<?xml version="1.0" encoding="UTF-8"?>
<rss xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" xmlns:taxo="http://purl.org/rss/1.0/modules/taxonomy/" version="2.0">
  <channel>
    <title>topic Re: Code portability to Alpha Unix in Software Archive</title>
    <link>https://community.intel.com/t5/Software-Archive/Code-portability-to-Alpha-Unix/m-p/936120#M15926</link>
    <description>Our UNIX compilers and the CVF compiler are the same underneath, so other than OS-specific features, the compilers are as compatible as can be.  DFLIB (MSFLIB is just a synonym for this) isn't supported on UNIX, but CVF does support many UNIX-compatible routines in module DFPORT (in many cases, you may not need to USE DFPORT to use these.)&lt;BR /&gt;&lt;BR /&gt;Jugoslav, you'll like our new INT_PTR_KIND() intrinsic in 6.5A (and 5.4 on UNIX).&lt;BR /&gt;&lt;BR /&gt;Steve</description>
    <pubDate>Mon, 22 Jan 2001 02:28:12 GMT</pubDate>
    <dc:creator>Steven_L_Intel1</dc:creator>
    <dc:date>2001-01-22T02:28:12Z</dc:date>
    <item>
      <title>Code portability to Alpha Unix</title>
      <link>https://community.intel.com/t5/Software-Archive/Code-portability-to-Alpha-Unix/m-p/936118#M15924</link>
      <description>A simple question for which I think I know the answer, but my customer would like confirmation.  &lt;BR /&gt;  &lt;BR /&gt;Currently I have source code that compiles using Compaq 6.1 for Windows, the only external libraries used are MSFLIB and DFLIB.  My customer inquired if our source code could be compiled on the UNIX Alpha version.  They have not purchased the compiler yet. &lt;BR /&gt;  &lt;BR /&gt;I don't expect the MSLIB files to work, but I'm only using QuickWin for run-time status and summaries, it can go.  Are there other code portability issues between the two compilers?  Is there anything else I should tell my customer? (note, I'm not fluent with UNIX, but I can relay the information).  &lt;BR /&gt;  &lt;BR /&gt;Thanks,  &lt;BR /&gt;Dennis Hussey</description>
      <pubDate>Sat, 20 Jan 2001 08:03:59 GMT</pubDate>
      <guid>https://community.intel.com/t5/Software-Archive/Code-portability-to-Alpha-Unix/m-p/936118#M15924</guid>
      <dc:creator>Intel_C_Intel</dc:creator>
      <dc:date>2001-01-20T08:03:59Z</dc:date>
    </item>
    <item>
      <title>Re: Code portability to Alpha Unix</title>
      <link>https://community.intel.com/t5/Software-Archive/Code-portability-to-Alpha-Unix/m-p/936119#M15925</link>
      <description>By chance, we encountered porting our existing code written in CVF5 &lt;BR /&gt;into F90/TruUNIX/Alpha and we had no serious problems related with &lt;BR /&gt;compilers (we had problems, though, due to different architecture of &lt;BR /&gt;systems). Both compilers support most of the same intrinsics and extensions. We protected the OS-dependent stuff with preprocessor &lt;BR /&gt;(!DEC$IF DEFINED works fine on both) and there were only few &lt;BR /&gt;functions to replace (luckily, in few source file which are compiled only under &lt;BR /&gt;UNIX anyway). The porting problems we encountered were: &lt;BR /&gt; &lt;BR /&gt;- Case sensitivity in file names and different folder separator (/ and ) &lt;BR /&gt; &lt;BR /&gt;- There were some cray pointers involved, so we had to define a &lt;BR /&gt;  constant PTR_SIZE: &lt;BR /&gt; &lt;BR /&gt;  &lt;PRE&gt; 
  !DEC$IF DEFINED _UNIX 
  INTEGER, PARAMETER:: PTR_SIZE=8 
  !DEC$ELSE 
  INTEGER, PARAMETER:: PTR_SIZE=4 
  !DEC$ENDIF 
  &lt;/PRE&gt; &lt;BR /&gt; &lt;BR /&gt;and declare corresponding pointers as INTEGER(PTR_SIZE); &lt;BR /&gt; &lt;BR /&gt;- We also needed just a few interfaces to standard C library routines &lt;BR /&gt;  (e.g. localtime -- but that one is specific for interfacing the ported &lt;BR /&gt;  code with our partners' application environment) &lt;BR /&gt; &lt;BR /&gt;In total, we have only about ten !DEC$IF DEFINED (_UNIX) throughout &lt;BR /&gt;the 40,000 lines code, where only 2-3 were absolutely necessary (others &lt;BR /&gt;are related with functional differences). So, I'd say that with a little luck &lt;BR /&gt;the porting can be done very fast. &lt;BR /&gt; &lt;BR /&gt;Actually, the most of the problems were in synchronizing the data structures interchanged between our F90 and their C code (alignment, shared memory, POINTERs and similar), but for a stand-alone application &lt;BR /&gt;I believe there should be no such problems. &lt;BR /&gt; &lt;BR /&gt;HTH &lt;BR /&gt; &lt;BR /&gt;Jugoslav</description>
      <pubDate>Sun, 21 Jan 2001 01:13:27 GMT</pubDate>
      <guid>https://community.intel.com/t5/Software-Archive/Code-portability-to-Alpha-Unix/m-p/936119#M15925</guid>
      <dc:creator>Jugoslav_Dujic</dc:creator>
      <dc:date>2001-01-21T01:13:27Z</dc:date>
    </item>
    <item>
      <title>Re: Code portability to Alpha Unix</title>
      <link>https://community.intel.com/t5/Software-Archive/Code-portability-to-Alpha-Unix/m-p/936120#M15926</link>
      <description>Our UNIX compilers and the CVF compiler are the same underneath, so other than OS-specific features, the compilers are as compatible as can be.  DFLIB (MSFLIB is just a synonym for this) isn't supported on UNIX, but CVF does support many UNIX-compatible routines in module DFPORT (in many cases, you may not need to USE DFPORT to use these.)&lt;BR /&gt;&lt;BR /&gt;Jugoslav, you'll like our new INT_PTR_KIND() intrinsic in 6.5A (and 5.4 on UNIX).&lt;BR /&gt;&lt;BR /&gt;Steve</description>
      <pubDate>Mon, 22 Jan 2001 02:28:12 GMT</pubDate>
      <guid>https://community.intel.com/t5/Software-Archive/Code-portability-to-Alpha-Unix/m-p/936120#M15926</guid>
      <dc:creator>Steven_L_Intel1</dc:creator>
      <dc:date>2001-01-22T02:28:12Z</dc:date>
    </item>
  </channel>
</rss>

