Hi,

I have trouble getting to run the test program included in the Intel MPI package.

Specially, I've installed MPI in my home directory (which is NSF-mounted on both host1 and host2 machines) and then compiled test.c using mpicc 

mpicc -show -o test test.c
gcc -o 'test''test.c' -I/home/jyli/intel/compilers_and_libraries_2017.2.174/linux/mpi/intel64/include -L/home/jyli/intel/compilers_and_libraries_2017.2.174/linux/mpi/intel64/lib/release_mt -L/home/jyli/intel/compilers_and_libraries_2017.2.174/linux/mpi/intel64/lib -Xlinker --enable-new-dtags -Xlinker -rpath -Xlinker /home/jyli/intel/compilers_and_libraries_2017.2.174/linux/mpi/intel64/lib/release_mt -Xlinker -rpath -Xlinker /home/jyli/intel/compilers_and_libraries_2017.2.174/linux/mpi/intel64/lib -Xlinker -rpath -Xlinker /opt/intel/mpi-rt/2107.0.0/intel64/lib/release_mt -Xlinker -rpath -Xlinker /opt/intel/mpi-rt/2017.0.0/intel64/lib -lmpifort -lmpi -lmpigi -ldl -lrt -lpthread

I checked that the machines connect to each other fine

mpirun -ppn 1 -n 2 -hosts host1,host2 hostname
host1
host2

However, when I run the test program, I encountered the following errors:

mpirun -ppn 1 -n 2 -hosts host1,host2 ./test

[0] MPI startup(): Intel(R) MPI Library, Version 2017 Update 2  Build 20170125 (id: 16752)
[0] MPI startup(): Copyright (C) 2003-2017 Intel Corporation.  All rights reserved.
[0] MPI startup(): Multi-threaded optimized library

===================================================================================
=   BAD TERMINATION OF ONE OF YOUR APPLICATION PROCESSES
=   PID 5582 RUNNING AT host2
=   EXIT CODE: 11
=   CLEANING UP REMAINING PROCESSES
=   YOU CAN IGNORE THE BELOW CLEANUP MESSAGES
===================================================================================

===================================================================================
=   BAD TERMINATION OF ONE OF YOUR APPLICATION PROCESSES
=   PID 5582 RUNNING AT host2
=   EXIT CODE: 11
=   CLEANING UP REMAINING PROCESSES
=   YOU CAN IGNORE THE BELOW CLEANUP MESSAGES
===================================================================================
   Intel(R) MPI Library troubleshooting guide:
https://software.intel.com/node/561764

===================================================================================

I then attached gdb to the core file generated, and here's the trace

gdb ./test core
(gdb) bt
#0  __GI_____strtol_l_internal (nptr=0x0, endptr=0x0, base=10, group=<optimized out>, loc=0x7faf932e2060 <_nl_global_locale>)
    at ../stdlib/strtol_l.c:298
#1  0x00007faf9368e11a in atoi (__nptr=<optimized out>) at /usr/include/stdlib.h:286
#2  i_mpi_numa_nodes_compare (a=0x0, b=0x0) at ../../src/mpid/ch3/src/mpid_init.c:62
#3  0x00007faf92f5b419 in msort_with_tmp (p=0x7fff5c532aa0, b=0xac7b30, n=2) at msort.c:83
#4  0x00007faf92f5b6cc in msort_with_tmp (n=2, b=0xac7b30, p=0x7fff5c532aa0) at msort.c:45
#5  __GI_qsort_r (b=0xac7b30, n=2, s=8, cmp=0x7faf9368e100 <i_mpi_numa_nodes_compare>, arg=<optimized out>) at msort.c:297
#6  0x00007faf936911af in MPID_nem_impi_create_numa_nodes_map () at ../../src/mpid/ch3/src/mpid_init.c:1305
#7  0x00007faf93692284 in MPID_Init (argc=0x0, argv=0x0, requested=10, provided=0x0, has_args=0x7faf932e2060 <_nl_global_locale>, has_env=0xac7db1)
    at ../../src/mpid/ch3/src/mpid_init.c:1732
#8  0x00007faf9362872b in MPIR_Init_thread (argc=0x0, argv=0x0, required=10, provided=0x0) at ../../src/mpi/init/initthread.c:717
#9  0x00007faf93615e2b in PMPI_Init (argc=0x0, argv=0x0) at ../../src/mpi/init/init.c:253
#10 0x0000000000400a6e in main ()
(gdb) 

Does anyone have some clue as to what is going on? Thank you very much in advance!

Jenny

I am just getting started with Phi and wanted to know what is the best way to get coding and testing that code?

I have been looking around for the last couple of days on the intel site and have found reference to starter kits from a couple of years ago and I see that you can get remote time on nodes too.  Just wanted to see what would be a good efficient (read as cheap) starting point so that I can have some valid performance data for proof of concept before investing in a full server.

Thanks,

Phil

Hi, comparison of code on i7 boxes with HDD against two different KNL systems (at least one of which has SSD) I see that the time for former to read two 2.6 GB files line by line is <20 seconds but for the KNL systems it is 70-90 seconds. Can anybody explain eg is it that KNL reads from disk over internal PCIe? Yours, M

Hi,

This is probably a simple, setup issue, but I am stumped.  I am compiling a simple application that queries the Phi board.  Currently, I am running this on my laptop that doesn't have any card installed.  After everything is compiled, I ran the application thinking that it would indicate that there are no cards installed and then terminate, but it doesn't work.  Windows is complaining that the application didn't start correctly.  What gives?  I suspected that the application will just return, indicating that there was no Phi card installed...

Note that I am also having problems with all the tutorial code that is found in the SDK as well.  They all return the same error...

Thanks,

John

Hello!

Running on Windows Server 2012, we encountered a situation where the micx64 driver started to log the message "Driver detected an internal error in its data structures for ." about 80 times per second. This filled up the event log rapidly, leaving only about 15 minutes of history with the default event log size.

Restarting the system resolved the issues, but I'd like to get to the bottom of this. For one I want to make sure everything is in order with the Xeon Phi cards. Furthermore, I don't want to lose event log data again due to the driver going on a logging spree.

The Server in question has four Xeon Phi 31S1P cards installed and is running MPSS 3.8.1 under a fully patched Windows Server 2012. The output of micinfo is attached to this post.

We also have a second server with the same hardware configuration where this issue did not occur. It might be worth noting that the machines are usually rebooted once a month for security updates, so this might be something that gets more likely with higher uptime.

Does anyone have suggestion on how to proceed to debug this? Are there any people here using Xeon Phi cards on Windows Server and if so, did you encounter any similar issues?

Kind regards,
Stefan

C:\Program Files\Intel\MPSS\bin>micctrl --start
Error manipulating coprocessors: Cannot find MpssLib.dll in loaded modules!

C:\Program Files\Intel\MPSS\bin>miccheck
MicCheck 3.8.1.3681
Copyright (c) 2016, Intel Corporation.

Executing default tests for host
  Test 0: Check number of devices the OS sees in the system ... pass
  Test 1: Check mic driver is loaded ... pass
  Test 2: Check number of devices driver sees in the system ... pass
Executing default tests for device: 0
  Test 3 (mic0): Check device is in online state and its postcode is FF ... fail
    device is not online, current state is 0

Status: FAIL
Failure: A device test failed

Hi,

I use PAPI to read the L2 miss count data on KNL. When I run two processes (MPI) on the tile (mapping one process onto each core), and read the L2 miss counter data on each process, how does the count data on each process is reported. The 2 cores on the KNL share the same L2 cache, how are the L2 misses which potentially originate from both the processes are distinguished?

A similar related issue is when multiple hyper threads are used on a core and when each thread reads the L1 miss count data. This issue I suppose can be handled given that process context saved between the thread context switches. However, it is not clear to me how does the earlier mentioned L2 counter sharing between the cores is handled?

Hi there!

I'm trying to offload some code to the Xeon Phi in my MFC application, however when I run the application and get to the offload command I get an 0xC0000005: Access violation reading location exception. The project is built with the Visual C++ compiler and I use the Intel C++ compiler for the file that has the offload code. If I do the exact same thing in a win32 console application everything works fine. 

I have also tried creating a static library with the Offload code in a separate project with the Intel C++ compiler and linking that to my MFC project, this gave the exact same result: also the access violation exception. The same issue with a DLL instead of the static library. 

I couldn't find any info if it is even possible to use the Xeon Phi in combination with MFC, does anyone know if and how this can be done?

The project properties are a standard MFC project with the following changes:

- C/C++>Language>Run Time Type Information> NO (/GR-)

- C/C++>General>Additional Include Directories> $(WindowsSDK_IncludePath)

If I run this application on the actual machine with the Xeon Phi it results in a BSOD with WHEA_UNCORRECTABLE_ERROR.

Any help is greatly appreciated!

Gerco

Hello,

Does anyone know how to adjust MCDRAM frequency on KNL?

I have confirmed that the DDR channel frequency can be controlled by the BIOS, but I don't know how to control that of the MCDRAM.

Do you think that it is possible to adjust MCDRAM frequency from MSR or CSR register etc.?

I desperately want to decrease MCDRAM frequency to half the level.

Is there any idea?

Thank you,

Daejin.

My intel Phi 7120 doesn't seem to work when I go to modprobe mic it says module mic not found.  I have 4G enabled and lspci sees the mic, and I am running at the 5GTs x16 width as shown on the flowchart but the card doesn't want to work.  I have shifted slots and changed the power.  My NVIDIA K40 seems to work with the same power connectors so it isn't that.  The blue light is also flashing on the back.

I have some cases where the new update doesn't want to auto-vectorize for a correctly assessed vector speedup of 1.5x to 2.0x, so now

#if  __MIC__

!dir$ simd

#endif

(avoiding array assignment synttax) is needed to retain performance.

Similar cases for previous 2017 builds in C have been requiring conditional compilation with

#if ! __MIC__

#pragma omp simd safelen(1)

#endif

because the compiler is too aggressive when compiling for AVX2, and the update didn't require a change in source code.

Cilk plus still makes the optimum decision about vectorization of CEAN for both KNC and AVX2.

I am currently a new user to intel xdk. And now I am developing a new app for android from Bootstrap plugin. Everything is clear to me but I can't insert image from local drive . In source folder there is only option like uss/images/startbug.jpg and other, but no option to upload from local drive. Can anyone help me with this?

Hi

I am working on processing a batch of small matrix multiplications on Knights Landing. Since the MKL's performance is not good for small matrix, I use libxsmm. However I find there are a lot of cache misses by using Vtune. The L1 cache miss rate is about 10%, and L2 cache miss rate is about 17%. 
The gflops it achieves is less than 20 for single thread. I also write a sample program to test the performance under a ideal condition (no or very few cache miss), it can achieve 50 gflops for single thread.

The code is:

 for(i = 0; i < batch; i++){
         const float *  inA = A +  i*Arows*Acols;
         const float *  inB = B +  i*Bcols*Brows;
         float * oto = out + i*Orows*Ocols;
	 libxsmm_smm_16_16_16(inA,inB,oto);
	// libxsmm_smm_16_16_16(A,B,out); //ideal condition, no or very few cache miss
}

In this case, it seems that the hardware prefetchers do not work well. So I am very curious about why the hardware prefetchers can not always prefetch the next matrix? Each matrix has size of 16*16. So for each gemm, the input matrices and the result can fit into L1 cache. The memory access is consecutive, so the prefether should prefetch data for the next matrix multiplication. However, the prefetchers are not, otherwise there should not be so many cache misses.

According to the Intel Xeon Phi Book, the hardware prefetcher will not stream across a 4KB boundary. Is that the problem?
Does the 4KB boundary mean the page size boundary? I also used the huge page, i.e., 2M page, through hugetlbfs. However, huge page is not going to help.

I also check the assembly code, even though the compiler’s software prefetching is enable, I do not see prefetch instruction in the assembly code. So I think I may need to trigger the software prefecther manually.

Any idea to optimize the program? 

Thanks！

Best regards,

Zhen

When i compile my codes with the following instructions,it will occur such problem.

#!/bin/bash
OP="-fomit-frame-pointer -O3 -funroll-loops -W -Wall -DARCH_INTEL64 -mavx -fopenmp"
#OPT="-DPARAM_OPT -DPOW_OPT -DCYG_OFFLOAD -DCYG_REACTION_OFFLOAD  -DCYG_MARCH_OFFLOAD -DNO_OUTPUT -fomit-frame-pointer -O2 -funroll-loops -DARCH_INTEL64 -fopenmp"
#OPT="-O3 -fopenmp"
source /WORK/app/osenv/ln1/set2.sh
#source /opt/intel/composer_xe_2013_sp1.2.144/bin/ifortvars.sh intel64
rm -rf *.o *.a ForestPose.exe
#module load gcc/4.9.2
mpic++ $OPT -c meanshift.cpp `pkg-config --cflags --libs gtk+-2.0` -I/WORK/app/opencv/3.0.0/include
mpic++ $OPT -c CRTree.cpp `pkg-config --cflags --libs gtk+-2.0` -I/WORK/app/opencv/3.0.0/include
mpic++ $OPT -c ForestPose.cpp `pkg-config --cflags --libs gtk+-2.0` -I/WORK/app/opencv/3.0.0/include
mpic++ $OPT -c CRTree_mpi.cpp
mpic++ $OPT *.o -o ForestPose.exe `pkg-config --cflags --libs gtk+-2.0`  -L/WORK/app/opencv/3.0.0/lib -lopencv_highgui -lopencv_core -lopencv_imgproc -lopencv_imgcodecs -Wl,-rpath-link /WORK/app/libgphoto2/2.5.8/lib

problem:

ld: ForestPoseMIC.o: Relocations in generic ELF (EM: 181)
ld: ForestPoseMIC.o: Relocations in generic ELF (EM: 181)
ld: ForestPoseMIC.o: Relocations in generic ELF (EM: 181)
ld: ForestPoseMIC.o: Relocations in generic ELF (EM: 181)
ld: ForestPoseMIC.o: Relocations in generic ELF (EM: 181)
ld: ForestPoseMIC.o: Relocations in generic ELF (EM: 181)
ld: ForestPoseMIC.o: Relocations in generic ELF (EM: 181)
ld: ForestPoseMIC.o: Relocations in generic ELF (EM: 181)
ForestPoseMIC.o: could not read symbols: File in wrong format

 I change the way of compile.when I run, another problem occurs.

instrutions:

#!/bin/bash
#OP="-fomit-frame-pointer -O3 -funroll-loops -W -Wall -DARCH_INTEL64 -mavx -fopenmp"
OPT="-fomit-frame-pointer -O2 -funroll-loops -DARCH_INTEL64 -fopenmp -W -Wall  -g"
#OPT="-O3 -fopenmp"
source /WORK/app/osenv/ln1/set2.sh
rm -rf *.o *.a ForestPose.exe
#module load gcc/4.9.2
/usr/local/mpi3-dynamic/bin/mpic++ $OPT -shared -c meanshift.cpp `pkg-config --cflags --libs gtk+-2.0` -I/WORK/app/opencv/3.0.0/include
/usr/local/mpi3-dynamic/bin/mpic++ $OPT -shared -c CRTree.cpp `pkg-config --cflags --libs gtk+-2.0` -I/WORK/app/opencv/3.0.0/include
/usr/local/mpi3-dynamic/bin/mpic++ $OPT -shared -c ForestPose.cpp `pkg-config --cflags --libs gtk+-2.0` -I/WORK/app/opencv/3.0.0/include
/usr/local/mpi3-dynamic/bin/mpic++ $OPT -shared -c CRTree_mpi.cpp
xiar -qoffload-build rcs libCRTree.a meanshift.o CRTree.o ForestPose.o
/usr/local/mpi3-dynamic/bin/mpic++ $OPT -shared CRTree_mpi.o -o ForestPose.exe libCRTreeMIC.a libCRTree.a `pkg-config --cflags --libs gtk+-2.0`  -L/WORK/app/opencv/3.0.0/lib -lopencv_highgui -lopencv_core -lopencv_imgproc -lopencv_imgcodecs -Wl,-rpath-link /WORK/app/libgphoto2/2.5.8/lib
#/usr/local/mpi3-dynamic/bin/mpic++ $OPT -shared CRTree_mpi.o -o ForestPose.exe libCRTree.a `pkg-config --cflags --libs gtk+-2.0`  -L/WORK/app/opencv/3.0.0/lib -lopencv_highgui -lopencv_core -lopencv_imgproc -lopencv_imgcodecs -Wl,-rpath-link /WORK/app/libgphoto2/2.5.8/lib

problem:

yhrun: error: cn10540: tasks 4-5: Segmentation fault

Hello,
We have a Phi 31s1p [1] with a Tyan 7059[2] board. There is a compatibility list from Tyan [3] which we conclude our board supports such device. In fact, the BIOS has the "above 4G decoding" option. So, we are totally sure that the board supports Phi. I have to mention that our phi is not new though!!

We have installed the Phi on one of the PCI slots.There is also another (low end) nVidia 280 on the board. When we power on the board, the LED for the nVidia blinks and the color is red. According to the manual (page 93) [2] that means it is working with lower speed. However, the LED for the Phi is solid red which is not explained in the manual (page 93) [2]. Also, there is a blue LED on the Phi and it blinks repeatedly.

Next, with the 'lspci' command on Ubuntu, there is no sign of Coprocessor or Phi.

1- What does the blue blinking LED on the Phi mean?
2- What should we do for more investigations?

Thanks.

[1] https://ark.intel.com/products/79539/Intel-Xeon-Phi-Coprocessor-31S1P-8G...
[2] http://www.tyan.com/manuals/FT77B-B7059_UG_v1.0a.pdf
[3] http://www.tyan.com/EN/solution/coprocessor

I am successfully able to compile and execute my code on Xeon, but when I try to compile it for executing on a co-processor (natively by adding a -mmic flag), I get the following error:

x86_64-k1om-linux-ld: cannot find -lpardiso500-GNU481-X86-64
x86_64-k1om-linux-ld: cannot find -lblas
x86_64-k1om-linux-ld: cannot find -llapack
x86_64-k1om-linux-ld: cannot find -lgfortran

My .sh file for compilation:

export LD_LIBRARY_PATH=/usr/local/lib:/opt/intel/compilers_and_libraries_2016.3.210/linux/compiler/lib/intel64_lin_mic/:$LD_LIBRARY_PATH
icc -mmic PIC.c -o PIC -pg -g -lpardiso500-GNU481-X86-64 -lblas -llapack -lgfortran -fopenmp -lpthread -lm -O2 -vec-report2

$LD_LIBRARY_PATH looks like:

/usr/local/lib:
/opt/intel/compilers_and_libraries_2016.3.210/linux/compiler/lib/intel64_lin_mic/:
/opt/intel/compilers_and_libraries_2016.3.210/linux/compiler/lib/intel64:
/opt/intel/compilers_and_libraries_2016.3.210/linux/compiler/lib/intel64_lin:
/opt/intel/compilers_and_libraries_2016.3.210/linux/mpi/intel64/lib:
/opt/intel/compilers_and_libraries_2016.3.210/linux/mpi/mic/lib:
/opt/intel/compilers_and_libraries_2016.3.210/linux/ipp/lib/intel64:
/opt/intel/mic/coi/host-linux-release/lib:
/opt/intel/mic/myo/lib:
/opt/intel/compilers_and_libraries_2016.3.210/linux/compiler/lib/intel64:
/opt/intel/compilers_and_libraries_2016.3.210/linux/mkl/lib/intel64:
/opt/intel/compilers_and_libraries_2016.3.210/linux/tbb/lib/intel64/gcc4.4:
/opt/intel/debugger_2016/libipt/intel64/lib:
/opt/intel/compilers_and_libraries_2016.3.210/linux/daal/lib/intel64_lin:
/opt/intel/compilers_and_libraries_2016.3.210/linux/daal/../compiler/lib/intel64_lin

Looking at previous posts on the forum, I tried adding '/usr/lib64/' to the $LD_LIBRARY_PATH but then it gives me the following error when I try to compile:-

relocation error: /usr/lib64/libc.so.6: symbol _dl_starting_up, version GLIBC_PRIVATE not defined in file ld-linux-x86-64.so.2 with link time reference

Also, it was suggested to avoid adding '/usr/lib64/'.

What sould I do next?

Regards

Keval Shah

Hello,

I am having trouble with Xeon Phi X200 co-processor in Windows.  I have tried both Server 2012 R2 and Windows 10 Pro (fresh installs), both attempts resolved in the same issue.  When I run "miccheck" the output is "Test 0: Check number of devices the OS sees in the system ... pass", "Test 1: Check required drivers are loaded ... fail".  If I run the command "net start mic" it returns "The Intel Xeon Phi x200 co-processor services was started successfully".  Micctrl --start the output is "Cannot find any Intel Xeon Phi x200 coprocessor.  Please check if MPSS has been installed properly."

I have successfully configured Knights Corner co-processors (with an older version of MPSS) in Windows, which was as easy as running the MPSS installers and having the prerequisite programs. Using the Knights Landing co-processor and MPSS 4.4.0.4724 I can't get the cards to start.  In device manager there is a yellow exclamation point next to the device, I have tried forcing the "MIC" driver but Windows rejects this, stating it's already using the most appropriate driver. 

I have followed the instructions at the following URL's verbatim:

https://software.intel.com/en-us/articles/xeon-phi-coprocessor-x200-quic...

http://registrationcenter-download.intel.com/akdlm/irc_nas/11597/readme-...

Python 2.7.5 and Pywin32 are both installed and working correctly.  I have also used Windows Management Console to attempt to resolve any permission issues that may exist, this did not help.  There is only one user account which has administrator privileges.  

Are there any known issues I'm unaware of that could be causing this problem?  Is there an undocumented step, or something I'm missing during installation?  

Thanks. 

Hello,

Are any of the following 3 configurations NUMA systems?

1. Server Dell PowerEdge R530 - Rack 2U - 2x Intel Xeon E5-2620v3 6C/12T 2.4GHz, 8GB (1x8GB) DDR4-2133 RDIMM, DVD-ROM, 2x 300GB 10K SAS (max. 8 x 3.5" hot-plug HDD), RAID PERC H730 Ctrl 12GBps 1GB Cache, iDRAC8 Enterprise, Hot-plug PS (1+0) 750W, ReadyRails

2. Lenovo x3500 M5, 2x Xeon 6C E5-2609v3 6C/6T  85W 1.9GHz/1600MHz/15MB, 1x8GB, HDD 1TB 7.2K 6Gbps NL SAS 2.5in, SR M1215, Multi-Burner, 550W p/s, Tower

3. Lenovo x3650 M5, Form Factor 2U, 2x Xeon 6C E5-2620v3 85W 2.4GHz/1866MHz/15MB, 8GB 2133MHz LP RDIMM, SerrverRAID , M5210(no flash/cache), 2x300GB, HS 2.5in SAS(8), Multiburner, Power 2x550W HS PSU

Microsoft Windows Server 2016 Standard Edition, 16 cores, 2VMs ROK for Dell servers

Windows Server CAL 2012 English 1pk DSP OEI 5 Clt Device CAL

Windows Server CAL 2012 English 1pk DSP OEI 5 Clt User CAL

Windows Server CAL 2012 English 1pk DSP OEI 1 Clt Device CAL

Windows Server CAL 2012 English 1pk DSP OEI 1 Clt User CAL

Thank you!

My code as below:

      include 'param.h'
      dimension alpha1(mx)
      open(99,file='checking.dat')

and there is an error: Error 5 error #6591: An automatic object is invalid in a main program.[alpha1]

I cannot understand what I can fix it. Can someone help me?

Thank you.

Hi,

I installed MPSS 3.8.2 on a CentOS 7.3 system with kernel 3.10.0-514.16.1

I built the kernel module package for this kernel version. The mic module loads, but it reports "no MIC boards present"

I also tried booting with the default kernel, 3.10.0-514 (and the module from the MPSS rpm package). Again, the module loads but reports "no MIC boards present"

In lspci I can see the Xeon Phi card, and it seems to have memory assigned. On the other hand, miccheck fails on first test:

Executing default tests for host
  Test 0: Check number of devices the OS sees in the system ... fail
    no Intel(R) Xeon Phi(TM) coprocessors devices detected

Status: FAIL

I have attached a tgz with the output of lspci, dmesg, etc

Thanks in advance and kind regards,