Rogue Wave have announced the latest version of TotalView (8.13), its parallel debugging software for HPC systems. The scalable architecture allows debugging on a large scale; it has shown increased performance on Cray and IBM BlueGene supercomputers, as well as on Linux clusters.
TotalView is scalable debugging software for parallel applications written in C, C++, and Fortran, designed to improve developer productivity. TotalView provides a combination of capabilities for pinpointing and fixing hard-to-find bugs, such as race conditions, memory leaks, and memory overruns.
TotalView 8.13 adds support for Cuda 5.0 and 5.5 including dynamic parallelism capabilities. This expanded support allows TotalView to work with Nvidia’s latest hardware offering, the Kepler K40 GPU accelerator.
The program also expands support for debugging on Intel Xeon Phi coprocessors, for the first time, users can debug applications running in Xeon Phi symmetric mode with seamless transitions when the code shifts between processor and coprocessor. TotalView’s MemoryScape is the first debugging solution available for Xeon Phi applications.
