Rogue Wave Unveils New Solutions for Code Testing in HPC Environments

By CIOReview | Wednesday, November 4, 2015
776
1334
258

LOUISVILLE, CO: Rogue Wave Software, provider of cross-platform software development tools and embedded components, introduces CodeDynamics to help developers identify and rectify bugs, memory issues, and crashes while implementing complex C and C++ applications. The solution resides on debugger engine in TotalView for high-performance computing (HPC).

CodeDynamics, a purpose-built solution, is mainly designed for multithreaded applications which provide better visibility into running programs, complete control over thread states and an exceptional conceptual view that supports the analysis process. The solution hands over developers the complete power over breakpoints and stepping commands as it speeds up the crash detection process even in large, long-running programs.

“The use for large-scale data modeling and analytics is growing, driving the adoption of technologies traditionally reserved for HPC into new enterprise and industrial applications,” explains Marty Bakal, Principal Product Manager, Rogue Wave.

Rogue Wave’s TotalView keeps a track on everything happening on HPC and supercomputing applications like unprecedented control over processes, thread execution and visibility program states and data. The solution is developed for multicore and parallel computing. Furthermore, TotalView for HPC 8.15.10 will soon be unveiled with support features for Linux PowerLE, NVIDIA CUDA 7.5, improvements for C++11 and additional compiler support .

“The need is the same: extremely fast systems, highly optimized code, and innovative algorithms - all requirements for big data and similar computing solutions in financial, energy, science, government, and other commercial services,” adds Bakal. Being parallel with

TotalView for HPC, CodeDynamics incorporates the same ReplayEngine feature that simplifies troubleshooting features. The engine records and saves executed programs aiding developers to locate the root cause for a crash, debug or error eliminating the need for restarts and stops thus dramatically bringing down the time consumed for troubleshooting. It also records and stores these data for developers to scrutinize crashes anytime.

“With these two products we’re building on our commitment to parallel, multicore computing, while continuing to help developers understand their applications better and isolate faults faster,” comments Bakal.