Mapping Iterative Medical Imaging Algorithm on Cell Accelerator
2011
Mapping Medical Imaging Algorithms on Cell Accelerator
publication
Evidence: moderate
Author Information
Author(s): Xu Meilian, Thulasiraman Parimala
Primary Institution: University of Manitoba
Hypothesis
Can the OS-SART algorithm be effectively mapped on the Cell Broadband Engine to improve performance in medical imaging?
Conclusion
The proposed OS-SART algorithm on the Cell BE outperforms traditional methods, achieving significant speed improvements.
Supporting Evidence
- The Cell BE is five times faster than the AMD Opteron dual-core processor.
- The speedup of the algorithm on Cell BE increases with the number of SPEs.
- The proposed algorithm significantly reduces the computation time compared to traditional methods.
Takeaway
This study shows how a special computer chip can make medical imaging faster and better by using smart algorithms.
Methodology
The study involved mapping the OS-SART algorithm onto the Cell Broadband Engine and comparing its performance against traditional CPU methods.
Limitations
The limited memory storage on each SPE of the Cell BE can be a bottleneck for performance.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website