ac6-training, un département d'Ac6 SAS
 
Site displayed in English (GB)
Site affiché en English (GB)View the site in FrenchVoir le site en English (USA)
go-up

ac6 >> ac6-training >> Programming >> Languages >> OpenCL Inquire Download as PDF Write us

L9 OpenCL

Parallel programming with OpenCL-1.2

formateur
Objectives
  • Learn parallel programming with OpenCL.
  • Know what (not) to expect from parallel programming.
  • Understand heavy multithreading and how it is mapped to the hardware.
  • Measure OpenCL code performance, locate and solve bottlenecks.
  • Write efficient OpenCL code.
Depending on the hardware environment, exercises will be run on either multi-core CPUs, nVidia or ATI GPUs.
  • Theoretical course
    • PDF course material (in English) supplemented by a printed version.
    • The trainer answers trainees' questions during the training and provide technical and pedagogical assistance.
  • Practical activities
    • Practical activities represent from 40% to 50% of course duration.
    • Code examples, exercises and solutions
    • One PC (Linux ou Windows) for the practical activities with, if appropriate, a target board.
      • One PC for two trainees when there are more than 6 trainees.
    • For onsite trainings:
      • An installation and test manual is provided to allow preinstallation of the needed software.
      • The trainer come with target boards if needed during the practical activities (and bring them back at the end of the course).
  • Downloadable preconfigured virtual machine for post-course practical activities
  • At the start of each session the trainer will interact with the trainees to ensure the course fits their expectations and correct if needed
  • Good knowledge of the C language
  • Any embedded systems engineer or technician with the above prerequisites.
  • The prerequisites indicated above are assessed before the training by the technical supervision of the traineein his company, or by the trainee himself in the exceptional case of an individual trainee.
  • Trainee progress is assessed in two different ways, depending on the course:
    • For courses lending themselves to practical exercises, the results of the exercises are checked by the trainer while, if necessary, helping trainees to carry them out by providing additional details.
    • Quizzes are offered at the end of sections that do not include practical exercises to verifythat the trainees have assimilated the points presented
  • At the end of the training, each trainee receives a certificate attesting that they have successfully completed the course.
    • In the event of a problem, discovered during the course, due to a lack of prerequisites by the trainee a different or additional training is offered to them, generally to reinforce their prerequisites,in agreement with their company manager if applicable.

Course Outline

  • History
    • OpenCL 1.0
    • OpenCL 1.1
    • OpenCL 1.2
    • OpenCP/EP (Embedded Profile)
  • Design goals of OpenCL
    • CPUs, GPUs and GPGPUs
    • Data-parallel and Task-parallel
    • Hardware related and portable
  • Terminology
    • Host / Device
    • Memory model
    • Execution Model
  • The OpenCL Architecture
    • Platform Model
    • Execution Model
    • Memory Model
    • Programming Model
  • The OpenCL Software Stack
  • Example
Exercise:  Installation and test of the OpenCL SDK
  • Platform layer
    • Querying and selecting devices
    • Managing compute devices
    • Managing computing contexts and queues
    • The host objects: program, kernel, buffer, image
Exercise:  Write a platform discovery and analysis program (displaying CPUs, GPUs, versions...)
  • Runtime
    • Managing resources
    • Managing memory domains
    • Executing compute kernels
Exercise:  Write an image loader program, transferring image to/from compute devices
  • Compiler
    • The OpenCL C programming language
    • Online compilation
    • Offline compilation
  • How code is executed on hardware
    • Compute kernel
    • Compute program
    • Application queues
  • OpenCL Data-parallel execution
    • N-dimensional computation domains
    • Work-items and work-groups
    • Synchronization and communication in a work-group
    • Mapping global work size to work-groups
    • Parallel execution of work-groups
Exercise:  Compile and execute a program to square an array on the platform computing nodes
  • Restrictions from C99
  • Data types
    • Scalar
    • Vector
    • Structs and pointers
    • Type-conversion functions
    • Image types
Exercise:  Rewrite the square program to use vector operations
  • Required built-in functions
    • Work-item functions
    • Math and relational
    • Input/output
    • Geometric functions
    • Synchronization
  • Optional features
    • Atomics
    • Rounding modes
Exercise:  Write and execute an image manipulation program (Blur filter)
  • Profiling
Exercise:  Enhance the image manipulation program to measure kernel computation time
  • The OpenCL Memory Model
    • Global Memory
    • Local Memory
    • Private Memory
  • OpenCL Task-parallel execution
    • Optional OpenCL feature
    • Native work-items
Exercise:  Simulate the N-Body problem, displaying data using OpenGL
  • When (not) to use OpenCL
  • Code design guidelines
  • Explicit vectorization
Exercise:  Explore vectorisation on an image rotation kernel
  • Memory latency and access patterns
    • ALU latency
    • Using local memory
Exercise:  Enhance the Blur filter program to investigate memory optimisations
  • Synchronizing threads
  • Warps/Wavefronts, work groups, and GPU cores