mirror of
https://github.com/NVIDIA/cuda-samples.git
synced 2024-12-01 15:19:18 +08:00
100 lines
3.1 KiB
XML
100 lines
3.1 KiB
XML
<?xml version="1.0" encoding="UTF-8"?>
|
|
<!DOCTYPE entry SYSTEM "SamplesInfo.dtd">
|
|
<entry>
|
|
<name>conjugateGradientMultiDeviceCG</name>
|
|
<cflags>
|
|
<flag>-ewp</flag>
|
|
<flag>-maxrregcount=64</flag>
|
|
<flag>--std=c++11</flag>
|
|
</cflags>
|
|
<cuda_api_list>
|
|
<toolkit>cudaHostAlloc</toolkit>
|
|
<toolkit>cudaMemPrefetchAsync</toolkit>
|
|
<toolkit>cudaFree</toolkit>
|
|
<toolkit>cudaLaunchCooperativeKernel</toolkit>
|
|
<toolkit>cudaMallocManaged</toolkit>
|
|
<toolkit>cudaSetDevice</toolkit>
|
|
<toolkit>cudaGetDeviceCount</toolkit>
|
|
<toolkit>cudaGetDeviceProperties</toolkit>
|
|
<toolkit>cudaFreeHost</toolkit>
|
|
<toolkit>cudaMemset</toolkit>
|
|
<toolkit>cudaStreamCreate</toolkit>
|
|
<toolkit>cudaStreamSynchronize</toolkit>
|
|
<toolkit>cudaDeviceEnablePeerAccess</toolkit>
|
|
<toolkit>cudaMemAdvise</toolkit>
|
|
<toolkit>cudaOccupancyMaxActiveBlocksPerMultiprocessor</toolkit>
|
|
<toolkit>cudaDeviceCanAccessPeer</toolkit>
|
|
</cuda_api_list>
|
|
<description><![CDATA[This sample implements a conjugate gradient solver on multiple GPUs using Multi Device Cooperative Groups, also uses Unified Memory optimized using prefetching and usage hints.]]></description>
|
|
<devicecompilation>whole</devicecompilation>
|
|
<includepaths>
|
|
<path>./</path>
|
|
<path>../</path>
|
|
<path>../../../Common</path>
|
|
</includepaths>
|
|
<keyconcepts>
|
|
<concept level="basic">Unified Memory</concept>
|
|
<concept level="advanced">Linear Algebra</concept>
|
|
<concept level="basic">Cooperative Groups</concept>
|
|
<concept level="advanced">MultiDevice Cooperative Groups</concept>
|
|
<concept level="advanced">CUBLAS Library</concept>
|
|
<concept level="advanced">CUSPARSE Library</concept>
|
|
</keyconcepts>
|
|
<keywords>
|
|
<keyword>CUDA</keyword>
|
|
<keyword>Sparse Matrix</keyword>
|
|
<keyword>Unified Memory</keyword>
|
|
<keyword>Multi-GPU</keyword>
|
|
<keyword>CPP11</keyword>
|
|
</keywords>
|
|
<libraries>
|
|
<library>cudadevrt</library>
|
|
</libraries>
|
|
<librarypaths>
|
|
</librarypaths>
|
|
<nsight_eclipse>true</nsight_eclipse>
|
|
<primary_file>conjugateGradientMultiDeviceCG.cu</primary_file>
|
|
<required_dependencies>
|
|
<dependency>UVM</dependency>
|
|
<dependency>MDCG</dependency>
|
|
<dependency>CPP11</dependency>
|
|
</required_dependencies>
|
|
<scopes>
|
|
<scope>1:CUDA Advanced Topics</scope>
|
|
<scope>3:Linear Algebra</scope>
|
|
</scopes>
|
|
<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>
|
|
<sm-arch>sm90</sm-arch>
|
|
<supported_envs>
|
|
<env>
|
|
<arch>x86_64</arch>
|
|
<platform>linux</platform>
|
|
</env>
|
|
<env>
|
|
<arch>ppc64le</arch>
|
|
<platform>linux</platform>
|
|
</env>
|
|
<env>
|
|
<platform>windows</platform>
|
|
</env>
|
|
<env>
|
|
<platform>aarch64</platform>
|
|
</env>
|
|
<env>
|
|
<platform>sbsa</platform>
|
|
</env>
|
|
</supported_envs>
|
|
<supported_sm_architectures>
|
|
<from>6.0</from>
|
|
</supported_sm_architectures>
|
|
<title>conjugateGradient using MultiDevice Cooperative Groups</title>
|
|
<type>exe</type>
|
|
</entry>
|