![]() "The new features of Cuda 4.0 are designed for the programmer who doesn't want to dive into some of the details that were required for the previous releases," Russell said. ![]() I have been able to remove gnome but I am unsure, if this is a proper removal so I leave the question open for now. The gnome comes from the cuda toolkit I guess. Routines like parallel sorting are to 5 to 100 times faster with Standard Template Library and Threading Building Blocks, which is Intel's C++ template library. Using the headless nvidia Driver did not help. Developers can coordinate work across multiple GPUs.Īlso, Thrust C++ Template Performance Primitives Libraries in version 4.0 provide open source C++ parallel algorithms and data structures to make it easier to program in C++, Nvidia said. In addition, multi-GPU sharing by a single CPU thread is enabled, letting a single CPU host thread access all GPUs in a system. This makes it easier for multithreaded applications to share a single GPU. Version 4.0 enables multithread sharing of GPUs in which multiple CPU host threads can share contexts on a single GPU. Also, a Unified Virtual Addressing capability offers a merged-memory address space for the main system memory and GPU memories, for easier parallel programming, Nvidia said. "You can now have direct transfers between the GPUs," said Sanford Russell, director of Cuda marketing at Nvidia. Among the key features in version 4.0 is the company's GPUDirect 2.0 technology, which supports peer-to-peer communications among GPUs within a single processor or workstation, enabling faster multi-GPU programming and application performance, the company said.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |