Alleviating Scalability Limitation of Accelerator-Based Platforms

Scalability of Platforms for Data Warehouse Applications

In the past, academics and commercial vendors have categorised parallel DBMS software architectures on the basis of hardware resource sharing. This approach fails to give insight into the DBMS's performance characteristics. Our alternative approach to categor-ising DBMS architectures is based on the way that processes and threads are organised to cooperate in complex query processing associated...

On the Stochastic Scalability of Information Sharing Platforms

Recently, centrally controlled information distribution systems are rapidly emerging to decentralized structures. This tendency can be observed in software distribution applications using BitTorrent or information sharing platforms based on distributed hash table structures like Chord or Kademlia. To ensure that the emerging platforms will function properly with a growing number of users and se...

Scalability of parallel finite element algorithms on multi-core platforms

The speedup of element-by-element FEM algorithms depends not only on peak processor performance but also on access time to shared mesh data. Eliminating memory boundness would significantly speed up unstructured mesh computations on hybrid multi-core architectures, where the gap between processor and memory performance continues to grow. The speedup can be achieved by ordering unknowns so that ...

Workloads, Scalability, and QoS Considerations in CMP Platforms

Execution Environment Support for Many Core Heterogeneous Accelerator Platforms

We are seeing the advent of large scale, heterogeneous systems comprised of homogeneous general purpose cores intermingled with customized heterogeneous cores and interconnected to diverse memory hierarchies. The presence of accelerators requires support for new programming abstractions and run-time environments that can efficiently harvest platform resources comprised of general purpose and sp...