cuda-samples/Samples/6_Performance/UnifiedMemoryPerf/NsightEclipse.xml

<?xml version="1.0" encoding="UTF-8"?> 
<!DOCTYPE entry SYSTEM "SamplesInfo.dtd">
<entry>
  <name>UnifiedMemoryPerf</name>
  <cuda_api_list>
    <toolkit>cudaStreamDestroy</toolkit>
    <toolkit>cudaFree</toolkit>
    <toolkit>cudaMallocHost</toolkit>
    <toolkit>cudaMallocManaged</toolkit>
    <toolkit>cudaMemPrefetchAsync</toolkit>
    <toolkit>cudaStreamCreate</toolkit>
    <toolkit>cudaStreamAttachMemAsync</toolkit>
    <toolkit>cudaFreeHost</toolkit>
    <toolkit>cudaMalloc</toolkit>
    <toolkit>cudaMemcpyAsync</toolkit>
    <toolkit>cudaStreamSynchronize</toolkit>
    <toolkit>cudaHostGetDevicePointer</toolkit>
    <toolkit>cudaMemcpy</toolkit>
    <toolkit>cudaGetDeviceProperties</toolkit>
  </cuda_api_list>
  <description><![CDATA[This sample demonstrates the performance comparision using matrix multiplication kernel of Unified Memory with/without hints and other types of memory like zero copy buffers, pageable, pagelocked memory performing synchronous and Asynchronous transfers on a single GPU.]]></description>
  <devicecompilation>whole</devicecompilation>
  <includepaths>
    <path>./</path>
    <path>../</path>
    <path>../../../Common</path>
  </includepaths>
  <keyconcepts>
    <concept level="basic">CUDA Systems Integration</concept>
    <concept level="basic">Unified Memory</concept>
    <concept level="basic">CUDA Streams and Events</concept>
    <concept level="basic">Pinned System Paged Memory</concept>
  </keyconcepts>
  <keywords>
    <keyword>CUDA</keyword>
    <keyword>Unified Memory</keyword>
    <keyword>Pinned Memory</keyword>
    <keyword>Zero copy buffer</keyword>
    <keyword>UVM</keyword>
    <keyword>Streams</keyword>
  </keywords>
  <libraries>
  </libraries>
  <librarypaths>
  </librarypaths>
  <nsight_eclipse>true</nsight_eclipse>
  <primary_file>matrixMultiplyPerf.cu</primary_file>
  <required_dependencies>
    <dependency>UVM</dependency>
  </required_dependencies>
  <scopes>
    <scope>1:CUDA Basic Topics</scope>
    <scope>1:CUDA Systems Integration</scope>
    <scope>1:Unified Memory</scope>
  </scopes>
  <sm-arch>sm35</sm-arch>
  <sm-arch>sm37</sm-arch>
  <sm-arch>sm50</sm-arch>
  <sm-arch>sm52</sm-arch>
  <sm-arch>sm53</sm-arch>
  <sm-arch>sm60</sm-arch>
  <sm-arch>sm61</sm-arch>
  <sm-arch>sm70</sm-arch>
  <sm-arch>sm72</sm-arch>
  <sm-arch>sm75</sm-arch>
  <sm-arch>sm80</sm-arch>
  <sm-arch>sm86</sm-arch>
  <sm-arch>sm87</sm-arch>
  <supported_envs>
    <env>
      <arch>x86_64</arch>
      <platform>linux</platform>
    </env>
    <env>
      <arch>x86_64</arch>
      <platform>macosx</platform>
    </env>
    <env>
      <platform>windows7</platform>
    </env>
    <env>
      <arch>arm</arch>
    </env>
    <env>
      <arch>sbsa</arch>
    </env>
    <env>
      <arch>aarch64</arch>
    </env>
    <env>
      <arch>ppc64le</arch>
      <platform>linux</platform>
    </env>
  </supported_envs>
  <supported_sm_architectures>
    <from>3.5</from>
  </supported_sm_architectures>
  <title>Unified and other CUDA Memories Performance</title>
  <type>exe</type>
</entry>