Dimension scores are derived from public data and fields; weighted into the composite. Reference only.
dispy is a distributed and parallel computing framework for Python, designed to run compute tasks on a single multi-core machine, a LAN cluster, a grid, or cloud servers. Its core use case is data-parallel/SIMD workloads: applying the same Python function or standalone program to different datasets, with little to no communication between tasks. If inter-task communication is required, the documentation recommends using pycos Distributed Communicating Processes.
dispy consists of components such as client, dispynode, dispyscheduler, dispynetrelay, and dispyadmin. Developers can use JobCluster for exclusive node scheduling, or use SharedJobCluster together with dispyscheduler to let multiple clients share a cluster. dispy can automatically distribute dependencies such as Python functions, classes, modules, and files, and return results, stdout, stderr, and exception traces. It also supports intermediate results, transferring files back from nodes to the client, job status notifications, cluster status notifications, and fault recovery. When a node fails, if the computation is declared reentrant, tasks can be automatically resubmitted to other nodes.
dispy supports Python 2.7+ and Python 3.1+, and has been tested on Linux, OS X, and Windows. The documentation also mentions PyPy support. It depends on pycos, while some resource monitoring and networking features may rely on psutil, netifaces, pywin32, and related packages. Deployment is largely self-hosted: each compute node runs dispynode; dispyscheduler is needed for shared scheduling; and dispynetrelay can be used across networks. It can also be used with Amazon EC2, Google Cloud, and Microsoft Azure, and supports Docker containers for isolating compute environments.
The main documentation does not mention commercial pricing. The project can be installed via PyPI, its source code is hosted on GitHub, and it can also be downloaded from SourceForge, making it clearly oriented as an open-source tool. The documentation is relatively strong, covering installation, dependencies, API parameters, NAT/firewall setups, SSH forwarding, SSL, cloud computing, monitoring and administration, plus many examples. That said, the documentation is information-dense, and the Python version and source installation instructions are somewhat complex, so new users may need time to understand how the components fit together.
Its strengths are that it is Python-native, flexible to deploy, and able to turn existing machines into a private compute cluster quickly. It also includes practical engineering features such as result collection, status notifications, web-based administration, and fault recovery. Its weaknesses are that it is better suited to independent batch-processing tasks, and less suitable for workloads requiring frequent communication or complex DAG scheduling. The multi-component architecture also adds operational overhead. dispy is a good fit for scientific computing, parameter sweeps, batch data processing, and making use of private cluster compute capacity. The source text provides no information about accessibility from China; access to PyPI, GitHub, and SourceForge may vary depending on the network environment. Teams looking for a more modern ecosystem may also want to evaluate Ray, Dask, Spark, or Celery.
⚠ This review is compiled from public sources and does not constitute a purchase recommendation. Verify all facts on the vendor's official site. Verify on dispy.org official site.
dispy.org is an United States Dev Tools provider. TG4G tracks its product information, an overall rating of 7.0/10, and a China-accessibility score of China direct-connect friendly. Click "Visit Official Site" to reach dispy.org directly.