Relations with ISV and Open Source Stephane Requena GENCI Stephane.requena@genci.fr
Agenda of this session 09:15 09:30 Prof. Hrvoje Jasak: Director, Wikki Ltd. «HPC Deployment of OpenFOAM in an Industrial Setting» 09:30 09:45 S. Requena (GENCI) & C. Zannoni (CINECA) «ISV & Opensource : PRACE, business models and services» 09:45 10:15 Interactive Q&A part with the audience 2
The PRACE Mission: Implement the ESFRI Vision of a European HPC service European HPC-facilities at the top of an HPC provisioning pyramid Tier-0: 3-6 European Centres Tier-1: National Centres Tier-2: Regional/University Centres Tier-0 Tier-1 Tier-2 Provide services through a persistent Research Infrastructure Create a single European entity (international non profit association with seat in Brussels) Enable world-class science through large scale simulations Creation of a European HPC ecosystem involving all stakeholders HPC service providers on all tiers & Grid Infrastructures Scientific and industrial user communities The European HPC hard- and software industry 3
PRACE Tier-0 Systems 1 st Tier-0 System provides cycles since August 1, 2010 Jugene: BlueGene/P in GCS@Juelich 72 Racks, 1 PFlop/s Peak 35% of capacity provided to PRACE 2 nd Tier-0 System provides cycles since Jan 11, 2011 Curie: Bull Cluster (Intel CPUs) in GENCI@CEA 1.6 PFlop/s peak start Oct. 2011 (1 st step in 2010) 80% of capacity provided to PRACE Next Tier0 Systems (in alphabetical order) BSC, CINECA, GCS@HLRS, GCS@LRZ Procurement plan based on analysis of user requirements and market 4
Engaging relations with industrial users Establishing a strong link with industrial users Organisation of regular industrial seminars Face to face meetings PRACE User Forum Working on industry needs and expectations Industrial seminars & dedicated face to face meetings Assessment of key industrial applications (Open Source, ISV, in-house) Selection of industrial pilots for using the PRACE RI facilities Elaborating a proposal for a PRACE industrial offer Business models and associated services for industry To be decided by the PRACE AISBL Council 5
Feedback from previous industrial seminars Interest on PRACE infrastructure and services Primary : Acces for open or pre competitive R&D needs Some interest for commercial activites : studies/production or real business (ISV for SaaS business models for example) Need to exchange and to collaborate with academia, numerics and hpc experts Needs on training, petascaling and technological watch Network and security issues important (access ok for industrial users, encryptation of data, redundancy and relialibilty of systems, archive,...) Major open issue in relation with ISVs : (the key point!) Licences cost and scalability of ISV apps Availability of commercial apps on Tier0/Tier1 systems Support of Open Source applications 6
DRAFT PRACE industrial offer Target Groups Industrial users at the tera-/ petascale level Self-developed code High level of expertise in using HPC systems Industrial users relying on ISV codes Not really interested in HPC as such Just want shorter time to result ideally in similar (certified environment) as at home ISV Vendors/Tool Vendors Validation/development of software and supporting infrastructure at large scale (e.g. license server for thousands of nodes) Hardware Vendors Aim for expertise for assessment of planned products Co-Development approaches 7
DRAFT Elements of the PRACE offer Differentiation Strategy Unique resources in terms of size and architecture (tier-0/1 systems) High level of expertise in emerging technologies and experimental technologies (e.g. WP9 prototypes) Layer integration Unification of access models, application procedures across several systems across tier-0/tier-1 Co-Desgin Collaboration approaches aiming to shorten time to market in Software, Hardware, Solutions/ Architecture Training and Consulting Lower the barrier to exploit available resources Support selection of most appropriate system for the problem size (tier-x) Training of trainers 8
DRAFT ISV Potential business models investigated Phase 0: Open R&D, Pre-competitive research/ Feasibility study (non-profit) In early stages research consortia that might involve also industrial users are investigating new methods (mathematical, numerical, ) Phase 1: Product development (profit orientation) Findings of phase 0 are integrated in ISV code/ Tool (could be free access if results are available under favourable conditions later) Phase 2: Product use (profit realisation) ISV code / Tools are offered on PRACE resources with dynamic license models (e.g. on demand tokens bought by users directly at the vendor) 9
DRAFT ISV Vendors / Tool vendors Caveat Why should an ISV/Tool vendor collaborate? Is there a market for >10.000 core ISV use? How to setup conditions for unlocking some current non scalable ISV models (performance, licences)? How does a reasonable price for >10.000 cores jobs impact the major market segment of 128 cores? Dynamic licensing is solved technically but is the pay-per-use model a business proposition for ISVs? Is the market large enough to support exotic hardware for an ISV (if not are they willing to allow others to do this for them?) 10
DRAFT PRACE high value services for industry Training : Dedicated industrial support PRACE seasonal schools PRACE Advanced Training Centers Online Training portal Application enabling Access to HPC ressources for perf evaluation, porting, optimisation Benefits from the expertise of all the PRACE members Ongoing pilots launched Technological watch : STRATOS Access to knowledge on emerging HPC Technologies CoDesign & participate to joint hardware/software prototypes Exchanges of expertise with other HPC users Best practises in energy- and cooling-efficiency, high density, parallelism, 11 reliability,
PRACE PABS applications Name QCD Quantum Espresso CPMD CP2K Gromacs Octopus GPAW Code Saturne NS3D GADGET TORB/EUTERPE PEPC WRF NEMO TRIPOLI-4 ELMER ALYA SPECFEM3D AVBP NAMD BSIT HELIUM Scientific domain Particle Physics Chemistry Chemistry Chemistry Chemistry Chemistry Chemistry CFD CFD Astrophysics Physique des plasmas Physique des plasmas atmospheric modeling atmospheric modeling Engineering Engineering computational mechanics Geophysics CFD Chemistry Geophysics computational physics
Application enabling services An essential activity in PRACE : >40% of the workforce Benefits from the expertise of all the PRACE members Petascaling Mixed-mode parallelisation Load balancing Minimisation of communication overheads Parallel I/O Checkpointing, Optimisation Optimising serial performance CPU Memory Optimising both for general-purpose architectures and specialised architectures Algorithmic optimisations, Services addressed by PRACE in a 2 ways vision : By adresssing needs of single users in a code by code approach è PRACE Preparatory access calls (6 months) By working in a mid term collaboration with users communities 13
PRACE support to industrial challenges Dr Cinzia Zannoni (CINECA)
Support to applications relevant for industrial research Objectives: Support industrial challenges Enable applications used in research projects relevant for industry to run on Tier0 and Tier1 system 15
Support to applications relevant for industrial research Tasks: Petascaling of relevant codes together with industrial users and developers community Continue and strengthen the relations with industrial users and ISVs Enabling emerging industry relevant open source codes for HPC systems 16
Petascaling of codes Evaluation and selection of open source applications from industrial R&D, design or production processes; Petascaling of the more relevant aplications, based on the maturity, interest and commitment expressed by industrial users and developers community; coverage of different industrial fields: automotive, drug design, mechanics, CFD, etc. The objective is to enable and petascale application codes from an Industrial R&D stage, to run in production on the PRACE RI Tier 0 /1 systems 17
Support for emerging opportunities Selection of emerging applications of high potential impact on industrial research and innovation, but not yet ready for industrial production Evaluate the promising applications both in terms of industrial interest and HPC challenge Porting and enabling to suitable systems within the PRACE RI (Tier-1 or Tier-0) Work in close collaboration with code owners and committed industrial users The open source application may become the nucleus for a business opportunity for a spinoff or an SME to create an industrial strength offering. 18