Full Text Search: The Key to Better Natural Language Queries for NoSQL in Node.js
NVIDIA GPUDirect 2.0 Technology -- Offers support for peer-to-peer communication among GPUs within a single server or workstation. This enables easier and faster multi-GPU programming and application performance.
Unified Virtual Addressing (UVA) -- Provides a single merged-memory address space for the main system memory and the GPU memories, enabling quicker and easier parallel programming.
Thrust C++ Template Performance Primitives Libraries -- Provides a collection of powerful open source C++ parallel algorithms and data structures that ease programming for C++ developers. With Thrust, routines such as parallel sorting are 5X to 100X faster than with Standard Template Library (STL) and Threading Building Blocks (TBB).
Unified Virtual Addressing, GPU-to-GPU Communication and Enhanced C++ Template Libraries Enable More Developers to Take Advantage of GPU Computing