NVIDIA_CUDA_ProgrammingGuideRevisionHistory -...

Info iconThis preview shows pages 1–3. Sign up to view the full content.

View Full Document Right Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: --------------------------------------------------------------------------------NVIDIA CUDA Programming GuideRevision History----------------------------------------------------------------------------------------------------------------------------------------------------------------Version 2.3--------------------------------------------------------------------------------- Section 4.2- Updated now that all GPUs in SLI mode are accessible via the CUDA driverand runtime as separate devices- Section Warned about mis-aligned reads producing incorrect results - New appendix D on C++ language constructs supported in device code--------------------------------------------------------------------------------Version 2.3 Beta--------------------------------------------------------------------------------- Section 3.2.10- Mentioned CUDA-GDB- Section B.7- Listed functions affected by -use_fast_math- Section C.2.1- Added __float2ll and __float2ull--------------------------------------------------------------------------------Version 2.2.1--------------------------------------------------------------------------------- Modified all driver API code samples to use the correct way of passingdevice pointers as parameters to kernel launches- Section 3.2.3- Mentioned cudaThreadExit()- Section 3.2.5- Mentioned that the simple zero-copy SDK sample comes with a detaileddocument on the page-locked memory APIs- Section 3.3- Updated to recommend loading ptx for applications that want to run onfuture device architectures- Sections A.1- Added more GPUs- Sections A.1.1- Mentioned the 16K limit for local memory--------------------------------------------------------------------------------Version 2.2--------------------------------------------------------------------------------- Sections and 3.3.8- Mentioned how to check support for mapped page-locked host memory- Section 3.2.7- Mentioned cudaWGLGetDevice() and Quadro's high performance OpenGLinteroperability in multi-device configurations- New sections 3.3.12 and 3.3.12 on error handling.- New section 3.4 on versioning and compatibility- New section 3.5 on compute modes- Section Throughput now expressed in instructions per clock cycles- Expanded on slow and fast paths of argument reduction code fortrigonometric functions- Section A.2- Mentioned single-precision IEEE-compliant software implementations--------------------------------------------------------------------------------Version 2.2 Beta--------------------------------------------------------------------------------- Chapter on programming interface reorganized and rewritten with more codesamples- Chapter on hardware implementation moved to after chapter on programminginterface- Section 3.2.4 and B.8 modified to reflect new support for texturing fromlinear memory (2D textures, texture filtering, addressing modes, andnormalized texture coordinates)- Section 3.2.5- Updated with new page-locked host memory functions and features- Section 3.3.3- Section 3....
View Full Document

This note was uploaded on 02/09/2010 for the course CS 152 taught by Professor Staff during the Spring '98 term at UCLA.

Page1 / 11

NVIDIA_CUDA_ProgrammingGuideRevisionHistory -...

This preview shows document pages 1 - 3. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online