Computación con clústeres

Hybrid OpenMP+MPI: How to synchronize access to MPI shared RMA window by multiple OpenMP threads ?

Dear Collegues,

I've developed parallel code that implements both OpenMP + MPI hybrid parallelization and allocates a shared RMA window by calling MPI_Win_allocate_shared function to store the data shared among multiple OpenMP threads executed withing several MPI processes.

Is there any chance to synchronize access to the shared RMA window by the number of OpenMP threads executed within multiple MPI processes.?

Is there any pattern for hybrid OpenMP + MPI programming ?

Thanks for your replies in advance.

Cheers, Arthur.

Get the Power Consumption Info When Running a Program


When I run a program using offload mode, I want to collect the information of power consumption of the MIC. Is there any way to do that?

I found that I can use the micsmc to get the Total Power info in the command line/GUI interface.( Could I add a function from any libraries inside a program to collect the power consumption when use offload mode?

MPI_FILE_SET_VIEW produces a seg fault in Windows 10

I have a large CFD code that uses a parallel MPI write routine. The code compiles and runs on our Windows 7 machines (Intel Fortran 16 and Intel MPI 5.1.2), but the code fails under Windows 10. The failure always occurs in the call MPI_FILE_SET_VIEW. I wrote a short program to demonstrate the problem. This program runs on a Windows 7 machine and fails under Windows 10, regardless of which platform we compile on.

Remote procedure call, license of code call to scif library

We have developed a low latency, remote procedure call (rpc) Framework  for Xeon <--> Xeon Phi based on SCIF, which is very successful for our applications.  We think it could be helpful for other Intel Xeon Phi  users.  We would like to share the code to communities as open-source project. However we does not know if  there are any problem with Intel's license, copyright or copyright infringement, which we believe do not. Would you please help us? Thanks, Minh

execvp error on file, The requested operation requires elevation.

This is Tansel. I am not a direct MPI user. Someone from my team created an application using MPI, using a single batch file to run his executable on mpiexec.

The problem is, I need to integrate this into an automated system (written in Java) and my software will call this batch at some point. However, when I do, I get the following error: (batch file launches the software in the first line, [0+] are errors)
(Program runs fine if I run the batch file w/ double click or from any command line)

My MPI program doesn't work (hangs) when you launch processes on different nodes (hosts)

My MPI program doesn't work (hangs) when you launch processes on different nodes (hosts). In my program I use MPI_Win_allocate_shared function to allocate shared memory using RMA window. And I'm wondering what is the possible cause why my program doesn't work. Do I actually need to implement intercommunicators for that purpose? Here's the code:

offload error: cannot offload to MIC - device is not available

I am a newer in MIC.

I install and config Phi follow this link "".It seems that everything is OK.But when I run program,I get this information "offload error: cannot offload to MIC - device is not available".I source intel64 before running program.My compiler version is "Intel(R) C Intel(R) 64 Compiler for applications running on Intel(R) 64, Version Build 20151021".

I input "micctrl -s",then I get follow output:   

Effect of -g on

I'm realizing that including -g in my compile statement causes my MPI C++ program executable to link with rather than (v13/impi/ Can anyone explain exactly why that is so and whether there is necessarily an associated impact on performance?

A simple C++ code calling only MPI_Init compiled with mpicxx and with or without -g is sufficient to demonstrate the behavior.

offload error: dlopen() failed


When I try to run my offload code application, I got this error:

On the remote process, dlopen() failed. The error message sent back from the sink is /tmp/coi_procs/1/3640/load_lib/icpcoutmdD0mj: undefined symbol: _ZSt3maxIiERKT_S2_S2_
On the sink, dlopen() returned NULL. The result of dlerror() is "/tmp/coi_procs/1/3640/load_lib/icpcoutmdD0mj: undefined symbol: _ZSt3maxIiERKT_S2_S2_"
offload error: cannot load library to the device 0 (error code 20)

My compiler version is :

Suscribirse a Computación con clústeres