<?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: ILLEGAL INSTRUCTION : ippiDCT8x8Inv_16s_C1I in Intel® Integrated Performance Primitives</title>
    <link>https://community.intel.com/t5/Intel-Integrated-Performance/ILLEGAL-INSTRUCTION-ippiDCT8x8Inv-16s-C1I/m-p/956988#M19172</link>
    <description>&lt;DIV&gt;please accept my apologies.sorry for taking your time for this thread. i've just discovered that the cpu is a via!!!#$#%&amp;amp;!&lt;/DIV&gt;
&lt;DIV&gt;&lt;/DIV&gt;
&lt;DIV&gt;i've trusted eMbedded C++ tool which says that the CPU is a PIII, but when i used cpuid, it revealed that it was not. :smileymad:&lt;/DIV&gt;
&lt;DIV&gt;&lt;/DIV&gt;
&lt;DIV&gt;fortunately, this via cpu supports mmx and hopefully i can use a "mmx-only" ipp library for my dct calls.&lt;/DIV&gt;
&lt;DIV&gt;&lt;/DIV&gt;
&lt;DIV&gt;does v5.0 supports mmx-only? if yes, which specific library should be used? if not, which version does?&lt;/DIV&gt;
&lt;DIV&gt;&lt;/DIV&gt;
&lt;DIV&gt;thanks again.&lt;/DIV&gt;
&lt;DIV&gt;&lt;/DIV&gt;</description>
    <pubDate>Thu, 22 Dec 2005 14:03:11 GMT</pubDate>
    <dc:creator>andybc</dc:creator>
    <dc:date>2005-12-22T14:03:11Z</dc:date>
    <item>
      <title>ILLEGAL INSTRUCTION : ippiDCT8x8Inv_16s_C1I</title>
      <link>https://community.intel.com/t5/Intel-Integrated-Performance/ILLEGAL-INSTRUCTION-ippiDCT8x8Inv-16s-C1I/m-p/956984#M19168</link>
      <description>&lt;DIV&gt;hello there.&lt;/DIV&gt;
&lt;DIV&gt;&lt;/DIV&gt;
&lt;DIV&gt;im just a newbie with these cool libraries. im just wondering what could be possibly wrong when i encounter the above illegal instruction. im using the trial version of ipp v.5 and trying to run the application in pentium3 under wince .net 4.2.&lt;/DIV&gt;
&lt;DIV&gt;&lt;/DIV&gt;
&lt;DIV&gt;any comments will be appreciated.&lt;/DIV&gt;
&lt;DIV&gt;&lt;/DIV&gt;
&lt;DIV&gt;cheers!&lt;/DIV&gt;
&lt;DIV&gt;andy&lt;/DIV&gt;</description>
      <pubDate>Mon, 19 Dec 2005 19:30:17 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-Integrated-Performance/ILLEGAL-INSTRUCTION-ippiDCT8x8Inv-16s-C1I/m-p/956984#M19168</guid>
      <dc:creator>andybc</dc:creator>
      <dc:date>2005-12-19T19:30:17Z</dc:date>
    </item>
    <item>
      <title>Re: ILLEGAL INSTRUCTION : ippiDCT8x8Inv_16s_C1I</title>
      <link>https://community.intel.com/t5/Intel-Integrated-Performance/ILLEGAL-INSTRUCTION-ippiDCT8x8Inv-16s-C1I/m-p/956985#M19169</link>
      <description>have you use an aligned memory buffer to push your dct data into the ipp function ...&lt;BR /&gt;&lt;BR /&gt;try to use aligned memory on 16-byte boundary with all optimized function ipp or other mmx/sse/sse2 optimized function,&lt;BR /&gt;&lt;BR /&gt;it could be solve your issue</description>
      <pubDate>Tue, 20 Dec 2005 16:11:53 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-Integrated-Performance/ILLEGAL-INSTRUCTION-ippiDCT8x8Inv-16s-C1I/m-p/956985#M19169</guid>
      <dc:creator>abuatnext04</dc:creator>
      <dc:date>2005-12-20T16:11:53Z</dc:date>
    </item>
    <item>
      <title>Re: ILLEGAL INSTRUCTION : ippiDCT8x8Inv_16s_C1I</title>
      <link>https://community.intel.com/t5/Intel-Integrated-Performance/ILLEGAL-INSTRUCTION-ippiDCT8x8Inv-16s-C1I/m-p/956986#M19170</link>
      <description>&lt;P&gt;Could you please provide simple test case to demostrate this wrong behaviour?&lt;/P&gt;
&lt;P&gt;Regards,&lt;BR /&gt; Vladimir&lt;/P&gt;
&lt;DIV&gt;&lt;/DIV&gt;</description>
      <pubDate>Wed, 21 Dec 2005 00:57:51 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-Integrated-Performance/ILLEGAL-INSTRUCTION-ippiDCT8x8Inv-16s-C1I/m-p/956986#M19170</guid>
      <dc:creator>Vladimir_Dudnik</dc:creator>
      <dc:date>2005-12-21T00:57:51Z</dc:date>
    </item>
    <item>
      <title>Re: ILLEGAL INSTRUCTION : ippiDCT8x8Inv_16s_C1I</title>
      <link>https://community.intel.com/t5/Intel-Integrated-Performance/ILLEGAL-INSTRUCTION-ippiDCT8x8Inv-16s-C1I/m-p/956987#M19171</link>
      <description>&lt;P&gt;thanks for the comments. but the 16-bit allignment did not solve the problem. originally the code is bit alligned as it runs with other platforms. the same source code runs perfectly with a pentium 3 stepping 8 (can anybody tell me what is meant by this stepping nomenclature. :smileyhappy:) but with a pentium 3 stepping 3, it does not. both have the same os. wince .net 4.2.&lt;/P&gt;
&lt;P&gt;the following is the dissassembly code in which the illegal instruction occurs. it stops at pshufw instruction. hope this helps to furtherly analyze the problem.&lt;/P&gt;
&lt;P&gt;_a6_dct_8x8_inv_16s:&lt;BR /&gt;1E03C920 mov ecx,dword ptr [esp+4]&lt;BR /&gt;1E03C924 mov edx,dword ptr [esp+8]&lt;BR /&gt;1E03C928 movq mm0,mmword ptr [ecx]&lt;BR /&gt;1E03C92B movq mm1,mmword ptr [ecx+8]&lt;BR /&gt;1E03C92F movq mm2,mm0&lt;BR /&gt;1E03C932 movq mm3,mmword ptr ds:[67CE0h]&lt;BR /&gt;1E03C939 pshufw mm0,mm0,88h&lt;/P&gt;
&lt;P&gt;thanks in advance.&lt;/P&gt;
&lt;P&gt;&lt;/P&gt;
&lt;DIV&gt;&lt;/DIV&gt;</description>
      <pubDate>Wed, 21 Dec 2005 16:53:33 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-Integrated-Performance/ILLEGAL-INSTRUCTION-ippiDCT8x8Inv-16s-C1I/m-p/956987#M19171</guid>
      <dc:creator>andybc</dc:creator>
      <dc:date>2005-12-21T16:53:33Z</dc:date>
    </item>
    <item>
      <title>Re: ILLEGAL INSTRUCTION : ippiDCT8x8Inv_16s_C1I</title>
      <link>https://community.intel.com/t5/Intel-Integrated-Performance/ILLEGAL-INSTRUCTION-ippiDCT8x8Inv-16s-C1I/m-p/956988#M19172</link>
      <description>&lt;DIV&gt;please accept my apologies.sorry for taking your time for this thread. i've just discovered that the cpu is a via!!!#$#%&amp;amp;!&lt;/DIV&gt;
&lt;DIV&gt;&lt;/DIV&gt;
&lt;DIV&gt;i've trusted eMbedded C++ tool which says that the CPU is a PIII, but when i used cpuid, it revealed that it was not. :smileymad:&lt;/DIV&gt;
&lt;DIV&gt;&lt;/DIV&gt;
&lt;DIV&gt;fortunately, this via cpu supports mmx and hopefully i can use a "mmx-only" ipp library for my dct calls.&lt;/DIV&gt;
&lt;DIV&gt;&lt;/DIV&gt;
&lt;DIV&gt;does v5.0 supports mmx-only? if yes, which specific library should be used? if not, which version does?&lt;/DIV&gt;
&lt;DIV&gt;&lt;/DIV&gt;
&lt;DIV&gt;thanks again.&lt;/DIV&gt;
&lt;DIV&gt;&lt;/DIV&gt;</description>
      <pubDate>Thu, 22 Dec 2005 14:03:11 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-Integrated-Performance/ILLEGAL-INSTRUCTION-ippiDCT8x8Inv-16s-C1I/m-p/956988#M19172</guid>
      <dc:creator>andybc</dc:creator>
      <dc:date>2005-12-22T14:03:11Z</dc:date>
    </item>
    <item>
      <title>Re: ILLEGAL INSTRUCTION : ippiDCT8x8Inv_16s_C1I</title>
      <link>https://community.intel.com/t5/Intel-Integrated-Performance/ILLEGAL-INSTRUCTION-ippiDCT8x8Inv-16s-C1I/m-p/956989#M19173</link>
      <description>&lt;P&gt;Hi,&lt;/P&gt;
&lt;P&gt;You need to use PX libraries, if your cpu is not 100% Intel compatible. Note, IPP dispatcher chooses PX libraries automatically in such cases, but seems you linked directly with A6 code, which is Pentium III specific.&lt;/P&gt;
&lt;P&gt;Regards,&lt;BR /&gt; Vladimir&lt;/P&gt;
&lt;DIV&gt;&lt;/DIV&gt;</description>
      <pubDate>Fri, 23 Dec 2005 01:53:30 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-Integrated-Performance/ILLEGAL-INSTRUCTION-ippiDCT8x8Inv-16s-C1I/m-p/956989#M19173</guid>
      <dc:creator>Vladimir_Dudnik</dc:creator>
      <dc:date>2005-12-23T01:53:30Z</dc:date>
    </item>
  </channel>
</rss>

