2 <name>CUDA Parallel Prefix Sum with Shuffle Intrinsics (SHFL_Scan)</name>
3 <description><![CDATA[This example demonstrates how to use the shuffle intrinsic __shfl_up to perform a scan operation across a thread block. A GPU with Compute Capability SM 3.0. is required to run the sample]]></description>
4 <whitepaper></whitepaper>
5 <minspec>KEPLER SM 3.0</minspec>
6 <keywords>GPGPU, CUDA, scan, parallel prefix sum,Data-Parallel Algorithms</keywords>
7 <scope>1:CUDA Advanced Topics,1:Data-Parallel Algorithms,1:Performance Strategies</scope>
11 <keyconcepts level="advanced">Data-Parallel Algorithms</keyconcepts>
12 <keyconcepts level="advanced">Performance Strategies</keyconcepts>
14 <!-- *** Additional project definition information *** -->
15 <!-- Libraries that the project should be linked against. Document may contain any number of "libraries" elements. -->
16 <!-- Attributes "os" and "arch" are optional and may be used in any combination. -->
17 <!-- "framework" attribute will be ignored on platforms other then Mac OS X -->
19 <!-- Either absolute or relative to sample -->
20 <library-path os="windows"></library-path>
22 <!-- Either absolute or relative to sample -->
23 <include-path>./</include-path>
24 <include-path>../</include-path>
25 <include-path>../../common/inc</include-path>
27 <!-- Can be either "separate" or "whole", omitting this element means that sample can be compiled either way -->
28 <devicecompilation>whole</devicecompilation>
30 <!-- These options will be passed to NVCC compiler as is and for all files -->
32 <nvcc-compiler file="shfl_scan.cu"></nvcc-compiler>
33 <sm-arch>sm30</sm-arch>
35 <!-- One of "exe|dll|lib" (exe is default) -->
38 <!-- By default IDE will open file "samplename.(cu|cpp) -->
39 <primary-file>shfl_scan.cu</primary-file>
41 <!-- Collection of [OS][:arch] tokens -->
42 <supported-env>linux:x86_64, linux:i686, windows7, macosx:x86_64, macosx:i686, :arm</supported-env>