2. WHO WILL BENEFIT FROM THIS TALK TOPICS WHAT YOU’LL LEAVE WITH Developers looking to build applications that analyze big data. Developers building applications running in the cloud and on-premises. What is “Big Data”? Introducing LINQ to HPC. Demos. Conclusions. Q&A. A clear understanding of how big data is processed. How LINQ to HPC works. Some great examples of LINQ to HPC in action.

3. What is “Big Data”?

4. 4

7. RUNTIME var logentries = from line in logs where !line.StartsWith("#") select new LogEntry(line); var user = from access in logentries where access.user.EndsWith(@"\Ade") select access; var accesses = from access in user group access by access.page into pages select new UserPageCount(“Ade", pages.Key, pages.Count()); var htmAccesses = from access in accesses where access.page.EndsWith(".htm") orderby access.count descending select access;

9. Introducing LINQ to HPC

10. MPI Optimize CPU utilization for tightly coupled problems like climate modeling, car crash simulation, etc. 10 Optimize CPU utilization for loosely coupled problems like financial product pricing, etc. SOA Data-focused Compute-focused Optimize for data locality rather than CPU utilization to support jobs that are primarily bound by disk I/O. LINQ TO HPC NEW

11. Distributed runtimes Cluster and cloud services Platform Languages and Libraries Tools HPC provisioning, management, etc. MPI SOA LINQ to HPC Runtime Windows Server Azure Distributed Storage Catalog Bind individual NTFS shares together to support the distributed runtime LINQ to HPC C#, VB.NET, F#, … LINQ to HPC C#, VB.NET, F#, … Visual Studio.NET, LINQ, Excel, etc. Visual Studio.NET, LINQ, Excel, etc. New in HPC Server

12. HPC Head Node Application or command line creates a new DSC file set. DSC creates a new file set entry in it’s database. 1 1 HPC Compute Nodes DSC Application finishes adding files and Seals the DSC file set. 2b 2a Application adds a file. DSC returns a WritePath and the application writes the file. 2a 3 Once the DSC file set is sealed it becomes read- only and its files are replicated. 3

13. var logentries = from r in context.FromDsc ("Logs") where !r.Line.StartsWith("#") select new LogEntry(r.Line); var users = from access in logentries where access.User.EndsWith(@"\ade") select access; var accesses = from access in user group access by access.Page into pages select new UserPageCount(“ade", pages.Key, pages.Count()); var htmAccesses = from access in accesses where access.Page.EndsWith(".htm") orderby access.Count descending select access; htmAccesses.ToDsc("Result"); Output Compute and resort Read & compute Compute Input

14. HPC Head Node Application that calls LINQ to HPC APIs Submit LINQ to HPC Job 1 1 HPC Compute Nodes Graph Manager starts/stops Vertices 3a DSC A LINQ to HPC job starts 1 basic task assigning a node as the GM 2a LINQ to HPC Vertices read and write files 3b The LINQ to HPC job also starts a set of tasks across the rest of the nodes as VH 2b Graph Manager Vertex Host

23. Conclusions…

26. 452 - Building and Running HPC Applications in Windows Azure RELATED SESSIONS DOCUMENTATION & ARTICLES CONTACT: ade.miller@microsoft.com Introducing LINQ to HPC http://www.microsoft.com/download/ en/details.aspx?id=13615 http://www.microsoft.com/download/ en/details.aspx?id=13615 LINQ to HPC downloads: https://connect.microsoft.com/hpc https://connect.microsoft.com/hpc HPC Team Blog: http://blogs.technet.com/b/windows hpc/ http://blogs.technet.com/b/windows hpc/ My Blog http://ademiller.com/tech/ http://ademiller.com/tech/

27. Questions?

31. WHO WILL BENEFIT FROM THIS TALK TOPICS WHAT YOU’LL LEAVE WITH Developers looking to build applications that analyze big data. Developers building applications running in the cloud and on-premises. What is “Big Data”? Introducing LINQ to HPC. Demos. Conclusions. Q&A. A clear understanding of how big data is processed. How LINQ to HPC works. Some great examples of LINQ to HPC in action.

32. What is “Big Data”?

33. 33

36. RUNTIME var logentries = from line in logs where !line.StartsWith("#") select new LogEntry(line); var user = from access in logentries where access.user.EndsWith(@"\Ade") select access; var accesses = from access in user group access by access.page into pages select new UserPageCount(“Ade", pages.Key, pages.Count()); var htmAccesses = from access in accesses where access.page.EndsWith(".htm") orderby access.count descending select access;

38. Introducing LINQ to HPC

39. MPI Optimize CPU utilization for tightly coupled problems like climate modeling, car crash simulation, etc. 39 Optimize CPU utilization for loosely coupled problems like financial product pricing, etc. SOA Data-focused Compute-focused Optimize for data locality rather than CPU utilization to support jobs that are primarily bound by disk I/O. LINQ TO HPC NEW

40. Distributed runtimes Cluster and cloud services Platform Languages and Libraries Tools HPC provisioning, management, etc. MPI SOA LINQ to HPC Runtime Windows Server Azure Distributed Storage Catalog Bind individual NTFS shares together to support the distributed runtime LINQ to HPC C#, VB.NET, F#, … LINQ to HPC C#, VB.NET, F#, … Visual Studio.NET, LINQ, Excel, etc. Visual Studio.NET, LINQ, Excel, etc. New in HPC Server

41. HPC Head Node Application or command line creates a new DSC file set. DSC creates a new file set entry in it’s database. 1 1 HPC Compute Nodes DSC Application finishes adding files and Seals the DSC file set. 2b 2a Application adds a file. DSC returns a WritePath and the application writes the file. 2a 3 Once the DSC file set is sealed it becomes read- only and its files are replicated. 3

42. var logentries = from r in context.FromDsc ("Logs") where !r.Line.StartsWith("#") select new LogEntry(r.Line); var users = from access in logentries where access.User.EndsWith(@"\ade") select access; var accesses = from access in user group access by access.Page into pages select new UserPageCount(“ade", pages.Key, pages.Count()); var htmAccesses = from access in accesses where access.Page.EndsWith(".htm") orderby access.Count descending select access; htmAccesses.ToDsc("Result"); Output Compute and resort Read & compute Compute Input

43. HPC Head Node Application that calls LINQ to HPC APIs Submit LINQ to HPC Job 1 1 HPC Compute Nodes Graph Manager starts/stops Vertices 3a DSC A LINQ to HPC job starts 1 basic task assigning a node as the GM 2a LINQ to HPC Vertices read and write files 3b The LINQ to HPC job also starts a set of tasks across the rest of the nodes as VH 2b Graph Manager Vertex Host

52. Conclusions…

55. SAC-452T - Building and running HPC apps in Windows Azure RELATED SESSIONS DOCUMENTATION & ARTICLES Introducing LINQ to HPC http://www.microsoft.com/download/ en/details.aspx?id=13615 http://www.microsoft.com/download/ en/details.aspx?id=13615 LINQ to HPC downloads: https://connect.microsoft.com/hpc https://connect.microsoft.com/hpc HPC Team Blog: http://blogs.technet.com/b/windows hpc/ http://blogs.technet.com/b/windows hpc/

56. Questions?