Parallel Performance Wizard (PPW) is a performance analysis tool designed for partitioned global-address-space (PGAS) programs, in particular UPC and SHMEM programs. The tool features an easy-to-use interface and tight integration with PGAS programming models via the GASP interface.
Updates:
PPW version 3.2 was released on 10/01/2012. PPW 3.2 is a stable release following 3.0. This release includes new features such as selective measurment for memory transfer to improve scalability. This release also includes support for memory leak detection feature for UPC programs in addition to number of bug fixes and improvements. Get a copy of this version from our download page.
Java Web Start link to test drive the latest version of PPW coming soon.
Sample performance data files are available by choosing Help > Examples from the menu bar.
If the webstart links above do not work for you, you need to install a Java JRE. To do so, please visit the Sun Java website and install version 1.5 of the Java JRE or higher.