📄 lmbench.html
字号:
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"><html><head> <title></title> <link rel="stylesheet" media="screen" type="text/css" href="./style.css" /> <link rel="stylesheet" media="screen" type="text/css" href="./design.css" /> <link rel="stylesheet" media="print" type="text/css" href="./print.css" /> <meta http-equiv="Content-Type" content="text/html; charset=utf-8" /></head><body><a href=start.html>start</a></br><div class="toc"><div class="tocheader toctoggle" id="toc__header">Table of Contents</div><div id="toc__inside"><ul class="toc"><li class="clear"><ul class="toc"><li class="level2"><div class="li"><span class="li"><a href="#tools_for_performance_analysis" class="toc">Tools for Performance Analysis</a></span></div><ul class="toc"><li class="level3"><div class="li"><span class="li"><a href="#compiling_things_for_blackfin" class="toc">Compiling Things for Blackfin</a></span></div></li><li class="level3"><div class="li"><span class="li"><a href="#bandwidth_benchmarks" class="toc">Bandwidth Benchmarks</a></span></div></li><li class="level3"><div class="li"><span class="li"><a href="#latency_benchmarks" class="toc">Latency Benchmarks</a></span></div></li><li class="level3"><div class="li"><span class="li"><a href="#results" class="toc">results</a></span></div></li></ul></li></ul></li></ul></div></div><h2><a name="tools_for_performance_analysis" id="tools_for_performance_analysis">Tools for Performance Analysis</a></h2><div class="level2"><ul><li class="level1"><div class="li"> Project <a href="http://www.bitmover.com/lmbench/" class="urlextern" title="http://www.bitmover.com/lmbench/" rel="nofollow">Home Page</a> <a href="http://sourceforge.net/projects/lmbench" class="urlextern" title="http://sourceforge.net/projects/lmbench" rel="nofollow">Project Page</a></div></li></ul><p> lmbench includes the following nifty tests, which can run on Blackfin:</p><ul><li class="level1"><div class="li"> Bandwidth benchmarks </div><ul><li class="level2"><div class="li"> Cached file read </div></li><li class="level2"><div class="li"> Memory copy (bcopy) </div></li><li class="level2"><div class="li"> Memory read </div></li><li class="level2"><div class="li"> Memory write </div></li><li class="level2"><div class="li"> Pipe </div></li><li class="level2"><div class="li"> TCP </div></li></ul></li><li class="level1"><div class="li"> Latency benchmarks </div><ul><li class="level2"><div class="li"> Context switching. </div></li><li class="level2"><div class="li"> Networking: connection establishment, pipe, TCP, UDP, and <acronym title="Remote Procedure Call">RPC</acronym> hot potato </div></li><li class="level2"><div class="li"> File system creates and deletes. </div></li><li class="level2"><div class="li"> Process creation. </div></li><li class="level2"><div class="li"> Signal handling </div></li><li class="level2"><div class="li"> System call overhead </div></li><li class="level2"><div class="li"> Memory read latency </div></li></ul></li><li class="level1"><div class="li"> Miscellanious </div><ul><li class="level2"><div class="li"> Processor clock rate calculation </div></li></ul></li></ul></div><!-- SECTION [1-705] --><h3><a name="compiling_things_for_blackfin" id="compiling_things_for_blackfin">Compiling Things for Blackfin</a></h3><div class="level3"><p> lmbench uses an “interesting” build system, which is not really comptaible with building for embedded targets. It is not as simple as doing a:</p><pre class="code">make OS=bfin-uclinux-gnu CC='bfin-uclinux-gcc -Wl,-elf2flt'</pre><p>:(</p></div><!-- SECTION [706-970] --><h3><a name="bandwidth_benchmarks" id="bandwidth_benchmarks">Bandwidth Benchmarks</a></h3><div class="level3"></div><h4><a name="bw_file_rd" id="bw_file_rd">bw_file_rd</a></h4><div class="level4"><p>time the reading and summing of a file </p></div><h4><a name="bw_mem_cp" id="bw_mem_cp">bw_mem_cp</a></h4><div class="level4"><p>time memory copy speeds</p></div><h4><a name="bw_mem_rd" id="bw_mem_rd">bw_mem_rd</a></h4><div class="level4"><p>time memory read rate (with overhead) </p></div><h4><a name="bw_mem_wr" id="bw_mem_wr">bw_mem_wr</a></h4><div class="level4"><p>time memory write rate (with overhead) </p></div><h4><a name="bw_mmap_rd" id="bw_mmap_rd">bw_mmap_rd</a></h4><div class="level4"><p>time the reading and summing of a file </p></div><h4><a name="bw_pipe" id="bw_pipe">bw_pipe</a></h4><div class="level4"><p>time data movement through pipes </p></div><h4><a name="bw_tcp" id="bw_tcp">bw_tcp</a></h4><div class="level4"><p>time data movement through TCP/IP sockets </p></div><!-- SECTION [971-1377] --><h3><a name="latency_benchmarks" id="latency_benchmarks">Latency Benchmarks</a></h3><div class="level3"></div><h4><a name="lat_connect" id="lat_connect">lat_connect</a></h4><div class="level4"><p>measure interprocess connection latency via TCP/IP </p></div><h4><a name="lat_ctx" id="lat_ctx">lat_ctx</a></h4><div class="level4"><p>context switching benchmark </p></div><h4><a name="lat_fs" id="lat_fs">lat_fs</a></h4><div class="level4"><p>measure file system create/delete performance </p></div><h4><a name="lat_mem_rd" id="lat_mem_rd">lat_mem_rd</a></h4><div class="level4"><p>memory read latency benchmark</p></div><h4><a name="lat_mmap" id="lat_mmap">lat_mmap</a></h4><div class="level4"><p>csts of mmapping and unmmapping varying file sizes </p></div><h4><a name="lat_pagefault" id="lat_pagefault">lat_pagefault</a></h4><div class="level4"><p> - measure the cost of pagefaulting pages from a file </p></div><h4><a name="lat_pipe.8" id="lat_pipe.8">lat_pipe.8</a></h4><div class="level4"><p>measure interprocess communication latency through pipes </p></div><h4><a name="lat_proc.8" id="lat_proc.8">lat_proc.8</a></h4><div class="level4"><p>process creation tests</p></div><h4><a name="lat_rpc.8" id="lat_rpc.8">lat_rpc.8</a></h4><div class="level4"><p>measure interprocess communication latency via Sun <acronym title="Remote Procedure Call">RPC</acronym> </p></div><h4><a name="lat_syscall" id="lat_syscall">lat_syscall</a></h4><div class="level4"><p>time simple entry into the operating system</p></div><h4><a name="lat_tcp" id="lat_tcp">lat_tcp</a></h4><div class="level4"><p>measure interprocess communication latency via TCP/IP </p></div><h4><a name="lat_udp" id="lat_udp">lat_udp</a></h4><div class="level4"><p>measure interprocess communication latency via UDP/IP </p></div><!-- SECTION [1378-2167] --><h3><a name="results" id="results">results</a></h3><div class="level3"><p> For other architectures, they can be found on the project <a href="http://www.bitmover.com/lmbench/lmbench-summary" class="urlextern" title="http://www.bitmover.com/lmbench/lmbench-summary" rel="nofollow">web site</a>.</p></div><!-- SECTION [2168-] --></body></html>⌨️ 快捷键说明
复制代码
Ctrl + C
搜索代码
Ctrl + F
全屏模式
F11
切换主题
Ctrl + Shift + D
显示快捷键
?
增大字号
Ctrl + =
减小字号
Ctrl + -