Intel® Many Integrated Core Architecture

setup MIC to MIC passwordless ssh


Is it possible to setup MIC to MIC passwordless ssh?

When I ssh from one MIC to another:

root@mic0 ~]# ssh mic1
The authenticity of host 'mic1 (' can't be established.
ECDSA key fingerprint is -------------------------------------------------------
Are you sure you want to continue connecting (yes/no)? ^C
[root@mic0 ~]# 

What are the exact requirements for setup?

ofed-driver compilation error, centos 6.7

I am using Centos 6.7, kernel 2.6.32-573.3.1.el6.x86_64, OFED 3.18-1. From the release information I see that mpss 3.6 is supported on RH6.7, but I do not see the corresponding ofed-driver package in mpss-3.6/ofed/modules. I tried to build it myself, but I ran into the following error trying to rpmbuild ofed-driver in mpss-3.6/src

Compiling LTTng for the Xeon Phi


I am currently in trying to compile LTTng for the Xeon Phi for the first time, I started by trying to compile Userspace-RCU which should support the Xeon Phi, as x86_64-linux-k1om is recognized. I am using GCC provided in the MPSS release, as I do not need vectorisation for it. I am using a CentOs 6.6 computer with mpss 3.4.2 installed on it. Intel Parallel studio 2015 is also installed.

I am currently trying :

CC=/opt/mpss/3.4.2/sysroots/x86_64-mpsssdk-linux/usr/bin/k1om-mpss-linux/k1om-mpss-linux-gcc ./configure --host=x86_64-k1om-linux


[SCIF user guide] Alignment and DMA transfers question


in the subsection 4.1 in the SCIF user guide it is written the following.

“Lower performance will likely be realized if the source and destination base addresses are not cacheline aligned but are

separated by some multiple of 64.”

This sentence confuses me. What do you mean by separated. An example would really help.

Could somebody clarify this case?

Thank you in advance,



Offloading argv vector


I am trying to offload the argv vector on MIC and I was wondering what would be the right path to follow.

I found out here :

that "The offload run-time dynamically determines the length of each string element and transfers each element accordingly."

by using something like:

#pragma offload target(mic) in(argv[0:argc]) (with ompt support) cannot trigger ompt_intialize() in an offload openmp environment

I'm running an Intel MIC offload openmp application using openmp runtime (with ompt support, both the CPU side and the MIC side), while on the CPU side can trigger its ompt_intialize() and get profile data, on the MIC side cannot trigger its ompt_intialize(), so that I cannot get the profile data I want. I'm wondering where the problem is.

Optimizing matrix-vector multiplication

Hello everyone,

In the application I am developing a large matrix M exists of size nxm (n >> m) and a vector x of appropriate size (I will explain what I mean by appropriate). I have to perform two matrix-vector multiplications multiple times, but not with the whole matrix M. I have included a picture to help the discussion.

S’abonner à Intel® Many Integrated Core Architecture