| |
The ACCELLERANT SDKs (Software
Developers’ Kit) includes
algorithm-specific APIs, random number generation functions,
pre-processor services, utilities,
data
structuring utilities and data, object and structure
serialization services and partitioning algorithms,
templates, sample code and Best Practices guidelines.
These make it easy to adapt
or wrap existing applications to
leverage multi-CPUs and multiples systems in any work
station, server, cluster or grid environment as well as to
develop new applications without the need to focus on the
complexities of multi-threading and distributed processing
or dealing with grid infrastructure interfaces.
Each SDK is targeted to
enable the Architect, Designer, Programmer, Engineer, or
Administrator to focus on the application logic and design
with the assurance that the underlying subtleties of the
target execution environments will be efficiently an
dynamically handled by the APIs and the ACCELLERANT run-time
functionality.
The overall emphasis of the APIs are such
that the application can be designed, adapted or wrapped in a straight forward
manner without regards to the potential nuances and
complexities that a target parallel/distributed the
environment could present.
The application of ASPEED's APIs
and run-time engine ensure optimum performance every time
the application is run and insulates the the application
from the underlying hardware configuration, cluster or grid
infrastructure. This enables application developers
to focus their efforts on the business flow and logic while
significantly reducing the new application development time
and eliminating the need to re-engineer existing application
logic.
The SDK's include the tools required to identify
the relevant data objects, the serialization preprocessor, random number
generation models, and the application APIs. The APIs are used to:
-
identify the
processing assumption to apply to the ACCELLERANT
distribution algorithm (Independent, Mesh, Matrix, Tree,
Implicit and explicit dependent, Non-Linear,
etc.)
-
identify the data structures and objects for serialization
-
synchronize and
aggregation
as well as APIs for providing advance operational
control if desired.
|
Overview
