A new programmer tuition resource for Nvidia's CUDA parallel computing platform has emerged this month with a specific focus on the art of debugging CUDA. Allinea Software's DDT CUDA Education pack is intended for teaching CUDA debugging in academic institutions.
NOTE: As readers will generally know, CUDA is Nvidia's parallel computing platform designed to increase performance via the power of graphics processing units (GPUs).
"Many academic institutions find it difficult to include CUDA debugging into their already substantial curriculum, but acknowledge the need for its presence. The Allinea DDT CUDA Education pack helps meet this need and gives access to best-in-class tools to help academics deliver the best in-class experience for their students," commented Mike Fish, CEO of Allinea Software.
The pack features an advanced debugger for parallel programming via CUDA on Linux. It also contains annual subscriptions to a "classroom-sized set" of CUDA scalar workstation licenses, whitepapers on debugging CUDA, and complete lecture notes suitable for introductory CUDA debugging and hands-on walkthrough examples and exercises.
Monitor and control threads and processes individually and collectively — with process groups.
Intelligent aggregate views automatically compare and contrast program state at all scales. See process and thread anomalies instantly with the parallel stack display — the scalable view that narrows down problems for any level of concurrency, from a handful of processes and threads to hundreds of thousands.
