Transcription

HPC in the Cloud Built forAtmospheric ModelingKevin Van Workum, PhDSabalcore Computing ore.com

About Us HPC in the Cloud provider since 2000Focused on Engineering and Scientific application user baseServe the private-sector, academia, and government agenciesWeather Modeling, CFD, FEA, Material Science, MD, Finance, Life Sciences,Image Processing, etc.100% Bare-Metal hardwareSingle-Origin HPC providerwww.sabalcore.com

Atmospheric Research Applications WRF, WRF-Chem, WRFDA, NCAR ADCIRCSWANFVCOMwww.sabalcore.com

Clients: Department of DefenseDedicated HPC in the Cloud Services Provided reliable support and resources at critical timesITAR complianceDedicated HPC Cloud operations and maintenanceCOAMPS, WRF, NCAR GraphicsCustom software stackMulti year support contractwww.sabalcore.com

Clients: Weather Analytics, LLCHPC On-Demand Services Risk mitigation and predictive analytics productsOn-Demand HPC for post-event hindcast analysisCustom WRF executablesWorkflow automation and production supportMassive data storage and managementwww.sabalcore.com

HPC Cloud Adoption ChallengesYour application was notbuilt to run in the Cloud.Choose a Cloud built to runyour application.www.turnoff.us, licencewww.sabalcore.com

HPC Cloud Adoption Challenges Starting is hardEveryone is differentPerformance is critical Where’s my data?Steep learning curve Myriad of configuration options which is ideal for your application?what if you make the wrong choice?Managing Software Stack foreign environmentcompiler selectionsupporting libraries and compatibilityoptimizationJob Execution data stagingautomationmonitoring / notificationwww.sabalcore.com

HPC Cloud Adoption Challenges Starting is hardEveryone is differentPerformance is critical Application requirements Workflows software versionscustomized or in-house codeperformance, reproducible, or accuratedevelop, test, and evaluateautomated/scheduledunstructured approachKnowledge Level HPC power-usersapplication only expertshold my hand pleasewww.sabalcore.com

HPC Cloud Adoption Challenges Starting is hardEveryone is differentPerformance is critical High performance Infiniband interconnectsHigh performance parallel filesystemsDirect-attached storage domainsBare-metal coresCareful considerations supporting libraries and versionscompiler vendor (Intel, Portland, GNU, etc)build-time optimizationsmessage passing interfacenetwork tuningOS kernel tuningparallel file-system tuningapplication scalabilitywww.sabalcore.com

Solutions: Simple Familiar environment Managed software stack Works just like an in-house HPC facility orprivate clusterEasily switch software packages or versionsExperienced technical support (systemand applications)Tutorials get you up and running inminutessabalcore: Linuxsabalcore: sabalcore: sabalcore: uname vi wrf.pbs qsub wrf.pbs qstatwww.sabalcore.com

Solutions: Flexible Command-line and GUI Many software versions installed 20 WRF combinations (versions, compilers,models)request a custom installationBuild your own software terminal or remote desktop accessseveral compiler optionsinteractive jobsPre-Post processing online or on-deskAutomation and custom scriptingsabalcore: sabalcore: sabalcore: sabalcore: module load WRF/3.8.1mpiexec ./wrf.exemodule load WRF/3.6.1mpiexec ./wrf.exewww.sabalcore.com

Solutions: Powerful Already optimized at the software and hardware level Infiniband low-latency interconnectsuperior scaling for large number of corescritical for applications like WRFDirect-attached Storage Domain applications have been expertly compiledbenchmarked for speedHigh performance parallel file systemshandles large application level I/O loadsIntel Xeon processorswww.sabalcore.com

CCU v8.01a - Atmospheric Modeling Cluster 4800 cores (demand driven)Dual Intel Xeon E5-2667v4 Broadwell-EP 3.2GHz37.5 TB RAM, 8GB per core1 TB SSD local scratch driveInfiniband EDR Interconnect 100 Gbps throughput per portultra-low latencyBeeGFS parallel file system 80 Gbps sustained I/OScalable to 1.5PBwww.sabalcore.com

Our Data CentersTampa East N 1 RedundancySSAE-16 SOC 1 Type 1 certifiedSSAE-16 SOC 2 Type 1 certifiedISAE-3402NIST 800-171 and ITAR controlsAWS S3 Direct Connect, Internet-2Orlando N 1 RedundancySSAE-16 SOC 1 Type 1 certifiedSSAE-16 SOC 2 Type 1 certifiedISAE-3402NIST 800-171 and ITAR controlsFISMA (high) and FedRAMP Capable www.sabalcore.com

What’s Next Can our applications be re-designed for the Cloud? Decouple the execution kernel and user interface (MVC or server-client model)Decentralize data locality and deduplicationIs this feasible?Choose a Cloud built to run your applicationwww.sabalcore.com

Questions & Answers“We would like to thank Sabalcore’s team for the excellent, high-quality services (and support) provided since last year.”Dr.-Ing. Paulo B., WRF user, Federal University of Rio Grande do Sul, Brazil, January 2016“I ran my tests and the results were amazing. I got my code to run 3 times faster. You did a very good job compiling the code forme.”Wiktor w., Rockseis, Norway, December 2015“Jobs are running nicely now! Thanks for all the help today! Great service!”Havard M., Senior Engineer, Norway, August 2016“Thank you and I do appreciate the diligence in helping us “Martin P., Director, sensor manufacturing, UK, August 2016“I will definitely recommended [Sabalcore] in the future to anyone interested in HPC.”Chris C., Software Developer, Netherlands, March 2016www.sabalcore.com

HPC On-Demand Simple pricing: pay only for what you useScale to 1000 cores per jobSchedule automated daily jobs100 GB storage includedBandwidth includedRemote Visualization includedTechnical support includedSimple to learn and get startedIdeal for flexible requirementswww.sabalcore.com

Dedicated HPC in the Cloud Dedicated compute nodes and networksCustomize HPC designed for specific application and workloadScalable and off-loadable to On-Demand ServiceTechnical support includedBandwidth includedSpecial connectivity available (e.g. P2P, AWS Direct Connect, Internet2)Ideal for time-critical and large production workloadswww.sabalcore.com

SSAE-16 SOC 1 Type 1 certified SSAE-16 SOC 2 Type 1 certified ISAE-3402 NIST 800-171 and ITAR controls AWS S3 Direct Connect, Internet-2 Orlando N 1 Redundancy SSAE-16 SOC 1 Type 1 certified SSAE-16 SOC 2 Type 1 certified ISAE-3402 NIST 8