2014-05-24

Learning Machines - GPU vs CPU

Data center

BIG DATA ANALYTICS, DATA SCIENCE & MACHINE LEARNING

A growing number of customers are using GPUs for big data analytics to make better, real-time business decisions. This page highlights customer use cases and their techniques in big data analytics, such as machine learning, search, and sorting.

 
10x SPEED-UP ON IMAGE DETECTION USING NEURAL NETWORKS

10x SPEED-UP ON IMAGE DETECTION USING NEURAL NETWORKS 
Dr. Dan Ciresan, Swiss Al Lab IDSIA, Switzerland

WORLD'S LARGEST ARTIFICIAL NEURAL NETWORKS WITH GPUS

WORLD'S LARGEST ARTIFICIAL NEURAL NETWORKS WITH GPUS 
Adam Cotes et al Sanford Al Lab, U.S.A. – LEARN MORE

 

For information on key ISVs and applications, please visit the GPU Applications page.

 
News Highlights
> Researchers Deploy GPUs To Build World's Largest Artificial Neural Network (NV Press Release)
> Baidu using GPUs to train neural nets (Wired Article)
> Google using GPUs for convolution neural nets ( Wired Article)
> Fast Database Emerges from MIT Class, GPUs and Student's Invention (Data-informed article)
> NVIDIA GPUs, Big-Data Analytics And Search, Shazam, Salesforce, Cortexica (NV Press Release)
> IDSIA, How bio-inspired deep learning keeps winning competitions ( Kurzweil blog)
 
 
Technical Reports on using CUDA for Big Data Problems

Machine Learning

> Deep learning with COTS HPC systems, Coates (Stanford) (PDF)

> Mitosis Detection in Breast Cancer Histology Images using DNN, Ciresan (IDSIA)(PDF

Other competition winning papers and benchmarks (IDSIA website ) 

> Fast SVM Training and Classification on GPUs, Bryan (NVIDIA) (PDF

Data Mining & Analytics

>  GPU-accelerated Keyword Matching and Expression Evaluation for Real-time Text Search, Wood (Salesforce.com).  GTC13 (videoslides

>  Building Accelerated DSLs and GPU Compilers with libNVVM (note this includes an example with R), Lin (NVIDIA).  GTC13 (videoslides

>  GPU-Accelerated Large Scale Analytics, Wu (HP Laboratories) (PDF

> Exploiting Graphic Card Processor Technology to Accelerate Data Mining in SAP NetWeaver BIA, Weyerhaeuser (SAP) (PDF)

> Parallel Search on Video Cards, Kaldeway ( Oracle Corp) (PDF)

Search & Sorting

> Efficient Parallel Lists Intersection and Index Compression Algorithms using GPU, Ao (Baidu-Naikai Joint Lab) (PDF

> Scalable GPU Graph Traversal, Merrill (NVIDIA) (webpage)

> Efficient Sorting Algorithms for Manycore GPUs, Satish(Berkeley) (PDF)

> A Fast, Flexible Sorting Algorithm with CUDA, Chen. (PDF)

> Sorting using BItonic network wIth CUDA, Baraglia (PDF)

Databases

> Let Your GPU do the Heavy Lifting in Your Data Warehouse, Kaldewey (IBM), Mueller (IBM) (videoslides

> A GPU Database Architecture, Bakkum (Groupon) (videoslides)

> Faster Centrality Computations on GPUs, Catalyurek (Ohio State) (videoslides

>  Efficient Merge, Search and Set Operations on GPUs, Baxter, Merrill (videoslides

> GPU join processing revisited, Kaldewey(IBM, DaMon'12)  (PDF)

> GPU Accelerated Text Mining, Zhang. (PDF

Map-Reduce / Hadoop

> Multi-GPU MapReduce on GPU Clusters, Stuart (UC. Davis) (PDF

> Pipelined Multi-GPU MapReduce for Big-Data Processing, Chen(PDF

> Optimizing Map Reduce for GPUs with Effective Shared Memory, Chen (PDF

> MITHRA : Scaling CUDA to Clusters using MapReduce, Farivar (slides)

> A Map-Reduce Based Framework for Heterogeneous Processing Element Cluster Environments, Tan(PDF)

 

Source:

http://www.nvidia.com/object/data-science-analytics-database.html