Parallel Computing

Suggestion about memory-access-signaling mechanism

Hello,

while I was trying to solve some particular multi-thread problem, it occurred to me that it could be solved more efficiently with special assistance from the CPU.

The situation is as follows: say one thread needs to block until the content of a particular 4-byte (or can be other size) location in the memory is changed. (It think the usefulness of this is very obvious and there is no need to give concrete examples to demonstrate it).

What are the currently available options:

mpss-3.4.1 automount autofs missing

Hello,

I am just new in mic world :-) I build really small cluster with one master node and one computing node with xeon phi in my university.

My cluster is in network with private IPs.

- master node wmpc61 works like a NIS server with automount and has IP 192.168.0.1

- computing node wmpc41 has IP 192.168.0.2

- Xeon PHI mic0 is in wmpc41 and is connected through bridge br0 and has IP 192.168.0.3

- my OS is CENTOS 7 (I disabled NetworkManager and I manage network with network.service)

Run and debug Intel MPI application on cluster. (Grid Engine)

Hello,
 

I have a problem in debugging of Intel MPI application on a cluster.

Common question: How to debug the parallel application. Console debugger idbc is not convenient at all. Are there debuggers with GUI? better for free.

I tried using eclipse, I launched program using SGE script (below), but I cannot debug with the same way.

The script for RUN.
############## RUN ##############
#!/bin/bash
#$ -S /bin/bash

cd $SGE_O_WORKDIR

. /etc/profile.d/modules.sh

Performance issue with icpc

I'm currently experiencing some performance problems using the icpcp compiler in the composer_xe_2015 suite.
I attach the very simple driver I have used below: it has been compiled with the -O3 flag, the operating system is Suse 2011 SP1
 and the computer is a single node of a intel cluster with two soket Xeon X5675.
The driver takes nearly 33 seconds to run, if I remove the last two rows
  
  std::set<double> vect;
  vect.insert(2.0);

Usage of different IPP versions. Is it possible?

Hi all!

I use IPP 7.1 in my project.

Everything was fine before one of customers installed our application on PC with application of other vendor. This old application uses Intel Open MP.

Unfortunately dynamic-linking version of IPP uses Open MP too.

So we have 2 different versions of libiomp5md.dll inside PC.

Is it possible to set for IPP correct location of libiomp5md.dll inside C++ code file (without Control Panel settings modification)?

 

Best regards,

Roman

Finalizing results hangs with large .pdb file

I am experiencing an issue with Intel VTune Amplifier.

My executable has a rather large .pdb file (~390MB). After collecting data, the resolve step hangs while resolving symbols. I can fix the problem by removing un-referenced functions from the .exe and .pdb files (Visual Studio Linker option /OPT:REF). This shrinks the .pdb file to ~140MB and the .exe file from ~22MB to ~9MB.

Is this a known issue with V-Tune? Even with very small samples; V-Tune will hang if the symbol file is large.

 

Additional Details:

static linking with IPP 8.1 -> error ipps-8.0.dll is missing

Hello,

I have a dll project in VS2013 and IPP 8.1 installed. I'm linking the project statically with the ipp 8.1 by setting the option "Use Intel IPP" to "Single-threaded Static Library" in the VS2013 project settings.

The project compiles fine and all the files are found.

However, when I try to load the dll (the final dll of the project, not any intel dll) I get an error that the dll could not be loaded because:
"ipps-8.0.dll is missing".

incorrect types for Intrinsics for Packed Test Operations in the User and Reference Guide 15.0

In the User and Reference Guide for the Intel® C++ Compiler 15.0; the the Syntax of several intrinsics in the Intrinsics for Packed Test Operations section lists incorrect formal parameter types and/or return types:

https://software.intel.com/en-us/node/524147

Specifically:

Parallel Computing abonnieren