Name
Title | ||
|---|---|---|
A compute unified system architecture for graphics clusters incorporating data locality. |
Abstract | ||
|---|---|---|
We present a development environment for distributed GPU computing targeted for multi-GPU systems, as well as graphics clusters. Our system is based on CUDA and logically extends its parallel programming model for graphics processors to higher levels of parallelism, namely, the PCI bus and network interconnects. While the extended API mimics the full function set of current graphics hardware-including the concept of global memory-on all distribution layers, the underlying communication mechanisms are handled transparently for the application developer. To allow for high scalability, in particular for network-interconnected environments, we introduce an automatic GPU-accelerated scheduling mechanism that is aware of data locality. This way, the overall amount of transmitted data can be heavily reduced, which leads to better GPU utilization and faster execution. We evaluate the performance and scalability of our system for bus and especially network-level parallelism on typical multi-GPU systems and graphics clusters. |
Year | DOI | Venue |
|---|---|---|
2009 | 10.1109/TVCG.2008.188 | IEEE Trans. Vis. Comput. Graph. |
Keywords | Field | DocType |
gpu computing,multi-gpu system,network-level parallelism,data locality,gpu utilization,compute unified system architecture,graphics cluster,high scalability,graphics clusters incorporating data,graphics processor,pci bus,current graphics hardware,computer architecture,hardware architecture,parallel processing,concurrent programming,coprocessors,power generation,scalability,system architecture,programming language,parallel programming,computer graphic,hardware,distributed computing,computer graphics,parallel programming model,application development,software engineering,graphics hardware,concurrent computing,graphics,development environment | Graphics,Graphics hardware,GPU cluster,Computer science,CUDA,Parallel computing,Theoretical computer science,General-purpose computing on graphics processing units,Coprocessor,Computer graphics,Scalability | Journal |
Volume | Issue | ISSN |
15 | 4 | 1077-2626 |
Citations | PageRank | References |
16 | 1.61 | 17 |
Authors | ||
5 | ||
Name | Order | Citations | PageRank |
|---|---|---|---|
| Christoph Müller | 1 | 147 | 12.84 |
| Steffen Frey | 2 | 116 | 18.93 |
| Magnus Strengert | 3 | 221 | 13.43 |
| Carsten Dachsbacher | 4 | 1396 | 93.20 |
| Thomas Ertl | 5 | 4417 | 401.52 |