2012-12-06 Node-Level Performance Engineering (HNPF1W12)

Date:	Dec 6, 2012 10:00 ? 18:00 Dec 7, 2012 09:00 ? 17:00
Location:	LRZ Building, University campus Garching, near Munich Boltzmannstr. 1 Hörsaal H.E.009
Contents:	This course teaches performance engineering approaches on the compute node level. ?Performance engineering? as we define it is more than employing tools to identify hotspots and bottlenecks. It is about developing a thorough understanding of the interactions between software and hardware. This process must start at the core, socket, and node level, where the code gets executed that does the actual computational work. Once the architectural requirements of a code are understood and correlated with performance measurements, the potential benefit of optimizations can often be predicted. We introduce a ?holistic? node-level performance engineering strategy, apply it to different algorithms from computational science, and also show how an awareness of the performance features of an application may lead to notable reductions in power consumption. Introduction Intel and AMD x86 architectures ccNUMA Performance modeling & engineering approaches Our Approach Practical performance analysis The LIKWID tools Typical performance patterns Microbenchmarks and the memory hierarchy Understanding the memory hierarchy Data transfer between memory levels Write allocate vs. NT stores Modeling of cache hierarchies Contention NUMA effects ? anisotropy and asymmetry Typical node-level software overheads Cost of synchronization Work distribution Example Problem: The 3D Jacobi solver Core-level optimizations Blocking Non Temporal stores SIMD vectorization (SSE, AVX) Multithreading ? contention at different memory hierarchies Temporal Blocking Example Problem: The Lattice-Boltzmann Method (LBM) Introduction Roofline Model Data layout Non Temporal stores Model for in-cache data & multicore scaling Sparse representation and options for propagation Example Problem: Sparse Matrix-Vector Multiplication Data layouts Performance model ? CPU vs. GPU Bandwidth reduction Example Problem: A backprojection algorithm for CT reconstruction The algorithm Naïve analysis Detailed analysis and performance model Optimizations Energy & Parallel Scalability Energy consumption of modern processors The energy-to-solution metric Performance engineering == power engineering Case studies Between each module, there is time for Questions and Answers!
Prerequisites	Participants must have basic knowledge in programming with Fortran or C
Language:	English
Teacher:	Prof. Gerhard Wellen/RRZE, Dr. Georg Hager/RRZE et. al.
Registration:	Please register via the LRZ registration form (Please choose course HNPF1W12)