<?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 Boris, this is a known issue in Intel® oneAPI Math Kernel Library</title>
    <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/zgeev-problem/m-p/1137428#M26137</link>
    <description>&lt;P&gt;Boris, this is a known issue and we are planning to fix it into one of the next updates of MKL v.2020. We will keep you informed when this update would be released.&lt;/P&gt;</description>
    <pubDate>Tue, 10 Mar 2020 08:14:00 GMT</pubDate>
    <dc:creator>Gennady_F_Intel</dc:creator>
    <dc:date>2020-03-10T08:14:00Z</dc:date>
    <item>
      <title>zgeev problem</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/zgeev-problem/m-p/1137427#M26136</link>
      <description>&lt;P&gt;Seems&amp;nbsp;that zgeev routine could write past the end of the rwork array of the recommended size ( 2n ).&lt;/P&gt;&lt;P&gt;here is my code calling zgeev&lt;/P&gt;
&lt;PRE class="brush:cpp; class-name:dark;"&gt;using namespace std;
#define ComplexD std::complex&amp;lt;double&amp;gt;
bool zeigValVecA(ComplexD *Val, ComplexD *Vec, ComplexD *A, int *n, bool *brk)
{
	char left = 'N', right = 'V';
	int ok(0), LW(-1);

	vector&amp;lt;double&amp;gt; RWORK(size_t(*n)*2); // +1 fixes the access beyond the end of the array inside MKL
	ComplexD work;
	mkl_brk = brk;

	ZGEEV(&amp;amp;left,&amp;amp;right,n,A,n,Val,nullptr,n,Vec,n,&amp;amp;work,&amp;amp;LW,RWORK.data(),&amp;amp;ok);

	if (*brk) return *brk;

	LW=int(work.real());
	vector&amp;lt;ComplexD&amp;gt; WORK(LW);
	
	ZGEEV(&amp;amp;left,&amp;amp;right,n,A,n,Val,nullptr,n,Vec,n,WORK.data(),&amp;amp;LW,RWORK.data(),&amp;amp;ok);
	
	return ok!=0 || *brk;
}&lt;/PRE&gt;

&lt;P&gt;This code resides in a dll statically linked with MKL&amp;nbsp; and called from a main program written in Delphi. Dll is compiled with the latest Visual Studio 2019. Debugging the the main program I get the following debug output (presumably from VS C++ RTL)&lt;/P&gt;
&lt;P&gt;Debug Output: HEAP[ModalCol.exe]: &amp;nbsp;Process ModalCol.exe (13956)&lt;BR /&gt;Debug Output:&lt;BR /&gt;Heap block at 000000001525E490 modified at 000000001525F130 past requested size of c90&lt;/P&gt;
&lt;P&gt;Offset is beyond the end of the vector storage (n=201). Adding +1 to the RWORK length fixes the problem. The smallest n where i got this behavior is 17.&lt;/P&gt;
&lt;P&gt;Tests were made on Windows 10 1909 x64,&amp;nbsp; Intel Core i9-9900K processor,&amp;nbsp;INTEL_MKL_VERSION 20200000, code is 64bit.&lt;/P&gt;
&lt;P&gt;&amp;nbsp;&lt;/P&gt;</description>
      <pubDate>Sat, 07 Mar 2020 10:04:03 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/zgeev-problem/m-p/1137427#M26136</guid>
      <dc:creator>Usievich__Boris</dc:creator>
      <dc:date>2020-03-07T10:04:03Z</dc:date>
    </item>
    <item>
      <title>Boris, this is a known issue</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/zgeev-problem/m-p/1137428#M26137</link>
      <description>&lt;P&gt;Boris, this is a known issue and we are planning to fix it into one of the next updates of MKL v.2020. We will keep you informed when this update would be released.&lt;/P&gt;</description>
      <pubDate>Tue, 10 Mar 2020 08:14:00 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/zgeev-problem/m-p/1137428#M26137</guid>
      <dc:creator>Gennady_F_Intel</dc:creator>
      <dc:date>2020-03-10T08:14:00Z</dc:date>
    </item>
    <item>
      <title>Thanks, I suppose that for</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/zgeev-problem/m-p/1137429#M26138</link>
      <description>&lt;P&gt;Thanks, I suppose that for now my +1 fix is enough?&lt;/P&gt;</description>
      <pubDate>Tue, 10 Mar 2020 09:00:44 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/zgeev-problem/m-p/1137429#M26138</guid>
      <dc:creator>Usievich__Boris</dc:creator>
      <dc:date>2020-03-10T09:00:44Z</dc:date>
    </item>
    <item>
      <title>2*n+32 will be better. We</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/zgeev-problem/m-p/1137430#M26139</link>
      <description>&lt;P&gt;2*n+32 will be better. We will add some detailed instructions about this issue into the coming MKL 2020.1&lt;/P&gt;</description>
      <pubDate>Tue, 10 Mar 2020 13:39:22 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/zgeev-problem/m-p/1137430#M26139</guid>
      <dc:creator>Gennady_F_Intel</dc:creator>
      <dc:date>2020-03-10T13:39:22Z</dc:date>
    </item>
    <item>
      <title>the problem has been fixed in</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/zgeev-problem/m-p/1137431#M26140</link>
      <description>&lt;P&gt;the problem has been fixed in mkl v.2020.1&lt;/P&gt;</description>
      <pubDate>Thu, 02 Apr 2020 05:47:30 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/zgeev-problem/m-p/1137431#M26140</guid>
      <dc:creator>Gennady_F_Intel</dc:creator>
      <dc:date>2020-04-02T05:47:30Z</dc:date>
    </item>
    <item>
      <title>Can confirm that the problem</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/zgeev-problem/m-p/1137432#M26141</link>
      <description>&lt;P&gt;Can confirm that the problem has been fixed in 2020.1&lt;/P&gt;</description>
      <pubDate>Tue, 07 Apr 2020 12:40:00 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/zgeev-problem/m-p/1137432#M26141</guid>
      <dc:creator>Usievich__Boris</dc:creator>
      <dc:date>2020-04-07T12:40:00Z</dc:date>
    </item>
    <item>
      <title>thanks Boris</title>
      <link>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/zgeev-problem/m-p/1137433#M26142</link>
      <description>&lt;P&gt;thanks Boris&lt;/P&gt;</description>
      <pubDate>Tue, 07 Apr 2020 13:02:51 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/zgeev-problem/m-p/1137433#M26142</guid>
      <dc:creator>Gennady_F_Intel</dc:creator>
      <dc:date>2020-04-07T13:02:51Z</dc:date>
    </item>
  </channel>
</rss>

