QUADRO FOR EDUCATION & QUADRO FOR EDUCATION & RESEARCH - PowerPoint PPT Presentation

QUADRO FOR EDUCATION & QUADRO FOR EDUCATION & RESEARCH RESEARCH 6-2019

EDUCATION ENVIRONMENT CHALLENGES Applications. Frameworks, Libraries Research problems, Educational and Demand for skilled data IT budgets not growing research software scientists & AI expertise student projects more to match technology complex than ever, toolchains & requires students to hit infrastructure data sets & compute workloads increasing the ground running demands requirements growing in size and complexity after graduation exponentially 2

QUADRO HER VALUE PROPOSITION Only NVIDIA Quadro solutions provide: • Compute power, GPU memory capacity, and scalability required for today’s demanding education & research projects • Researcher and student access to professional level hardware & software – same used by professionals around the world RTX 6000 24GB • Enterprise level hardware & software support – maximize uptime and minimize IT support requirements • Server ready GPU solutions 3

QUADRO FOR DEEP LEARNING TRAINING INFERENCE Quadro value for DL training: Quadro value for DL inferencing: • Large memory, scalability for large • Fast inferencing performance for the training data sets, model optimization largest networks & datasets & validation Support for encode/decode of • • Scalability with NVLink to speed up multiple simultaneous video streams training tasks, provide development • Consumer cards limited to 2 simultaneous platform for server deployments streams Quadro solutions are validated, tested, and used daily in demanding professional DL environments 4

DEEP LEARNING TRAINING PERFORMANCE Quadro Provides Memory & Performance Required for DL Training ResNet-50 Training Mixed Precision OpenSeq2Seq (GNMT) Training Mixed Relative Performance Precision Relative Performance 1.8 1.69 1.6 2.50 1.36 1.4 2.03 2.00 1.2 1.0 1.50 0.8 1.00 0.6 0.4 0.50 0.2 * 0.00 0.0 0.00 RTX 2080 RTX 2080Ti RTX 6000 RTX 2080 RTX 2080Ti RTX 6000 *Did not run due to insufficient memory Quadro enables larger batch sizes to accelerate DL Large training datasets such as natural language training processing, machine translations, etc., don’t fit into consumer graphics memory Test run on Xeon Gold 6140@2.30GHz, NVIDIA driver 415.13, Tensor Flow 5 ResNet-50 and OpenSeq2Seq training benchmarks, mixed precision.

DEEP LEARNING INFERENCING PERFORMANCE Deep Networks with Many Layers can Take Advantage of Quadro Memory VGG16 Inference Mixed Precision VGG19 Inference INT8 Precision Relative Performance Relative Performance 1.40 1.33 1.40 1.31 1.29 1.25 1.20 1.20 1.00 1.00 0.80 0.80 0.60 0.60 0.40 0.40 0.20 0.20 0.00 0.00 RTX 2080 RTX 2080Ti RTX 6000 RTX 2080 RTX 2080Ti RTX 6000 Test run on Xeon Gold 6140@2.30GHz, NVIDIA driver 415.13, Tensor Flow 6 VGG16 and VGG19 inference benchmarks, mixed precision.

QUADRO – FASTER TIME TO SOLUTION* 1 Hourt Training Time (in minutes) RTX 6000 40.2 RTX 2080 Ti 0 10 20 30 40 50 60 1 Day Training Time (in Hours) RTX 6000 16.1 RTX 2080 Ti 0 4 8 12 16 20 24 1 Week Training Time (in Days) RTX 6000 4.7 RTX 2080 Ti 0 1 2 3 4 5 6 7 7 *based on RTX 6000 ResNet-50 performance 33% faster than 2080Ti

QUADRO REQUIRED FOR THE LARGEST WORKLOADS Data Science Data Sets Require GPU Memory Only Available with Quadro Data Science Workload Example: Sample data set using Home Mortgage data in the US for 2016. A single GeForce RTX 2080Ti can only load 3 months worth of data. A single Quadro RTX 6000 GPU can load 6 months of data at a time. Two RTX 6000’s can load the entire years worth of data in the combined GPU memory (dual 2080Ti’s can only load 6 months of data).* Max Data Set Size End To End Training Time Quadro value for data science (in months) (3 months of data, relative performance) 7 • Large Quadro memory lets 1.60 1.40 data scientists process 1.40 6 more data to improve 1.20 5 model training and 1.00 4 accuracy. 0.80 3 0.60 • Quadro performance 2 0.40 completes tasks faster 1 0.20 0 0.00 2080Ti RTX 6000 2080Ti RTX 6000 8 *A single RTX 8000 can load the entire year’s worth of data Test system: dual Gold 6140@2.30GHz 3.7GHz Turbo (Skylake), ETL with Dask + Pandas

QUADRO FOR HPC Quadro value for HPC: • Larger GPU memory for the largest data sets and compute tasks • Memory and compute power enables multi application workflows from compute to visualization Quadro is NGC Ready, tested and • validated for NVIDIA NGC HPC software 9

QUADRO FOR VIRTUALIZATION Quadro vDWS – Virtual Quadro Anywhere Quadro value for virtualization: • Get the features, performance of Quadro anywhere you need it. • Partition a single RTX 6000 into multiple high performance virtual workstations • Quadro vDWS only available on Tesla, Quadro RTX 6000/8000 GPUs Learn more about Quadro vDWS solutions: https://www.nvidia.com/quadro-vdws/ 10

NVIDIA NGC SUPPORT SERVICES Minimize Downtime And Maximize System Utilization Support Coverage Availability • NGC DL & ML containers • Exclusively for NGC-Ready • NVIDIA drivers workstations • NV-docker • Service agreement between • CUDA NVIDIA & customer • Purchase from OEM/System vendor 24x7 portal, phone and email access to Support by NVIDIA’s create support cases subject matter experts Live support during local, regional business hours for technical assistance 11

GeForce RTX 2080Ti Quadro RTX 6000 PC-based HPC, AI/DL/ML workloads, small to NVIDIA Desktop System Use Workstation-based HPC, AI/DL/ML workloads, largest data sets with up to 48GB medium size data sets, up to 22GB memory Case memory w/NVLink, pre-installed data science software stack. w/NVLink 11GB 24GB GPU Memory ECC Memory No Yes – error free compute 2-way NVLink, 2 & 3 slot bridges – fits into a wide variety of workstation 2-way NVLink, 3 & 4 slot bridges NVLink support chassis Cooling Solution Dual Axial Cooling Blower Active Fan – ideal cooling solution for multi-GPU configurations Pre-installed Data Science No Yes – up and running data science workloads within minutes Software Stack from OEMs Nvidia support N/A Enterprise level pre/post sales support OEM support N/A Enterprise level support No Yes – validated and tested for running NVIDIA NGC software NGC Ready NVIDIA NGC Support Services N/A Yes – support provided directly from NVIDIA Option Work with largest data sets to accelerate HPC, AI, ML workflows. NVIDIA and OEMs provide enterprise level testing, validation, and support. Additional NVIDIA NGC Quadro Advantage ฀ Support Services availability. GPU Cooling & NVLink solutions support widest range 12 of system configurations.

QUADRO FOR HIGHER EDUCATION & RESEARCH Quadro Brings Real Value • Enterprise grade hardware, enterprise level software & hardware testing, validation, & support – ready for workstation & server deployments • Performance & scalability – large GPU memory RTX 6000 24GB for large datasets & workloads today and room for growth in the future • Maximum flexibility with the option to provide multiple virtual Quadro workstations (Quadro vDWS) from a single Quadro GPU 13