08.00h – 9.00h | REGISTRATION WITH BREAKFAST
PARALLELISM, PERFORMANCE & OPTIMIZATION ON INTEL® ARCHITECTURE – WHAT YOU SHOULD KNOW!
The Intel® architecture offers parallelism at many different levels. Here we present the seven different levels of parallelism you should be thinking about when you write or optimize code.
SESSION 1: A STEP-BY-STEP APPROACH TO APPLICATION TUNING WITH INTEL® COMPILERS
In this first hands-on session, we follow the steps recommended by the Intel® Engineers that should be used to get the best performance out of the Intel® Compiler. Practical topics include general optimization, processor-specific options, inter procedural optimization, profile guided optimization, compiler optimization reports, and support for thread parallelism.
SESSION 2: INTEGRATING VTUNE AND ADVISOR IN YOUR OPTIMIZATION PROCESS
Starting from an un-optimized wave propagation kernel, we describe a general methodology to characterize and profile your code. We will see how threading, vectorization, but also memory optimization can affect the behaviour of your programs. The roofline model, dependency analysis and memory access pattern analysis from Intel® Advisor as well as the HPC Performance Characterization and Memory Access from Intel® VTuneTM Amplifier will be used to guide the developer to a more performant code.
18.00h | NETWORKING EVENING WITH DRINKS & FOOD
Leave a Reply
Want to join the discussion?Feel free to contribute!