 cores – CUDA 1.0, 1.1 Tesla 8-series based on G80, 128 cores – CUDA 1.0, 1.1

2008: Tesla 10-series based on GT 200, 240 cores – CUDA 2.0, 2.3 Tesla 10-series based on GT 200, 240 cores – CUDA 2.0, 2.3

2009: Tesla 20-series, code named “Fermi” up to 512 cores – CUDA SDK 3.0, 3.2 Tesla 20-series, code named “Fermi” up to 512 cores – CUDA SDK 3.0, 3.2

ANSYS 2011 Regional Conferences | 25 Aug 2011 | Palm Beach Gardens, FL

3 Years With 3 Generations

2

esearch Example: CFD for Gas Turbine Engines

esearch Example: CFD for Gas Turbine Engines University of Cambridge DARWIN Cluster CUDA Center of Excellence

University of Cambridge DARWIN Cluster

CUDA Center of Excellence Since 2008

DARWIN Cluster CUDA Center of Excellence Since 2008 GPU sub ‐ cluster: Dell T5500 servers, 32

GPU subcluster:

Dell T5500 servers, 32 dual ‐ socket CPUs T5500 servers, 32 dual socket CPUs

Tesla S1070 GPUs, 4 GPUs per socket for total 128 GPUs S1070 GPUs, 4 GPUs per socket for total 128 GPUs

urces: www.many-core.group.cam.ac.uk/ukgpucc2/talks/Brandvik.pdf

|

Siemens UK 3-Stage Compressor Rig

| Siemens UK 3-Stage Compressor Rig • 160 M Grid Points • 5 Revolutions • 23K
| Siemens UK 3-Stage Compressor Rig • 160 M Grid Points • 5 Revolutions • 23K

160 M Grid Points

5 Revolutions

23K Time Steps

24 hours for each

revolution on 32 GPUs

Turbostream Simulation Speed‐up 19x ~19x
Turbostream Simulation Speed‐up 19x
~19x

www.turbostream-cfd.com/

|

www.hpc.cam.ac.uk/services/darwin.html

ANSYS 2011 Regional Conferences | 25 Aug 2011 | Palm Beach Gardens, FL

3

VIDIA and ANSYS Collaboration Focus

NVIDIA GPU Status

Structural

Mechanics

Collaboration Focus NVIDIA GPU Status Structural Mechanics Fluid Dynamics Electro ‐ magnetics Available Today
Collaboration Focus NVIDIA GPU Status Structural Mechanics Fluid Dynamics Electro ‐ magnetics Available Today
Collaboration Focus NVIDIA GPU Status Structural Mechanics Fluid Dynamics Electro ‐ magnetics Available Today

Fluid

Dynamics

Electro magnetics

Structural Mechanics Fluid Dynamics Electro ‐ magnetics Available Today ANSYS Mechanical 13 SMP, Single GPU
Structural Mechanics Fluid Dynamics Electro ‐ magnetics Available Today ANSYS Mechanical 13 SMP, Single GPU
Available Today ANSYS Mechanical 13 SMP, Single GPU ANSYS Nexxim (Signal Integrity) Updates for 2011
Available
Today
ANSYS Mechanical 13
SMP, Single GPU
ANSYS Nexxim
(Signal Integrity)
Updates
for 2011
ANSYS Mechanical 14
DMP, Improved PCG
ANSYS CFD 14
Radiation HT (beta)

Product

Evaluation

Research

Evaluation

ANSYS Mechanical 15 Multi ‐GPU, Multi ‐node
ANSYS Mechanical 15
Multi ‐GPU, Multi ‐node
ANSYS CFD 15 Solver, other models
ANSYS CFD 15
Solver, other models

ANSYS HFSS

ANSYS Maxwell

NVIDIA Provides Business and Engineering Investments in ANSYS Technology Developments

ANSYS 2011 Regional Conferences | 25 Aug 2011 | Palm Beach Gardens, FL

4

ow ANSYS Software Works With GPUs

ANSYS computes the heavy workloads of matrix solvers on the GPU and other routines on the CPU the GPU and other routines on the CPU

ANSYS Mechanical GPU acceleration is user-transparentof matrix solvers on the GPU and other routines on the CPU Jobs launch and complete

CPU ANSYS Mechanical GPU acceleration is user-transparent Jobs launch and complete without additional user steps 1

Jobs launch and complete without additional user steps

1 1. ANSYS job launched on CPU 2. Solver operations sent to GPU 3. GPU
1
1. ANSYS job launched on CPU
2. Solver operations sent to GPU
3. GPU sends results back to CPU
2
4. ANSYS job completes on CPU
4
3

ANSYS 2011 Regional Conferences | 25 Aug 2011 | Palm Beach Gardens, FL

5

mportant Considerations for ANSYS and GPUs

m portant Considerations for ANSYS and GPUs Core ANSYS focus is on direct and iterative linear

Core ANSYS focus is on direct and iterative linear solvers

Others (models, mat. assembly) move to GPUs in progressive stages mat. assembly) move to GPUs in progressive stages

Most ANSYS software employs a domain parallel method

GPU computing fits this method, preserves DANSYS investments computing fits this method, preserves DANSYS investments

ANSYS 13 focus was SMP solvers; ANSYS 14 focus is DANSYS solvers 13 focus was SMP solvers; ANSYS 14 focus is DANSYS solvers

ANSYS software is parallel and scales well for multicore CPUs

Direct solvers use a scheme of computations on both GPU and CPU solvers use a scheme of computations on both GPU and CPU

Iterative solvers have computations on GPU, matrix assembly on CPU solvers have computations on GPU, matrix assembly on CPU

Investigations include GPU performance against multi ‐ core CPU only GPU performance against multicore CPU only

ANSYS 2011 Regional Conferences | 25 Aug 2011 | Palm Beach Gardens, FL

6

ANSYS Presentation at NVIDIA GTC 2010

Sep 20 — 23, 2010 San Jose Convention Center, San Jose, California, USA

2010 San Jose Convention Center, San Jose, California, USA Accelerating System Level Signal Integrity Simulation

Accelerating System Level Signal Integrity Simulation with GPU

Dr. Ekanathan Palamadai, ANSYS

Lower is better Speedup combines GPU and other SW changes
Lower
is
better
Speedup combines
GPU and other SW
changes

Nexxim 13.0 Convolution Results for Tesla C2050:

Intel Nehalem 8 core CPU, OpenMP: 108 H

NVIDIA Tesla C2050

GPU, OpenMP:

4 H

Single Precision ~27x

Double Precision ~13x

ANSYS 2011 Regional Conferences | 25 Aug 2011 | Palm Beach Gardens, FL

7

NSYS CFD 14.0 to Offer (Beta) GPU Capability

NOTE: Growing CPU time of view-factor computations inhibit proper inclusion of radiation HT effects NOTE:
NOTE: Growing CPU
time of view-factor
computations inhibit
proper inclusion of
radiation HT effects
NOTE: GPU time
remains low even
as view-factor
computations
grow very large
low even as view-factor computations grow very large ANSYS CFD preliminary results of radiation heat

ANSYS CFD preliminary results of radiation heat transfer view factor computation on GPUs vs. CPUs

Radiation HT Applications:

Underhood cooling

Cabin comfort HVAC

Furnace simulations

Solar loads on buildings

Combustor in turbine

Electronics passive cooling

Combustor in turbine ‐ Electronics passive cooling Other ANSYS CFD Evaluations: ‐ Models (e.g. disperse
Combustor in turbine ‐ Electronics passive cooling Other ANSYS CFD Evaluations: ‐ Models (e.g. disperse
Combustor in turbine ‐ Electronics passive cooling Other ANSYS CFD Evaluations: ‐ Models (e.g. disperse
Combustor in turbine ‐ Electronics passive cooling Other ANSYS CFD Evaluations: ‐ Models (e.g. disperse

Other ANSYS CFD Evaluations:

Models (e.g. disperse phase)

Implicit equation solvers

ANSYS 2011 Regional Conferences | 25 Aug 2011 | Palm Beach Gardens, FL

8

NSYS Announcement of NVIDIA CUDA Support

N SYS Announcement of NVIDIA CUDA Support "This initial development for GPU computing demonstrates our focus
N SYS Announcement of NVIDIA CUDA Support "This initial development for GPU computing demonstrates our focus
N SYS Announcement of NVIDIA CUDA Support "This initial development for GPU computing demonstrates our focus

"This initial development for GPU computing demonstrates our focus on evolving ANSYS software to take advantage of important technology trends in high performance computing." said Dipankar Choudhury, vice president of corporate product strategy and planning at ANSYS. "We work to achieve optimized software performance, across the full spectrum of HPC technologies, so that our customers get maximum value from their investment in HPC. Here, our technical collaboration with NVIDIA has resulted in a significant benefit for our mutual customers."

ANSYS 2011 Regional Conferences | 25 Aug 2011 | Palm Beach Gardens, FL

9

etails of ANSYS Mechanical for NVIDIA GPUs

e tails of ANSYS Mechanical for NVIDIA GPUs ANSYS Mechanical 13: Collaboration on SMP direct sparse

ANSYS Mechanical 13: Collaboration on SMP direct sparse and PCG/JCG iterative solvers CUDA 3.2 support in 13.0 SP2

Initial release for both Linux and Windows 64 bit, and single GPU per job – multiGPU under evaluation for future release:

Model limits for direct depend on largest front sizes: GPUs good for ~1M DOF to ~8M limits for direct depend on largest front sizes: GPUs good for ~1M DOF to ~8M DOF for 6GB Tesla C2075 or Quadro 6000

Model limits for iterative depend on GPU memory: GPUs good for ~1M DOF to ~5M DOF limits for iterative depend on GPU memory: GPUs good for ~1M DOF to ~5M DOF for 6GB Tesla C2075 or Quadro 6000

ANSYS Mechanical 14: Collaboration on DMP solvers – Q4 2011

ANSYS 2011 Regional Conferences | 25 Aug 2011 | Palm Beach Gardens, FL

10

NSYS Mechanical 13 Results of GPU Acceleration

NOTE: Results of ANSYS Mechanical for Tesla C2050 and Intel Xeon 5560

GPU Solver Kernel Speedups GPU Overall Simulation Speedups
GPU Solver
Kernel Speedups
GPU Overall
Simulation Speedups

ANSYS 2011 Regional Conferences | 25 Aug 2011 | Palm Beach Gardens, FL

Regional Conferences | 25 Aug 2011 | Palm Beach Gardens, FL From NAFEMS World Congress May

From NAFEMS World Congress May 2011 Boston, MA, USA

“Accelerate FEA Simulations with a GPU”

by Jeff Beisheim, ANSYS

System Configuration:

Xeon 5560, 2.8 GHz 2 sockets, 8 cores

32 GB memory

Win XP SP2 64bit

Tesla C2050 GPU

11

ANSYS Mechanical Times in Seconds

NSYS Mechanical 14 Preview on GPU Workstation

NOTE:

Results Based on ANSYS Mechanical 14.0 Preview 3 DMP Solver Aug 2011

3000

2000

1000

0

Lower is better
Lower
is
better

Xeon 5670 2.93 GHz Westmere (Dual Socket)3 DMP Solver Aug 2011 3000 2000 1000 0 Lower is better Xeon 5670 2.93 GHz

Xeon 5670 2.93 GHz Westmere + Tesla C2075Lower is better Xeon 5670 2.93 GHz Westmere (Dual Socket) Results from HP Z800 Workstation, 2

Results from HP Z800 Workstation, 2 x Xeon X5670 2.93GHz 48GB memory, CentOS 5.4 x64;
Results from HP Z800 Workstation, 2 x Xeon X5670 2.93GHz
48GB memory, CentOS 5.4 x64; Tesla C2075, CUDA 4.0.17
NOTE: Add a Tesla C2075 to use with 6 cores: now 46% faster than 12,
NOTE: Add a Tesla C2075
to use with 6 cores: now
46% faster than 12, with
6 available for other tasks
now 46% faster than 12, with 6 available for other tasks 2.1x 273 564 1848 4.2x

2.1x

273
273
564
564
1848
1848
than 12, with 6 available for other tasks 2.1x 273 564 1848 4.2x 444 1192 3.5x

4.2x

444
444
1192
1192
3.5x
3.5x
342
342

846

2.7x
2.7x
314
314
1.9x
1.9x
270
270
516
516
399
399
1 Core 2 Core 4 Core 6 Core 8 Core 12 Core 1 Socket 2
1 Core
2 Core
4 Core
6 Core
8 Core
12 Core
1 Socket
2 Socket

ANSYS 2011 Regional Conferences | 25 Aug 2011 | Palm Beach Gardens, FL

Regional Conferences | 25 Aug 2011 | Palm Beach Gardens, FL AVAILABLE Q4 2011 V13sp‐ 5

AVAILABLE Q4 2011

V13sp‐ 5 Model ‐ Turbine geometry ‐ 2,100 K DOF ‐ SOLID187 FEs ‐ Static,
V13sp‐ 5 Model
‐ Turbine geometry
‐ 2,100 K DOF
‐ SOLID187 FEs
‐ Static, nonlinear
‐ One load step
‐ Direct sparse

12

ow ANSYS is Licensed for NVIDIA GPUs

ANSYS Base License:

ANSYS HPC Pack:

Licensed for NVIDIA GPUs ANSYS Base License: ANSYS HPC Pack: Unlocks up to 2 CPU Cores

Unlocks up to 2 CPU Cores

Unlocks up to 8 CPU Cores Unlocks GPU Acceleration

* Academic customers: GPU acceleration is bundled

ANSYS 2011 Regional Conferences | 25 Aug 2011 | Palm Beach Gardens, FL

13

Factors Gain Over Base License Results

NSYS Cost Performance Gain > 4X vs. Base License

NOTE:

N SYS Cost Performance Gain > 4X vs. Base License NOTE: Based on ANSYS Mechanical 14.0

Based on ANSYS Mechanical 14.0 Preview 3 DMP Solver Aug 2011 and Model V13sp-5

5

4

3

2

1

0

Results from HP Z800 Workstation, 2 x Xeon X5670 2.93GHz 48GB memory, CentOS 5.4 x64;
Results from HP Z800 Workstation, 2 x Xeon X5670 2.93GHz
48GB memory, CentOS 5.4 x64; Tesla C2075, CUDA 4.0.17
CPU Speed‐ up GPU Speed‐ up Solution Cost
CPU Speed‐ up
GPU Speed‐ up
Solution Cost
1.0 1.0
1.0
1.0
CPU Speed‐ up GPU Speed‐ up Solution Cost 1.0 1.0 Base License 2 Core ANSYS 2011

Base License

2 Core

ANSYS 2011 Regional Conferences | 25 Aug 2011 | Palm Beach Gardens, FL

Solution Cost Basis ‐ ANSYS base license ‐ ANSYS HPC Pack ‐ Workstation ‐ Tesla
Solution Cost Basis
‐ ANSYS base license
‐ ANSYS HPC Pack
‐ Workstation
‐ Tesla C2075
Performance Basis
V13sp‐5 Model:
‐ 2,100 K DOF
‐ SOLID187 FEs
‐ Static nonlinear
‐ One load step
‐ Direct sparse
14

Factors Gain Over Base License Results

NSYS Cost Performance Gain > 4X vs. Base License

NOTE:

N SYS Cost Performance Gain > 4X vs. Base License NOTE: Based on ANSYS Mechanical 14.0

Based on ANSYS Mechanical 14.0 Preview 3 DMP Solver Aug 2011 and Model V13sp-5

5

4

3

2

1

0

Results from HP Z800 Workstation, 2 x Xeon X5670 2.93GHz 48GB memory, CentOS 5.4 x64;
Results from HP Z800 Workstation, 2 x Xeon X5670 2.93GHz
48GB memory, CentOS 5.4 x64; Tesla C2075, CUDA 4.0.17
CPU Speed‐ up GPU Speed‐ up Solution Cost
CPU Speed‐ up
GPU Speed‐ up
Solution Cost
2.3 1.23
2.3
1.23
2.1 1.23
2.1
1.23
1.0 1.0
1.0
1.0
GPU Speed‐ up Solution Cost 2.3 1.23 2.1 1.23 1.0 1.0 Base License 2 Core ANSYS
GPU Speed‐ up Solution Cost 2.3 1.23 2.1 1.23 1.0 1.0 Base License 2 Core ANSYS
GPU Speed‐ up Solution Cost 2.3 1.23 2.1 1.23 1.0 1.0 Base License 2 Core ANSYS

Base License

2 Core

ANSYS HPC

ANSYS HPC

Pack 6 Cores Pack 8 Cores

ANSYS 2011 Regional Conferences | 25 Aug 2011 | Palm Beach Gardens, FL

Solution Cost Basis ‐ ANSYS base license ‐ ANSYS HPC Pack ‐ Workstation ‐ Tesla
Solution Cost Basis
‐ ANSYS base license
‐ ANSYS HPC Pack
‐ Workstation
‐ Tesla C2075
Performance Basis
V13sp‐5 Model:
‐ 2,100 K DOF
‐ SOLID187 FEs
‐ Static nonlinear
‐ One load step
‐ Direct sparse
15

Factors Gain Over Base License Results

NSYS Cost Performance Gain > 4X vs. Base License

NOTE:

N SYS Cost Performance Gain > 4X vs. Base License NOTE: Based on ANSYS Mechanical 14.0

Based on ANSYS Mechanical 14.0 Preview 3 DMP Solver Aug 2011 and Model V13sp-5

5

4

3

2

1

0

Results from HP Z800 Workstation, 2 x Xeon X5670 2.93GHz 48GB memory, CentOS 5.4 x64;
Results from HP Z800 Workstation, 2 x Xeon X5670 2.93GHz
48GB memory, CentOS 5.4 x64; Tesla C2075, CUDA 4.0.17
CPU Speed‐ up GPU Speed‐ up Solution Cost
CPU Speed‐ up
GPU Speed‐ up
Solution Cost
2.1 1.23
2.1
1.23
4.4 1.28
4.4
1.28
3.8 1.28
3.8
1.28
2.3 1.23
2.3
1.23
1.0 1.0
1.0
1.0
Cost 2.1 1.23 4.4 1.28 3.8 1.28 2.3 1.23 1.0 1.0 Base License 2 Core ANSYS
Cost 2.1 1.23 4.4 1.28 3.8 1.28 2.3 1.23 1.0 1.0 Base License 2 Core ANSYS
Cost 2.1 1.23 4.4 1.28 3.8 1.28 2.3 1.23 1.0 1.0 Base License 2 Core ANSYS

Base License

2 Core

ANSYS HPC

Pack 6 Cores Pack 8 Cores 4 Cores + GPU

ANSYS HPC

ANSYS HPC Pack ANSYS HPC Pack

6 Cores + GPU

ANSYS 2011 Regional Conferences | 25 Aug 2011 | Palm Beach Gardens, FL

Solution Cost Basis ‐ ANSYS base license ‐ ANSYS HPC Pack ‐ Workstation ‐ Tesla
Solution Cost Basis
‐ ANSYS base license
‐ ANSYS HPC Pack
‐ Workstation
‐ Tesla C2075
Performance Basis
V13sp‐5 Model:
‐ 2,100 K DOF
‐ SOLID187 FEs
‐ Static nonlinear
‐ One load step
‐ Direct sparse
16

VIDIA Use of ANSYS Software for Product Design

ANSYS Icepak – active and passive cooling of IC packages

Icepak – active and passive cooling of IC packages ANSYS Mechanical – large deflection bending of
Icepak – active and passive cooling of IC packages ANSYS Mechanical – large deflection bending of

ANSYS Mechanical – large deflection bending of PCBs

ANSYS Mechanical – comfort and fit of 3D emitter glasses

Mechanical – comfort and fit of 3D emitter glasses ANSYS Mechanical – shock & vib of
Mechanical – comfort and fit of 3D emitter glasses ANSYS Mechanical – shock & vib of

ANSYS Mechanical – shock & vib of solder ball assemblies

ANSYS 2011 Regional Conferences | 25 Aug 2011 | Palm Beach Gardens, FL

17

VIDIA HPC Case Study: Performance Gain of 77x

ANSYS Mechanical simulations by NVIDIA for design of 3D emitter glasses

Simulation for prediction of comfort, fit, and handling for prediction of comfort, fit, and handling

Study optimized on CPU platform before applying GPU optimized on CPU platform before applying GPU

Once neglected model parameterization now practical neglected model parameterization now practical

GPU Once neglected model parameterization now practical ANSYS 2011 Regional Conferences | 25 Aug 2011 |
GPU Once neglected model parameterization now practical ANSYS 2011 Regional Conferences | 25 Aug 2011 |

ANSYS 2011 Regional Conferences | 25 Aug 2011 | Palm Beach Gardens, FL

18

ecommended System Configurations

e commended System Configurations Workstations Workstations with Tesla GPUs Existing System • Tesla C2050 (3 GB)

Workstations

Workstations with Tesla GPUs
Workstations with
Tesla GPUs

Existing System

• Tesla C2050 (3 GB)

• Tesla C2075 (6 GB)

New System Purchase

• Total 6-8 CPU cores

Total 48 GBs of CPU memory

• Disk with minimum 500 GB

• Tesla C2075

+ Quadro 2000 for pre/post -- OR --

• Quadro 6000 (6GB)

+ Quadro 2000 for pre/post -- OR -- • Quadro 6000 (6GB) Servers Existing System Servers

Servers Existing System

Servers with Tesla GPUs

6000 (6GB) Servers Existing System Servers with Tesla GPUs • Tesla S2050 (12 GB or 3
6000 (6GB) Servers Existing System Servers with Tesla GPUs • Tesla S2050 (12 GB or 3
6000 (6GB) Servers Existing System Servers with Tesla GPUs • Tesla S2050 (12 GB or 3

• Tesla S2050 (12 GB or 3 GB/GPU)

New System Purchase

• Total 4 CPUs, 6-8 CPU cores each

• Total 4 x16 PCIe (one for each GPU)

• Total 96 to128 GBs of CPU memory

• Disk with minimum 2000 GB (scratch)

• Tesla M2070 or Tesla M2090

ANSYS 2011 Regional Conferences | 25 Aug 2011 | Palm Beach Gardens, FL

19

ummary and Next Steps

ummary and Next Steps ANSYS Software supports NVIDIA GPUs for Computation ANSYS 13.0 since Nov 2010;

ANSYS Software supports NVIDIA GPUs for Computation

ANSYS Software supports NVIDIA GPUs for Computation ANSYS 13.0 since Nov 2010; New features coming in

ANSYS 13.0 since Nov 2010; New features coming in ANSYS 14.0

Joint Collaboration on ANSYS 13.0 is only the beginning

Collaboration ongoing in all disciplines of CSM, CFD and CEM ongoing in all disciplines of CSM, CFD and CEM

Learn more about ANSYS and NVIDIA GPU solution

More at: www.nvidia.com/object/tesla ‐ ansys ‐ accelerations.html

More at: www.nvidia.com/object/teslaansys accelerations.html

Want to try ANSYS on NVIDIA GPUs? Contact cae@nvidia.com

Want to try ANSYS on NVIDIA GPUs? Contact cae@nvidia.com

ANSYS 2011 Regional Conferences | 25 Aug 2011 | Palm Beach Gardens, FL

20

Thank You, Questions ? Stan Posey | CAE Industry Development | sposey@nvidia.com NVIDIA, Santa Clara,

Thank You, Questions ?

Stan Posey | CAE Industry Development | sposey@nvidia.com NVIDIA, Santa Clara, CA, USA

ANSYS 2011 Regional Conferences | 25 Aug 2011 | Palm Beach Gardens, FL

21