Gelato Home
  
About Gelato
Home > About > What is Gelato?
What Is Gelato?

The Gelato Federation is the global technical community dedicated to advancing Linux® on the Intel® Itanium® platform through collaboration, education, and leadership.

Co-founded by HP and seven of the world's leading research institutions, the Gelato Federation, also known as Gelato, is working to fulfill this mission by bringing together those developing and using Linux on Itanium for the purpose of fostering Itanium-specific research, focus groups, and collaborative discussions. An open exchange of ideas and solutions is nurtured through diverse lines of communication, including interactive areas of the Gelato portal, several interest-specific mailing lists, and regular meetings and conferences. Finally, Gelato facilitates the development and dissemination of Itanium-specific, open-source solutions, including a complete and robust tool chain. Gelato invites participation from all interested organizations and individuals.

Gelato Central Operations, under the direction of Professor Wen-mei Hwu, is housed within the Coordinated Science Laboratory (CSL) of the College of Engineering at the University of Illinois at Urbana-Champaign (UIUC).

History: The Gelato Manifesto
February 7, 2002

The academic, government, and industrial research community is increasingly turning to commodity computing – i.e., commodity OS'es running on commodity hardware – to develop, execute and share work.

While IA-32 Linux is the commodity research platform of choice today, we expect Linux on Itanium to succeed it in the near future. The speed by which this transition occurs will greatly depend on the availability of software solutions that exploit the architectural advantages of Itanium and further advance the capabilities of Linux – from the node, to the cluster, to the Grid.

There are many outstanding non-profit open-source organizations operating today that touch pieces of this problem, but none is dedicated solely to making Linux on Itanium a successful research platform.

In January 2002, a group of research organizations* from around the world agreed to create a federation, called Gelato, dedicated to enabling scalable, open-source commodity computing solutions for Linux on Itanium to address real-world problems in academic, government, and industrial research.

* Bioinformatics Institute, Singapore; Groupe ESIEE, France; Hewlett-Packard Company; National Center for Supercomputing Applications, USA; Tsinghua University, China; University of Illinois, Urbana-Champaign, USA; University of New South Wales, Australia; and University of Waterloo, Canada.

Gelato Values and Core Technologies

Gelato's primary objective is to be responsive to the needs and interests of Itanium technical communities in government, industry, and academe. The Federation seeks to attract people and institutions of the highest quality for membership and sponsorship, thus continuing its established tradition of excellence in the advancement of Linux Itanium.

The Gelato Federation delivers lasting value to Members and Sponsors through careful balance of short-term goals (meeting current real-world problems) and long-term goals (cutting edge research). Essential to Gelato is the notion of an open exchange of ideas and solutions, which is expressed in part by Gelato's commitment to facilitating the development of open-source software for the high performance computing community.

Gelato Members are looking for ways to apply new technology and solutions to specific applications in bioinformatics, high-energy physics, and atmospheric sciences, in direct response to Members' and Sponsors' core technologies and needs.

Governance: The Gelato Steering Committee

Gelato is headed by the Managing Director, who is sited at the University of Illinois at Urbana-Champaign, which serves as the worldwide hub for Gelato activities. The Steering Committee works with the Managing Director to review technical activity and administrative issues. The Committee sets priorities and determines the long-term direction of the Federation, provides high-level coordination and planning, and ensures that Gelato values prevail.

On the organizational side, the Steering Committee manages the growth of Gelato membership, evaluates and admits new Members and Sponsors, and establishes and oversees Gelatos public image and high-level message content.

Current Members and Their Interests
Members

Academia Sinica Computing Centre -- ASCC (Taiwan)
CRASA homology-based sequence annotation tool; BioPortal Web interface; mpiBLAST, an MPI wrapper of NCBI Blast

American University of Cairo -- AUC (Eqypt)
bioinformatics, mathematical function libraries, bio-medical modeling, stress analysis, and parallel ray tracing

Australian Partnership for Advanced Computing -- APAC (Australia)
management software for large-scale cluster: batch system with detailed utilization monitoring, NUMA-aware allocations and cpusets, a kernel module accurate job memory usage and page unreferencing of suspended jobs, parallel launch utility; application porting and tuning including large-scale chemistry packages

Bioinformatics Institute -- BII (Singapore)
bioinformatics application porting and tuning

CEA (France)
applications, HW administration, libraries

European Organization for Nuclear Research -- CERN (Switzerland)
high-energy physics library and application porting and tuning

Federal University of Rio de Janeiro -- UFRJ (Brazil)
develop parallel applications for clusters for graph partitioning, computational fluid dynamics, petroleum, data mining, data bases, and energy; deploy grid services for the Brazilian High Performance System; developed an accounting system for parallel jobs in PBS

French National Institute for Research in Computer Science and Control -- INRIA (France)
cluster and grid applications, tools, and filesystems

Fudan University (China)
grid software and tools, grid-based workflow system and data mining software

Georgia Institute of Technology -- GA Tech (USA)
filesystems, kernel development, performance and performance monitoring, security, cluster/grid software and tools

Huazhong University of Science and Technology -- HUST (China)
Clustering, filesystems, grid computing middleware, porting image processing and bioinformatics applications to Itanium

Institute for System Programming, Russian Academy of Sciences -- RAS-ISP (Russia)
cluster development web-based tools, GCC performance studies and extensions

Institute of Computing Technology, Chinese Academy of Sciences -- ICT (China)
Open Research Compiler (ORC) on Itanium, NBD driver on Itanium

Institute of High Performance Computing -- IHPC (Singapore)
performance benchmarking and optimization, remote visualization

Instituto Tecnologico de Estudios Superiores de Monterrey -- ITESM (Mexico)
An extension of User-Mode Linux called User Mode KU Real Time (UMK), Personal Digital Library (PDLib) with mobility support

Lawrence Livermore National Lab -- LLNL (USA)
IA-64 kernel and system management tools, clustering, filesystems, scientific applications

Leibniz Computing Centre -- LRZ (Germany)
providing relevant parts of the LRZ benchmark suite to measure the performance of supercomputers (e.g. CPU, memory, disk and network I/O, etc.)

National Aeronautics and Space Administration -- NASA (USA)
compilers, filesystems, kernel development, memory, performance, scheduling, security, applications, HW administration, libraries

National Center for Supercomputing Applications -- NCSA (USA)
OSCAR clustering software (OSCAR Gold), virtual machine interface (VMI)

National University of Defense Technology -- NUDT (China)
GCC compiler work: FORTRAN 95 syntax implementation and performance improvements for GCC; significant floating point performance improvements on Itanium 2 systems

National Grid Office -- NGO (Singapore)
Grid Accounting System (GAS); providing a portal to access grid resources

North Carolina State University -- NCSU (USA)
a hardware-assisted approach to support first-touch page placement on a large-scale Itanium-2 ccNUMA SMP, the SGI Altix; exploration of capabilities for the online selection of parallelizing optimizations

Northwestern Polytechnical University -- NPU (China)
optimize the performance of clusters using parallel programming techniques: develop job scheduling, a resource management suit, and a grid portal; scientific applications: develop applications for material design, computational fluid dynamics, and structural computing

Ohio Supercomputer Center -- OSC (USA)
remote heterogeneous cluster management and administration tools, parallel filesystems, training

Osaka University -- Osaka U (Japan)
grid software: developed a grid-aware secure filesystem by extending an existing distributed filesystem

Pacific Northwest National Lab -- PNNL (USA)
parallel file system development, job allocation management, management and monitoring software

Peking University -- PKU (China)
LilyTask parallel programming language, LilyTask runtime rupport, LilyHPC management suite, PKU grid software

Pittsburgh Supercomputing Center -- PSC (USA)
heterogeneous architecture & interoperability, job management & scheduling, maintenance monitoring, parallel file system development

Program Systems Institute, Russian Academy of Sciences -- RAS-PSI (Russia)
T-system migration tool for parallel computing on clusters and grids, which is designed to simplify both tasks of development of new parallel programs and parallelizing existing ones

Pontifical Catholic University of Rio Grande do Sul -- PUCRS (Brazil)
adaptation of cluster monitoring and management tools to work with small heterogeneous Linux Itanium clusters, Linux scalability

Pontifical Catholic University of Rio de Janeiro -- PUC-Rio (Brazil)
cluster and grid software: a set of tools to manage resources and applications in cluster and grid environments (CSGrid)

Purdue University (USA)
commercial printing/digital publishing applications and drivers for Itanium

Renmin University of China -- Renmin U (China)
parallel database management systems: developing a high-performance parallel and/or in-memory database management system in the MPP enviroment

Research Institute of Bioinformatics & Molecular Design -- BMD (Korea)
drug discovery, bioinformatics, and material design application tools porting and tuning

Rice University (USA)
In general, developing tools and software optimized for the Itanium platform. Specifically, HHPCToolkit, a performance tuninsg tools for IA-64, and porting superpages from FreeBSD for Alpha to the IA-64 Linux kernel.

RIKEN Genomic Sciences Center -- RIKEN (Japan)
Bioinformatics sequence analysis using Itanium

Royal Institute of Technology -- KTH (Sweden)
tools for application performance measurement using performance counters, development and support of scalable security infrastructures, adaptivity of codes to the grid computing environment

South China University of Technology -- SCUT (China)
cluster performance monitoring, bioinformatics analysis tools

Southern Partnership for Advanced Computational Infrastructure -- SPACI Consortium (Italy)
Grid libraries, Grid problem solving environments for remote sensing, bioinformatics, atmospheric and climate modeling, industry simulation

Supercomputing Center of Galicia -- CESGA (Spain)
performance and programmability improvement of HPC systems; software tools for HPC and grid facilities

Tsinghua University (China)
parallel debugger for PVM and MPI, checkpoint and rollback recovery resource manager for PVM, enable ORC as OpenMP compiler

Universidad Carlos III de Madrid -- UC3M (Spain)
MiMPI-a multithreaded implementation of MPI; EXPAND-a parallel filesystem based currently on NFS servers adequated for heterogeneous architectures and OSes; Clusterfile-a parallel file system for clusters of commodity computers; also a member os the Spanish IrisGrid Initiative

Universidade Federal de Campina Grande -- UFCG (Brazil)
OurGrid user-centric middleware, clustering, performance monitoring, scheduling and process management

University of Buenos Aires -- UBA (Argentina)
grid computing, performance, security

University of California, Berkeley (USA)
cluster and grid monitoring tools (Ganglia), compilers (UPC, Java), software performance tuning, math library development (LAPACK), networks (PlanetLab)

University of Chile (Chile)
Port of Climate Model MM5, a climate model programmed mostly in FORTRAN; port of Omnet++ Transportation Model, an open-source network simulator (discrete) programmed in C++; and an energy and telecommunications computing algorithm.

University of Copenhagen -- DIKU (Denmark)
Running the Badger Project, a study of the performance of database systems on modern computer platforms; testing new types of grid middleware, especially with a focus on minimal configuration and intrusion; and enhancing the performance of MySQL on Linux on Itanium

University of Delaware -- U Delaware (USA)
compilers: research, maintain, and develop Open64

University of Houston (USA)
storage, visualization, and networking facilities; adaptive FFT libraries; applications in life sciences, CFD, and seismic processing; adaptation of existing software to grid environments

University of Illinois at Urbana-Champaign -- UIUC (USA)
OpenIMPACT C compiler, recompilation of performance-critical libraries and applications

University of Karlsruhe (Germany)
parallel filesystem development, scalable servers, coordinated process scheduling, Java development environment

University of Manchester -- U Manchester (United Kingdom)
port and tune scientific applications for SGI Altix systems

University of Minnesota -- U Minnesota (USA)
ADORE, a transparent dynamic object code optimization system that can increase application execution performance on Linux-Itanium; enhancing the ORC compiler with data speculative optimizations such as data speculative register promotion and data speculative PRE

University of New South Wales -- UNSW (Australia)
performance and scalability improvements for the Linux kernel and libraries (user-level device drivers, new virtual memory infrastructure, preemption support, microstate accounting for multithreaded processes, POSIX threads, fast system calls, large block device support (> 1TB), TCP/IP performance analysis and improvement

University of Puerto Rico Mayaguez -- UPRM (Puerto Rico)
parallelize, port and tune signal processing, electromagnetics, acoustics and optics applications, software system for the automatic generation of high-performance, parallel multidimensional symmetric FFT libraries, Linux scalability

University of Sarajevo (Bosnia)
porting and tuning protein and nucleotide sequencing software for Itanium

University of Split (Craotia)
DIALIGN, a multiple DNA and protein sequence alignment program; HMMtop, a program for the prediction of transmembrane helices (based on the HMM model); AMBER, a molecular dynamics package; and GRALIC, a molecular visualization program

University of Tennessee -- U Tennessee (USA)
deploying a wide collection of mathematical software and tools on the Linux IA-64 architecture, including: LAPACK, ATLAS, BLAS, ScaLAPACK, and PAPI

University of Tokyo (Japan)
UTChem quantum chemical calculation program

University of Tromso -- U Tromso (Norway)
developing computational science codes, especially in the area of theoretical chemistry and oceanography; parallelizing the Dalton molecular electronic structure program on Itanium

University of Waterloo (Canada)
scalable event notification and delivery, concurrency-neutral system library

Xiamen University -- Xiamen U (China)
application for resource usage monitoring and performance tuning

Zhejiang University (China)
visual environment for simulation and visualization for multidisciplinary applications, and an information grid tool suite including specifications and middleware

Associate Members

French Engineering College of Electronics and Electrotechnology -- ESIEE (France)
Gelato portal development, hosting and maintenance; software packaging, integration, and testing

National Center for Atmospheric Research -- NCAR (USA)
efficient, modern, and portable spectral transform software libraries optimized for Itanium

San Diego Supercomputer Center -- SDSC (USA)
NPACI Rocks clustering software

Saint-Petersburg State University of Telecommunications -- SUT (Russia)
performance/performance monitoring and security

Swiss National Supercomputing Centre -- CSCS (Switzerland)
performance monitoring and benchmarking in an heterogeneous environment, scientific applications, visualization packages

University of Bristol (United Kingdom)
Porting of existing particle physics application software and libraries and grid middleware

University of Guadalajara -- U Guadalajara (Mexico)
grid software and security, probe and port open-source software for climate modeling using WRF, develop techniques for fabricating Fullerene Nanostructures and rendering (blender)

-> Interested in Participating?






Featured Gelato Leader
Terence Hung
Terence Hung is the manager of the Software and Computing Programme at the Institut of High Performance Computing.
-> Learn More

-> View other Gelato leaders


-
-
-
-
-

 

All content © copyright 2002-2006 Gelato Federation. Click here to view the Gelato Federation Privacy Policy and Terms of Service Agreement. If you have any questions or comments, please contact us.

Gelato Central Operations is housed within the Coordinated Science Laboratory (CSL) of the College of Engineering at the University of Illinois at Urbana-Champaign (UIUC).