Data-Intensive Computing: From Multi-Cores and GPGPUs to Cloud Computing and Deep Web

1. Data-Intensive Computing: From Multi-Cores and GPGPUs to Cloud Computing and Deep Web Gagan Agrawal u

2. Data-Intensive Computing • Simply put: scalable analysis of large datasets • How is it different from: related to • Databases: • Emphasis on processing of static datasets • Data Mining • Community focused more on algorithms, and not scalable implementations • High Performance / Parallel Computing • More focus on compute-intensive tasks, not I/O or large datasets • Datacenters • Use of large resources for hosting data, less on their use for processing

3. Why Now ? • Amount of data is increasing rapidly • Cheap Storage • Better connectivity, easy to move large datasets on web/grids • Science shifting from compute-X to X-informatics • Business intelligence and analysis • Google’s Map-Reduce has created excitement

4. Architectural Context • Processor architecture has gone through a major change • No more scaling with clock speeds • Parallelism – multi-core / many-core is the trend • Accelerators like GPGPUs have become effective • More challenges for scaling any class of applications

5. Grid/Cloud/Utility Computing • Cloud computing is a major new trend in industry • Data and computation in a Cloud of resources • Pay for use model (like a utility) • Has roots in many developments over the last decade • Service-oriented computing, Software as a Service (SaaS) • Grid computing – use of wide-area resources

6. My Research Group • Data-intensive computing on emerging architectures • Data-intensive computing in Cloud Model • Data-integration and query processing – deep web data • Querying low-level datasets through automatic workflow composition • Adaptive computation – time as a constraint

7. Personnel • Current students • 6 PhD students • 2 MS thesis students • Talking to several first year students • Past students • 7 PhDs completed between 2005 and 2008

8. Outline • FREERIDE: Data-intensive Computing on Cluster of Multi-cores • A system for exploiting GPGPUs for data-intensive computing • FREERIDE-G: Data-intensive computing on Cloud Environments • Quick overview of three other projects

9. FREERIDE - Motivation • Availability of very large datasets and it’s analysis (Data-intensive applications) • Adaptation of Multi-core and inevitability of parallel programming • Need for abstraction of difficulties of parallel programming.

10. FREERIDE • A middle-ware for parallelizing Data-intensive applications • Motivated by difficulties in implementing and performance tuning of Datamining applications • Based on observation of similar generalized reduction among datamining, OLAP and other scientific applications

11. Generalized Reduction structure

12. SMP Techniques • Full-replication(f-r) (obvious technique) • Locking based techniques • Full-locking (f-l) • Optimized Full-locking(o-f-l) • Fixed Locking(fi-l) • Cache-sensitive locking( Hybrid of o-f-l & fi-l)

13. Memory Layout of SMP techs

14. Experimental setup • Intel Xeon E5345 CPU • 2 Quad-core machine • Each core 2.33GHz • 6GB Main memory • Nodes in cluster connected by Infiniband

15. Experimental Results – K-means (CMP)

16. K-means (cluster)

17. Apriori (CMP)

18. Apriori (cluster)

19. E-M (CMP)

20. E-M (cluster)

21. Summary of Results • Both Full-replication and Cache-sensitive locking can outperform each other based on the nature of application • Cache-sensitive locking seems to have high overhead when there is little computation between updates in ReductionObject • MPI processes competes well with best of other two when run on smaller cores, but experiences communication overheads when run on larger number of cores

22. Background: GPU Computing • Multi-core architectures are becoming more popular in high performance computing • GPU is inexpensive and fast • CUDA is a high level language that supports programming on GPU

23. Architecture of GeForce 8800 GPU (1 multiprocessor)‏

24. Challenges of Data-intensive Computing on GPU • SIMD shared memory programming • 3 steps involved in the main loop • Data read • Computing update • Writing update

25. Complication of CUDA Programming • User has to have thorough knowledge of the architecture of GPU and the programming model of CUDA • Must specify the grid configuration • Has to deal with the memory allocation and copy • Need to know what data to be copied onto shared memory and how much shared memory to use • ……

26. Architecture of the Middleware • User input • Code analyzer • Analysis of variables (variable type and size) • Analysis of reduction functions (sequential code from the user) • Code Generator ( generating CUDA code and C++ code invoking the kernel function)

27. Architecture of the middleware Variable Analyzer Host Program Variable information Variable Access Pattern and Combination Operations Kernel functions Reduction functions Code Generator Grid configuration and kernel invocation Optional functions Code Analyzer( In LLVM) Executable

28. Variables to be used in the reduction function Values of each variable (typically specified as length of arrays)‏ A sequential reduction function Optional functions (initialization function, combination function…)‏ User Input

29. Analysis of Sequential Code • Get the information of access features of each variable • Figure out the data to be replicated • Get the operator for global combination • Calculate the size of shared memory to use and which data to be copied to shared memory

30. Experiment Results Speedup of k-means

31. Speedup of EM

32. Emergence of Cloud and Utility Computing • Group generating data • use remote resources for storing data • Already popular with SDSC/SRB • Scientist interested in deriving results from data • use distinct but remote resources for processing • Remote Data Analysis Paradigm • Data, Computation, and User at Different Locations • Unaware of location of other

33. Remote Data Analysis Advantages Flexible use of resources Do not overload data repository No unnecessary data movement Avoid caching process once data Challenge: Tedious details: Data retrieval and caching Use of parallel configurations Use of heterogeneous resources Performance Issues Can a Grid Middleware Ease Application Development for Remote Data Analysis and Yet Provide High Performance ?

34. Our Work FREERIDE-G (Framework for Rapid Implementation of Datamining Engines in Grid) Enable Development of Flexible and Scalable Remote Data Processing Applications Middleware user Repository cluster Compute cluster

35. Challenges • Support use of parallel configurations • For hosting data and processing data • Transparent data movement • Integration with Grid/Web Standards • Resource selection • Computing resources • Data replica • Scheduling and Load Balancing • Data Wrapping Issues

36. FREERIDE (G) Processing Structure KEY observation: most data mining algorithms follow canonical loop Middleware API: • Subset of data to be processed • Reduction object • Local and global reduction operations • Iterator Derived from precursor system FREERIDE While( ) { forall( data instances d) { (I , d’) = process(d) R(I) = R(I) op d’ } ……. }

37. FREERIDE-G Evolution FREERIDE data stored locally FREERIDE-G • ADR responsible for remote data retrieval • SRB responsible for remote data retrieval FREERIDE-G grid service Grid service featuring • Load balancing • Data integration

38. Evolution FREERIDE-G-ADR Application Data ADR SRB Globus FREERIDE FREERIDE-G-SRB FREERIDE-G-GT

39. FREERIDE-G System Architecture

40. Compute Node More compute nodes than data hosts Each node: • Registers IO (from index) • Connects to data host While (chunks to process) • Dispatch IO request(s) • Poll pending IO • Process retrieved chunks

41. FREERIDE-G in Action Compute Node Data Host I/O Registration Connection establishment SRB Agent While (more chunks to process) I/O request dispatched Pending I/O polled MCAT SRB Master Retrieved data chunks analyzed SRB Agent Compute Node

42. Implementation Challenges • Interaction with Code Repository • Simplified Wrapper and Interface Generator • XML descriptors of API functions • Each API function wrapped in own class • Integration with MPICH-G2 • Supports MPI • Deployed through Globus components (GRAM) • Hides potential heterogeneity in service startup and management

43. Experimental setup Organizational Grid: • Data hosted on Opteron 250 cluster • Processed on Opteron 254 cluster • Connected using 2 10 GB optical fibers Goals: • Demonstrate parallel scalability of applications • Evaluate overhead of using MPICH-G2 and Globus Toolkit deployment mechanisms

44. Deployment Overhead Evaluation Clearly a small overhead associated with using Globus and MPICH-G2 for middleware deployment. Kmeans Clustering with 6.4 GB dataset: 18-20%. Vortex Detection with 14.8 GB dataset: 17-20%.

45. 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 wangfa@cse.ohio-state.edu

46. Motivating Example Given a gene ERCC6, we want to know the amino acid occurring in the corresponding position in orthologous gene of non-human mammals dbSNP ERCC6 AA Positions for Nonsynonymous SNP Alignment Database Encoded Protein Entrez Gene Protein Sequence Sequence Database Encoded Orthologous Protein wangfa@cse.ohio-state.edu

47. Observations • Inter-dependences between sources • Time consuming if done manually • Intelligent order of querying • Implicit sub-goals in user query wangfa@cse.ohio-state.edu

48. Contributions • Formulate the query planning problem for deep web databases with dependences • Propose a dynamic query planner • Develop cost models and an approximate planning algorithm • Integrate the algorithm with a deep web mining tool wangfa@cse.ohio-state.edu

49. HASTE Middleware Design Goals • To Enable the Time-critical Event Handling to Achieve the Maximum Benefit, While Satisfying the Time 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

50. HASTE Middleware Design ICAC 2008