<?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 hi, did you get the problem in OpenCL* for CPU</title>
    <link>https://community.intel.com/t5/OpenCL-for-CPU/Kernel-not-linking-properly-on-Xeon-Phi-accelerator/m-p/962653#M2229</link>
    <description>&lt;P&gt;hi, did you get the problem sorted out? how did you fix it?&amp;nbsp;&lt;/P&gt;</description>
    <pubDate>Sun, 30 Mar 2014 21:34:06 GMT</pubDate>
    <dc:creator>Dave_O_</dc:creator>
    <dc:date>2014-03-30T21:34:06Z</dc:date>
    <item>
      <title>Kernel not linking properly on Xeon Phi accelerator</title>
      <link>https://community.intel.com/t5/OpenCL-for-CPU/Kernel-not-linking-properly-on-Xeon-Phi-accelerator/m-p/962650#M2226</link>
      <description>&lt;P&gt;Hi,&lt;/P&gt;
&lt;P&gt;My department has just installed a Xeon Phi accelerator in our research cluster and I want to test several simple OpenCL kernels on it.&lt;/P&gt;
&lt;P&gt;Consdering the OpenCL execution portability, I just changed the device type of my host source code from CPU or GPU to ACCELERATOR, but the execution crashes when the host program tries to build the binary from the kernel file.&lt;/P&gt;
&lt;P&gt;I have also tried to compile kernels offline using "ioc64", and it runs smoothly for CPUs:&lt;/P&gt;
&lt;P&gt;-bash-4.1$ ioc64 -cmd=build -input=dummy.cl -device=cpu&lt;BR /&gt;Using device: CPU&lt;/P&gt;
&lt;P&gt;Setting target instruction set architecture to: Default&lt;BR /&gt;Intel OpenCL CPU device was found!&lt;BR /&gt;Device name: Intel(R) Xeon(R) CPU E5-2660 0 @ 2.20GHz&lt;BR /&gt;Device version: OpenCL 1.2 (Build 67279)&lt;BR /&gt;Device vendor: Intel(R) Corporation&lt;BR /&gt;Device profile: FULL_PROFILE&lt;BR /&gt;Compilation started&lt;BR /&gt;Compilation done&lt;BR /&gt;Linking started&lt;BR /&gt;Linking done&lt;BR /&gt;Kernel &amp;lt;dummy&amp;gt; was successfully vectorized&lt;BR /&gt;Done.&lt;BR /&gt;Build succeeded!&lt;/P&gt;
&lt;P&gt;However, "ioc64" also fails, returning the following output:&lt;/P&gt;
&lt;P&gt;-bash-4.1$ ioc64 -cmd=build -input=dummy.cl -device=acc&lt;/P&gt;
&lt;P&gt;Setting target instruction set architecture to: Default&lt;BR /&gt;Intel OpenCL Xeon Phi Coprocessor device was found!&lt;BR /&gt;Device name: Intel(R) Many Integrated Core Acceleration Card&lt;BR /&gt;Device version: OpenCL 1.2 (Build 67279)&lt;BR /&gt;Device vendor: Intel(R) Corporation&lt;BR /&gt;Device profile: FULL_PROFILE&lt;BR /&gt;Compilation started&lt;BR /&gt;Compilation done&lt;BR /&gt;Linking started&lt;BR /&gt;Linking failed&lt;/P&gt;
&lt;P&gt;Build failed!&lt;/P&gt;
&lt;P&gt;Thank you so much in advance!&lt;/P&gt;</description>
      <pubDate>Mon, 29 Jul 2013 13:45:09 GMT</pubDate>
      <guid>https://community.intel.com/t5/OpenCL-for-CPU/Kernel-not-linking-properly-on-Xeon-Phi-accelerator/m-p/962650#M2226</guid>
      <dc:creator>Jorge_F_2</dc:creator>
      <dc:date>2013-07-29T13:45:09Z</dc:date>
    </item>
    <item>
      <title>Hi Jorge,</title>
      <link>https://community.intel.com/t5/OpenCL-for-CPU/Kernel-not-linking-properly-on-Xeon-Phi-accelerator/m-p/962651#M2227</link>
      <description>Hi Jorge,

Which version of MPSS do you have installed? Note that Intel® MPSS Update 3 or higher should be installed (available at &lt;A href="http://software.intel.com/en-us/articles/intel-manycore-platform-software-stack-mpss)" target="_blank"&gt;http://software.intel.com/en-us/articles/intel-manycore-platform-software-stack-mpss)&lt;/A&gt;. Could you please provide the logs from micinfo and miccheck utilities (/opt/intel/mic/bin/)?

Thanks,
Yuri</description>
      <pubDate>Mon, 29 Jul 2013 14:26:15 GMT</pubDate>
      <guid>https://community.intel.com/t5/OpenCL-for-CPU/Kernel-not-linking-properly-on-Xeon-Phi-accelerator/m-p/962651#M2227</guid>
      <dc:creator>Yuri_K_Intel</dc:creator>
      <dc:date>2013-07-29T14:26:15Z</dc:date>
    </item>
    <item>
      <title>Hi,</title>
      <link>https://community.intel.com/t5/OpenCL-for-CPU/Kernel-not-linking-properly-on-Xeon-Phi-accelerator/m-p/962652#M2228</link>
      <description>&lt;P&gt;Hi,&lt;/P&gt;
&lt;P&gt;Here is the output from "micinfo":&lt;/P&gt;
&lt;P&gt;-bash-4.1$ ./micinfo &lt;BR /&gt;MicInfo Utility Log&lt;/P&gt;
&lt;P&gt;Created Mon Jul 29 16:28:03 2013&lt;/P&gt;
&lt;P&gt;System Info&lt;BR /&gt; HOST OS : Linux&lt;BR /&gt; OS Version : 2.6.32-358.6.2.el6.x86_64&lt;BR /&gt; Driver Version : 6720-13&lt;BR /&gt; MPSS Version : 2.1.6720-13&lt;BR /&gt; Host Physical Memory : 65932 MB&lt;/P&gt;
&lt;P&gt;Device No: 0, Device Name: mic0&lt;/P&gt;
&lt;P&gt;Version&lt;BR /&gt; Flash Version : 2.1.02.0386&lt;BR /&gt; SMC Firmware Version : 1.14.4616&lt;BR /&gt; SMC Boot Loader Version : 1.8.4326&lt;BR /&gt; uOS Version : 2.6.38.8-g5f2543d&lt;BR /&gt; Device Serial Number : ADKC24201340&lt;/P&gt;
&lt;P&gt;Board&lt;BR /&gt; Vendor ID : 0x8086&lt;BR /&gt; Device ID : 0x2250&lt;BR /&gt; Subsystem ID : 0x2500&lt;BR /&gt; Coprocessor Stepping ID : 3&lt;BR /&gt; PCIe Width : Insufficient Privileges&lt;BR /&gt; PCIe Speed : Insufficient Privileges&lt;BR /&gt; PCIe Max payload size : Insufficient Privileges&lt;BR /&gt; PCIe Max read req size : Insufficient Privileges&lt;BR /&gt; Coprocessor Model : 0x01&lt;BR /&gt; Coprocessor Model Ext : 0x00&lt;BR /&gt; Coprocessor Type : 0x00&lt;BR /&gt; Coprocessor Family : 0x0b&lt;BR /&gt; Coprocessor Family Ext : 0x00&lt;BR /&gt; Coprocessor Stepping : B1&lt;BR /&gt; Board SKU : B1PRQ-5110P&lt;BR /&gt; ECC Mode : Enabled&lt;BR /&gt; SMC HW Revision : Product 225W Passive CS&lt;/P&gt;
&lt;P&gt;Cores&lt;BR /&gt; Total No of Active Cores : 60&lt;BR /&gt; Voltage : 963000 uV&lt;BR /&gt; Frequency : 1052631 kHz&lt;/P&gt;
&lt;P&gt;Thermal&lt;BR /&gt; Fan Speed Control : N/A&lt;BR /&gt; Fan RPM : N/A&lt;BR /&gt; Fan PWM : N/A&lt;BR /&gt; Die Temp : 45 C&lt;/P&gt;
&lt;P&gt;GDDR&lt;BR /&gt; GDDR Vendor : Elpida&lt;BR /&gt; GDDR Version : 0x1&lt;BR /&gt; GDDR Density : 2048 Mb&lt;BR /&gt; GDDR Size : 7936 MB&lt;BR /&gt; GDDR Technology : GDDR5 &lt;BR /&gt; GDDR Speed : 5.000000 GT/s &lt;BR /&gt; GDDR Frequency : 2500000 kHz&lt;BR /&gt; GDDR Voltage : 1501000 uV&lt;/P&gt;
&lt;P&gt;And here the output form "miccheck"&lt;/P&gt;
&lt;P&gt;-bash-4.1$ ./miccheck&lt;/P&gt;
&lt;P&gt;miccheck 2.1.6720-13, created 14:49:30 Apr 30 2013&lt;BR /&gt;Copyright 2011-2013 Intel Corporation All rights reserved&lt;/P&gt;
&lt;P&gt;Test 1 Ensure installation matches manifest : OK&lt;BR /&gt;Test 2 Ensure host driver is loaded : OK&lt;BR /&gt;Test 3 Ensure driver matches manifest : OK&lt;BR /&gt;Test 4 Detect all listed devices : OK&lt;BR /&gt;MIC 0 Test 1 Find the device : OK&lt;BR /&gt;MIC 0 Test 2 Check the POST code via PCI : OK&lt;BR /&gt;MIC 0 Test 3 Connect to the device : OK&lt;BR /&gt;MIC 0 Test 4 Check for normal mode : OK&lt;BR /&gt;MIC 0 Test 5 Check the POST code via SCIF : OK&lt;BR /&gt;MIC 0 Test 6 Send data to the device : OK&lt;BR /&gt;MIC 0 Test 7 Compare the PCI configuration : OK&lt;BR /&gt;MIC 0 Test 8 Ensure Flash version matches manifest : OK&lt;/P&gt;
&lt;P&gt;Status: OK.&lt;/P&gt;
&lt;P&gt;According to these logs, I suppose that the MPSS installed version is "2.1.6720-13", but I don't know if this number is directly related to the "update number" you mentioned in the last message.&lt;/P&gt;
&lt;P&gt;Thanks,&lt;/P&gt;
&lt;P&gt;Jorge.&lt;/P&gt;</description>
      <pubDate>Mon, 29 Jul 2013 14:37:11 GMT</pubDate>
      <guid>https://community.intel.com/t5/OpenCL-for-CPU/Kernel-not-linking-properly-on-Xeon-Phi-accelerator/m-p/962652#M2228</guid>
      <dc:creator>Jorge_F_2</dc:creator>
      <dc:date>2013-07-29T14:37:11Z</dc:date>
    </item>
    <item>
      <title>hi, did you get the problem</title>
      <link>https://community.intel.com/t5/OpenCL-for-CPU/Kernel-not-linking-properly-on-Xeon-Phi-accelerator/m-p/962653#M2229</link>
      <description>&lt;P&gt;hi, did you get the problem sorted out? how did you fix it?&amp;nbsp;&lt;/P&gt;</description>
      <pubDate>Sun, 30 Mar 2014 21:34:06 GMT</pubDate>
      <guid>https://community.intel.com/t5/OpenCL-for-CPU/Kernel-not-linking-properly-on-Xeon-Phi-accelerator/m-p/962653#M2229</guid>
      <dc:creator>Dave_O_</dc:creator>
      <dc:date>2014-03-30T21:34:06Z</dc:date>
    </item>
    <item>
      <title>Hi,</title>
      <link>https://community.intel.com/t5/OpenCL-for-CPU/Kernel-not-linking-properly-on-Xeon-Phi-accelerator/m-p/962654#M2230</link>
      <description>&lt;P&gt;Hi,&lt;/P&gt;

&lt;P&gt;First, sorry for not having closed the thread earlier, it&amp;nbsp;has been a long time since I opened it...&lt;/P&gt;

&lt;P&gt;&lt;SPAN style="font-size: 1em; line-height: 1.5;"&gt;Well, at last I got the problem sorted out.&amp;nbsp;&lt;/SPAN&gt;&lt;SPAN style="font-size: 1em; line-height: 1.5;"&gt;S&lt;/SPAN&gt;&lt;SPAN style="font-size: 1em; line-height: 1.5;"&gt;o, the solution I applied might be not very conventional, but it worked: since in our research cluster we also have some Nvidia Tesla K20 devices installed, I simply unloaded the module of Intel OpenCL driver and then I exchanged it for the CUDA one. By doing that the execution of the host program run smoothly for all kind of devices: CPUs, GPUs and the MIC too.&lt;/SPAN&gt;&lt;/P&gt;

&lt;P&gt;&lt;SPAN style="font-size: 1em; line-height: 1.5;"&gt;Thanks again to everyone,&lt;/SPAN&gt;&lt;/P&gt;

&lt;P&gt;Jorge,&lt;/P&gt;</description>
      <pubDate>Wed, 02 Apr 2014 08:07:36 GMT</pubDate>
      <guid>https://community.intel.com/t5/OpenCL-for-CPU/Kernel-not-linking-properly-on-Xeon-Phi-accelerator/m-p/962654#M2230</guid>
      <dc:creator>Jorge_F_2</dc:creator>
      <dc:date>2014-04-02T08:07:36Z</dc:date>
    </item>
    <item>
      <title>Okay - thanks!</title>
      <link>https://community.intel.com/t5/OpenCL-for-CPU/Kernel-not-linking-properly-on-Xeon-Phi-accelerator/m-p/962655#M2231</link>
      <description>&lt;P&gt;Okay - thanks!&lt;BR /&gt;
	&lt;BR /&gt;
	Well, my program runs on Intel CPU and MIC using intel opencl. However, when I moved the code to Tesla K20 cluster with Nvidia opencl loaded, it failed to work. The cluster has no intel opencl, only nvidia's. I might have to look into it again...&lt;/P&gt;</description>
      <pubDate>Wed, 16 Apr 2014 23:20:45 GMT</pubDate>
      <guid>https://community.intel.com/t5/OpenCL-for-CPU/Kernel-not-linking-properly-on-Xeon-Phi-accelerator/m-p/962655#M2231</guid>
      <dc:creator>Dave_O_</dc:creator>
      <dc:date>2014-04-16T23:20:45Z</dc:date>
    </item>
  </channel>
</rss>

