SWARM Case Studies
SWARM advantages over OpenMP
• Better load balancing
• Faster synchronization
• Expresses more parallelism classes
• Scalable across clusters
Embarrassingly Parallel Problems:
Even on embarrassingly parallel problems like Mandelbrot set generation, SWARM outperforms OpenMP by providing better load balancing and faster synchronization. SWARM can also parallelize across nodes without adding another programming model (e.g., MPI).
Irregular Applications:
For more complex and irregular applications such as N-body simulations (Barnes-Hut), SWARM outperforms OpenMP by significant margins.
SWARM advantages over MPI
• Improved resource utilization
• Reduced latency effects
• Lower type overhead
• Scalable across shared memory
ETI has ported the Graph 500 reference MPI implementation to SWARM and produced superior results with speed-up from 2-fold to 15-fold. For more details, see ETI's Graph 500 implementation with SWARM
For more information on SWARM, see:
White Paper NEW!
SWARM: Scalable Performance Optimization for Multi-core/Multi-node
SWARM beta
Register to download free BETA software
Contact ETI
Want to learn more? We can help.
Site Search
Find what you need on our site, quickly & easily
