Register | Login

Hi,    |  Logout




In a typical eCAD grid environment, clusters participate in a variety of workloads deployed in a multitude of sizes causing unpredictable storage performance and data confusion across the networks. For years there has been no real solution available to move data into the appropriate cluster in a just-in-time fashion. Today users use difficult to maintain scripts to accomplish this task. Using Perfaccel, data is intelligently cached on the cluster member servers allocated to a specific workload. IO latency is dramatically reduced while network storage IOPS are offloaded to the cache. With tools like grid management, grid level analytics, real-time graphs, and deep IO analytics, applications supported by PerfAccel enjoy data on demand with data locality dramatically improving storage performance within their network.


How does it work? With PerfAccel, there is no need to setup synchronization jobs between source code servers and grid nodes. Users can compile data on the NFS share while data is automatically refreshed on the compute node. Managing the data might seem complex – but with PerfAccel, managing data across a grid of thousands of nodes and storage devices has never been easier. As an application, OS, and storage agnostic technology, PerfAccel works within the eCAD environment in complete anonymity, causing zero disruption to existing jobs. From acceleration in EDA applications to grid level scalability, PerfAccel delivers powerful IO management infrastructure to the eCAD community in an easy-to-use system that deploys in minutes.