Download presentation
Presentation is loading. Please wait.
Published byGeorge Howard Modified over 8 years ago
1
Data-Intensive Computing: From Clouds to GPUs Gagan Agrawal June 1, 20161
2
2 Growing need for analysis of large scale data Scientific Commercial Data-intensive Supercomputing (DISC) Map-Reduce has received a lot of attention Database and Datamining communities High performance computing community Motivation
3
Motivation (2) Processor Architecture Trends Clock speeds are not increasing Trend towards multi-core architectures Accelerators like GPUs are very popular Cluster of multi-cores / GPUs are becoming common Trend Towards the Cloud Use storage/computing/services from a provider How many of you prefer gmail over cse/osu account for e-mail? Utility model of computation Need high-level APIs and adaptation June 1, 20163
4
My Research Group Data-intensive theme at multiple levels Parallel programming Models Multi-cores and accelerators Adaptive Middleware Scientific Data Management / Workflows Deep Web Integration and Analysis June 1, 20164
5
Personnel Currently 10 PhD students 4 MS thesis students Graduated PhDs 7 Graduated PhDs between 2005 and 2008 June 1, 20165
6
This Talk Parallel Programming API for Data-Intensive Computing An alternate API and System for Google’s Map- Reduce Show actual comparison Data-intensive Computing on Accelerators Compilation for GPUs Overview of other topics Scientific Data Management Adaptive and Streaming Middleware Deep web June 1, 20166
7
Map-Reduce Simple API for (data-intensive) parallel programming Computation is: Apply map on each input data element Produce ( key,value ) pair(s) Sort them using the key Apply reduce on the set with a distinct key values June 1, 20167
8
8 Map-Reduce Execution
9
June 1, 20169 Positives: Simple API Functional language based Very easy to learn Support for fault-tolerance Important for very large-scale clusters Questions Performance? Comparison with other approaches Suitability for different class of applications? Map-Reduce: Positives and Questions
10
Class of Data-Intensive Applications Many different types of applications Data-center kind of applications Data scans, sorting, indexing More ``compute-intensive’’ data-intensive applications Machine learning, data mining, NLP Map-reduce / Hadoop being widely used for this class Standard Database Operations Sigmod 2009 paper compares Hadoop with Databases and OLAP systems What is Map-reduce suitable for? What are the alternatives? MPI/OpenMP/Pthreads – too low level? June 1, 201610
11
Hadoop Implementation June 1, 201611 HDFS Almost GFS, but no file update Cannot be directly mounted by an existing operating system Fault tolerance Name node Job Tracker Task Tracker
12
June 1, 201612 FREERIDE: GOALS Framework for Rapid Implementation of Data Mining Engines The ability to rapidly prototype a high- performance mining implementation Distributed memory parallelization Shared memory parallelization Ability to process disk-resident datasets Only modest modifications to a sequential implementation for the above three Developed 2001-2003 at Ohio State June 1, 201612
13
FREERIDE – Technical Basis June 1, 201613 Popular data mining algorithms have a common canonical loop Generalized Reduction Can be used as the basis for supporting a common API Demonstrated for Popular Data Mining and Scientific Data Processing Applications While( ) { forall (data instances d) { I = process(d) R(I) = R(I) op f(d) } ……. }
14
June 1, 201614 Similar, but with subtle differences Comparing Processing Structure
15
Observations on Processing Structure Map-Reduce is based on functional idea Do not maintain state This can lead to sorting overheads FREERIDE API is based on a programmer managed reduction object Not as ‘clean’ But, avoids sorting Can also help shared memory parallelization Helps better fault-recovery June 1, 201615
16
June 1, 201616 Tuning parameters in Hadoop Input Split size Max number of concurrent map tasks per node Number of reduce tasks For comparison, we used four applications Data Mining: KMeans, KNN, Apriori Simple data scan application: Wordcount Experiments on a multi-core cluster 8 cores per node (8 map tasks) Experiment Design
17
June 1, 201617 KMeans: varying # of nodes Avg. Time Per Iteration (sec) # of nodes Dataset: 6.4G K : 1000 Dim: 3 Results – Data Mining
18
June 1, 201618 Results – Data Mining (II) June 1, 201618 Apriori: varying # of nodes Avg. Time Per Iteration (sec) # of nodes Dataset: 900M Support level: 3% Confidence level: 9%
19
June 1, 201619 June 1, 201619 KNN: varying # of nodes Avg. Time Per Iteration (sec) # of nodes Dataset: 6.4G K : 1000 Dim: 3 Results – Data Mining (III)
20
June 1, 201620 Wordcount: varying # of nodes Total Time (sec) # of nodes Dataset: 6.4G Results – Datacenter-like Application
21
June 1, 201621 KMeans: varying dataset size Avg. Time Per Iteration (sec) Dataset Size K : 100 Dim: 3 On 8 nodes Scalability Comparison
22
June 1, 201622 Wordcount: varying dataset size Total Time (sec) Dataset Size On 8 nodes Scalability – Word Count
23
June 1, 201623 Four components affecting the hadoop performance Initialization cost I/O time Sorting/grouping/shuffling Computation time What is the relative impact of each ? An Experiment with k-means Overhead Breakdown
24
June 1, 201624 Varying the number of clusters (k) Avg. Time Per Iteration (sec) # of KMeans Clusters Dataset: 6.4G Dim: 3 On 16 nodes Analysis with K-means
25
June 1, 201625 Varying the number of dimensions Avg. Time Per Iteration (sec) # of Dimensions Dataset: 6.4G K : 1000 On 16 nodes Analysis with K-means (II)
26
Observations June 1, 201626 Initialization costs and limited I/O bandwidth of HDFS are significant in Hadoop Sorting is also an important limiting factor for Hadoop’s performance
27
This Talk Parallel Programming API for Data-Intensive Computing An alternate API and System for Google’s Map- Reduce Show actual comparison Data-intensive Computing on Accelerators Compilation for GPUs Overview of other topics Scientific Data Management Adaptive and Streaming Middleware Deep web June 1, 201627
28
Background - GPU Computing Many-core architectures/Accelerators are becoming more popular GPUs are inexpensive and fast CUDA is a high-level language for GPU programming
29
CUDA Programming Significant improvement over use of Graphics Libraries But.. Need detailed knowledge of the architecture of GPU and a new language Must specify the grid configuration Deal with memory allocation and movement Explicit management of memory hierarchy
30
Parallel Data mining Common structure of data mining applications (FREERIDE) /* outer sequential loop *//* outer sequential loop */ while() { while() { /* Reduction loop */ /* Reduction loop */ Foreach (element e){ Foreach (element e){ (i, val) = process(e); (i, val) = process(e); Reduc(i) = Reduc(i) op val; Reduc(i) = Reduc(i) op val; } }
31
Porting on GPUs High-level Parallelization is straight-forward Details of Data Movement Impact of Thread Count on Reduction time Use of shared memory
32
Architecture of the System Variable information Reduction functions Optional functions Code Analyzer( In LLVM) Variable Analyzer Code Generator Variable Access Pattern and Combination Operations Host Program Grid configuration and kernel invocation Kernel functions Executable User Input
33
A sequential reduction function Optional functions (initialization function, combination function…) Values of each variable or size of array Variables to be used in the reduction function
34
Analysis of Sequential Code Get the information of access features of each variable Determine the data to be replicated Get the operator for global combination Variables for shared memory
35
Memory Allocation and Copy Copy the updates back to host memory after the kernel reduction function returns C.C.C.C. Need copy for each thread T0T1 T2 T3 T4 T61T62 T63T0T1 …… T0T1 T2T3T4 T61T62 T63T0T1 …… A.A.A.A. B.B.B.B.
36
Generating CUDA Code and C++/C code Invoking the Kernel Function Memory allocation and copy Thread grid configuration (block number and thread number) Global function Kernel reduction function Global combination
37
Optimizations Using shared memory Providing user-specified initialization functions and combination functions Specifying variables that are allocated once
38
Applications K-means clustering EM clustering PCA
39
Experimental Results Speedup of k-means
40
Speedup of k-means on GeForce 9800X2
41
Speedup of EM
42
Speedup of PCA
43
wangfa@cse.ohio-state.edu Deep Web Data Integration The emerge of deep web Deep web is huge Different from surface web Challenges for integration Not accessible through search engines Inter-dependences among deep web sources
44
Our Contributions Structured Query Processing on Deep Web Schema Mining/Matching Analysis of Deep web data sources P44
45
45 Adaptive Middleware To Enable the Time-critical Event Handling to Achieve the Maximum Benefit, While Satisfying the Time/Budget Constraint To be Compatible with Grid and Web Services To Enable Easy Deployment and Management with Minimum Human Intervention To be Used in a Heterogeneous Distributed Environment ICAC 2008
46
46 HASTE Middleware Design ICAC 2008
47
47 Workflow Composition System
48
Summary Growing data is creating new challenges in HPC, grid and cloud environments Number of topics being addressed Many opportunities for involvement 888 meets Thursdays 5:00 – 6:00, DL 280 June 1, 201648
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.