C-DAC,Pune : High-Perf. Comp. Frontier Technologies Exploration Group and CMSD, University of Hyderabad, Technology Workshop hyPACK (October 15-18), 2013

Overview Venue : CMSD, UoH Key-Note/Invited Talks Faculty / Speakers Proceedings Downloads Past Tech. Workshops Target Audience Benefits Organisers Accommodation Local Travel Sponsors Feedback Acknowledgements Contact Home

Topics of Interest Tech. Prog. Schedule Topic : Multi-Core Topic : ARM Proc. Topic : Coprocessors Topic : GPGPUs Topic : HPC Cluster Topic : App. Kernels. Topic : Lab. Session Key-Note / Invited Talks Home

Mode-1 Multi-Core Memory Allocators OpenMP Intel TBB Pthreads Java - Threads Charm++ Prog. Message Passing (MPI) MPI - OpenMP MPI - Intel TBB MPI - Pthreads Compilers - Opt. Features Threads-Perf. Math. Lib. Threads-Prof. & Tools Threads - I/O Perf. PGAS : UPC / CAF/ GA Power & Perf. Home

Mode-2 ARM Prog. Env Benchmarks Power & Perf. Home

Mode-3 Coprocessors Arch. Software Compiler & Vect. Prog. Env. Benchmarks Power & Perf. Home

Mode-4 GPGPUs NVIDIA - CUDA/OpenCL AMD APP - OpenCL GPGPUs - OpenCL GPGPUs : Power & Perf. Home

Mode-5 HPC Cluster HPC MPI Cluster GPU Cluster - NVIDIA GPU Cluster - AMD APP Cluster - Intel Coprocessors Cluster- Power & Perf. Home

Mode-6 App. Kernels PDE Solvers : FDM/FEM Image Processing - FFT Monte Carlo Methods String Srch. Seq. Analy. Video Process. Intr. Detcn. Sys App. Power & Perf. Home

Reg. Overview Pvt. Sector Pub. Sector Govt. Acad. Staff Students Reg. On-line Reg. Accommodation Contact Home

hyPACK 2013 Benefits

Teaching is provided to understand programming in Intel Xeon-Phi Coprocessors using different programming paradigms

Understand tuning & performance of sequential and parallel programs on Intel Xeon-Phi Coprocessors using offload pragmas

Teaches necessary to understand tuning and performance on Multi-Core processors on Linux platforms with Laboratory Session focusing on MPI, OpenMP 3.0, Intel TBB, offload pragmas, Mixed programming

Teaches Open Source Software tools on Distributed Shared Memory (DSMs), Multi-Core Processors and Laboratory Sessions with the help of Industry Experts

Exposure to write codes to measure power consumption in Milliwatts as well as performance of application kernels using Multi-core processors with coprocessors and accelerators

Address Perspective users of Distributed Computing, Parallel Computing, Visual Computing, Language Computing, and Scientific & Engineering Applications on HPC Cluster with coprocessors and accelerators

Offers advanced concepts on Mixed programming - Multi-Core Processors (Pthreads, OpenMP, Intel TBB, MPI) & OpenCL on GPGPUs /GPU Computing Platforms and offload pragmas of Intel Xeon Phi coprocessors

In depth exposure to Heterogeneous Programming OpenCL (NVIDIA & AMD-APP) & Laboratory Sessions.

In depth exposure to Hybrid Programmin based on Intel Xeon Phi Coprocessors (Offload Pragmas, Compilers and Vectorizatrion techinques) in Laboratory Sessions.

Tuning and performance ideas on HPC GPU Cluster (CUDA /OpenCL enabled NVIDIA GPUs, OpenCL on AMD APP with hands-on on numerical & Non-Numerical computations.

Understand HPC Cluster with coprocessors and Acclerators - cluster Management & Health Monitoring

Teaches basics to start programming on Intel Xeon Phi Coprocessors, GPGPU / GPU computing based on NVIDIA - CUDA, AMD-APP SDK, Brook+, and OpenCL with Laboratory Sessions with the help of Industry Experts

Teaches progamming on Partitioned Global Address Space (PGAS) based on UPC, CAF, Titanium and Laboratory Sessions with the help of C-DAC & Industry Experts

Teaches NVIDIA - GPU computational libraries such as CULA tools, CUBLAS, CUFFT, CUSPARSE and CUDPP with Laboratory Sessions

Teaches Hybrid Computing - Mixed Programming - To integrate multi-core software development techniques with GPGPUs to understand performance issues on hybrid computing platforms (HPC Cluster with Coprocessors and acclerators.

Sessions on Practical advice on Performance aspects of Hybrid Computing (HPC GPU Cluster) - Multi-Core Processors and GPUs from Application point of view

A comprehensive technology-training workshop offers basics and advanced concepts on Multi-Core Processors and GPU computing. The hands-on laboratories provide in-depth tuning and performance ideas and increase participants' knowledge and skills.