Archive notice: This article was originally published on January 30, 2012. Links and embedded videos are preserved as part of the historical record.
The Cuda tool is said to be essential when it comes to numerical work with GPUs. An LLVM compiler was integrated for the first time, which is intended to be specially designed for parallel GPU calculations and thus make CUDA accessible to new programming languages and alternative processor architectures.
The following enhancements in the update are intended to simplify parallel programming for GPU calculations: according to Nvidia, the Visual Profiler has been redesigned and now contains a new automated system for performance improvements carried out step by step. In addition, around 1000 new signal-processing functions are said to have been added to the NPP library for image processing under Nvidia GPUs.
The free update is available here.
About the author
Redakteurin der Digital Production
Article details
- Published
- Author
- Mirja Fürst
- Reading time
- About 1 min
- Word count
- 140