Memory Optimization (MemOp)
Abstract
The Memory Optimization (MemOp) program developed technology to optimize memory transactions in large scale computing systems. The demand for computing services is growing within both the U.S. Government and commercial industry. In response, new technical approaches were developed to provide massive computation efficiently and cost effectively. In particular, distributed data centers with high-speed interconnects and customizable hardware, including graphics processing units (GPU) and field programmable gate arrays (FPGAs), are being used by service providers to achieve greater efficiency and improved processing performance. MemOp explored new memory architectures that more fully leverage emerging customizable hardware to deliver computing services reliably and at reduced cost. The more promising MemOp memory architectures were implemented and evaluated in hardware and software. The technologies developed in MemOp provide enhanced efficiency and improved performance for large scale computing systems.
Document Details
- Document Type
- Accomplishment
- Publication Date
- Oct 01, 2025
- Source ID
- 0e703031de978c997b6ce7a9f105e2a4