A Task-Centric Memory Model for Scalable Accelerator Architectures
John H. Kelm, Daniel R. Johnson, Steven S. Lumetta et al. · IEEE Micro · 2010
This article presents a memory model for parallel compute accelerators with task-based programming models that uses a software protocol, working in collaboration with hardware caches, to maintain a coherent, single address space view of memory without requiring hardware cache coherence. The memory model supports visual computing applications, which are becoming an important class of workloads capable of exploiting 1,000-core processors.
Cited by 5 works. doi:10.1109/mm.2010.6
Authors
References
- A Bridging Model for Parallel Computation
- Scalable parallel programming with CUDA
- Scalable Parallel Programming with CUDA
- NVIDIA Tesla: A Unified Graphics and Computing Architecture
- The implementation of the Cilk-5 multithreaded language
- Larrabee
- TreadMarks: shared memory computing on networks of workstations
- Intel threading building blocks : outfitting C++ for multi-core processor parallelism
- The Midway distributed shared memory system
- Cache Consistency and Sequential Consistency