Changes between Version 1 and Version 2 of laguna/faq


Ignore:
Timestamp:
01/30/2018 01:13:23 PM (9 months ago)
Author:
Tim Harvey
Comment:

convert restored html to wiki markup

Legend:

Unmodified
Added
Removed
Modified
  • laguna/faq

    v1 v2  
    1 {{{#!html
    2           <div id="wikipage" class="trac-content"><p>
    3 </p><div class="wiki-toc">
    4 <ol>
    5   <li>
    6     <a href="#LagunaFAQFrequentlyAskedQuestions"><b style="color:#000;background:#ffcc99">Laguna FAQ (Frequently Asked Questions</b>)</a>
    7     <ol>
    8       <li>
    9         <a href="#BogoMIPsvaluesappearlowerthenexpected">BogoMIPs values appear lower then expected</a>
    10       </li>
    11     </ol>
    12   </li>
    13 </ol>
    14 </div><p>
    15 </p>
    16 <h1 id="LagunaFAQFrequentlyAskedQuestions"><b style="color:#000;background:#ffcc99">Laguna FAQ (Frequently Asked Questions</b>)</h1>
    17 <p>
    18 <span class="wikianchor" id="bogomips"></span>
    19 </p>
    20 <h2 id="BogoMIPsvaluesappearlowerthenexpected">BogoMIPs values appear lower then expected</h2>
    21 <p>
    22 The reported bogoMIPS value is indeed bogus in the case of the ARM11 mpcore and other more modern ARM processors. The processor has a hardware function that helps to make "tight loops" more efficient by implementing branch folding. When the software attempts to discover its bogomips it executes a function called <span class="underline">delay with a number of instructions it wants to delay.  The </span>delay function is a "tight loop" and the branch instruction gets "folded" away causing the number of cycles to be far fewer than expected, yielding an incorrect value for bogomips.
    23 </p>
    24 <p>
     1[[PageOutline]]
     2
     3= Laguna FAQ (Frequently Asked Questions) =
     4
     5== BogoMIPs values appear lower then expected ==
     6The reported bogoMIPS value is indeed bogus in the case of the ARM11 mpcore and other more modern ARM processors. The processor has a hardware function that helps to make "tight loops" more efficient by implementing branch folding. When the software attempts to discover its bogomips it executes a function called delay with a number of instructions it wants to delay. The delay function is a "tight loop" and the branch instruction gets "folded" away causing the number of cycles to be far fewer than expected, yielding an incorrect value for bogomips.
     7
    258For the Cavium cns3xxx processors, expect the following bogomips values:
    26 </p>
    27 <ul><li>dual-core 600Mhz cns3420 (GW2388): 239 bogomips each
    28 </li><li>single-core 300MHz cns3410 (GW2380): 119 bogomips
    29 </li></ul><p>
     9 * dual-core 600Mhz cns3420 (GW2388): 239 bogomips each
     10 * single-core 300MHz cns3410 (GW2380): 119 bogomips
     11
    3012References:
    31 </p>
    32 <ul><li><a class="ext-link" href="http://infocenter.arm.com/help/index.jsp?topic=/com.arm.doc.ddi0360f/ch06s02s04.html"><span class="icon">​</span>http://infocenter.arm.com/help/index.jsp?topic=/com.arm.doc.ddi0360f/ch06s02s04.html</a>
    33 </li><li><a class="ext-link" href="http://lists.infradead.org/pipermail/linux-arm-kernel/2010-January/008676.html"><span class="icon">​</span>http://lists.infradead.org/pipermail/linux-arm-kernel/2010-January/008676.html</a>
    34 </li><li><a class="ext-link" href="http://lists.infradead.org/pipermail/linux-arm-kernel/2010-January/008912.html"><span class="icon">​</span>http://lists.infradead.org/pipermail/linux-arm-kernel/2010-January/008912.html</a>
    35 </li></ul
    36 }}}
     13 * ​http://infocenter.arm.com/help/index.jsp?topic=/com.arm.doc.ddi0360f/ch06s02s04.html
     14​ * http://lists.infradead.org/pipermail/linux-arm-kernel/2010-January/008676.html
     15​ * http://lists.infradead.org/pipermail/linux-arm-kernel/2010-January/008912.html