Evaluating Mobile Graphics Processing Units (GPUs) for Real-Time Resource Constrained Applications
Modern graphics processing units (GPUs) can provide tremendous performance boosts for some applications beyond what a single CPU can accomplish, and their performance is growing at a rate faster than CPUs as well. Mobile GPUs available for laptops have the small form factor and low power requirements suitable for use in embedded processing. We evaluated several desktop and mobile GPUs and CPUs on traditional and non-traditional graphics tasks, as well as on the most time consuming pieces of a full hyperspectral imaging application. Accuracy remained high despite small differences in arithmetic operations like rounding. Performance improvements are summarized here relative to a desktop Pentium 4 CPU.
- Research Organization:
- Lawrence Livermore National Laboratory (LLNL), Livermore, CA
- Sponsoring Organization:
- USDOE
- DOE Contract Number:
- W-7405-ENG-48
- OSTI ID:
- 888600
- Report Number(s):
- UCRL-TR-217422
- Country of Publication:
- United States
- Language:
- English
Similar Records
Iterative methods in GPU-resident linear solvers for nonlinear constrained optimization
Flocking-based Document Clustering on the Graphics Processing Unit