Understanding throughput-oriented architectures
Michael Garland, David B. Kirk · Communications of the ACM · 2010
For workloads with abundant parallelism, GPUs deliver higher peak computational throughput than latency-oriented CPUs.
Cited by 154 works. doi:10.1145/1839676.1839694
Authors
References
- Very high-speed computing systems
- The CRAY-1 computer system
- Simultaneous multithreading
- Scalable parallel programming with CUDA
- Scalable Parallel Programming with CUDA
- NVIDIA Tesla: A Unified Graphics and Computing Architecture
- Niagara: A 32-Way Multithreaded Sparc Processor
- The Tera computer system
- CUDA by Example: An Introduction to General-Purpose GPU Programming
- The Illiac IV system