7. Heterogeneous computing in HSA
• Traditionally computers uses CPUs to process
applications in step by step task
• Computers also uses GPUs which handles
graphics processing and handles large block of
data done in parallel
• Now in order to achieve high performance
engineers started to work with CPUs along
with GPUs in a single chip
8.
9. GPU
• Originally designed for games and graphics
applications
• GPU specializes in many small tasks in
parallel
• Millions of calculations per second
• GPU processors can be used for typical CPU
computations with HSA
• Improved performance in Cryptography ,
Gesture recognition and much more
10.
11.
12. Programming for HSA
• New software design is needed for full
advantage of HSA
• Common technologies to help program for
heterogeneous system
– openCL
– CUDA
– hUMA
15. Real world applications
• Intel CPUs with HD graphics
• AMD APUs
• Smart phones
• Laptops
• Gaming consoles like playstation3 and
playstation4
16. • Benefits
– Small and efficient processor
– Better battery life
– Lower cost
– Increased performance for parallel applications
• Drawbacks
– A discrete CPU and GPU can provide more power
– New software design models
– Less programmers know how to program GPUs
17. Heterogeneous system ERA
• The design and software implementation for
different types of processors on a single chip
• CPU is great for sequential tasks
• GPU is great for small tasks in parallel
• openCL,CUDA and hUMA helps to write
programming for HSA easier
• HSA can improve performance, lower costs
and improve battery life
18. HSA Foundation
• Founded in June 2012
• Developing a new platform for
heterogeneous systems
• www.hsafoundation.com
18