Intel® VTune™ Amplifier XE

Calling Python Developers - High performance Python powered by Intel MKL is here!

We are introducing a Technical Preview of Intel® Distribution of Python*, with packages such as NumPy* and SciPy* accelerated using Intel MKL. Python developers can now enjoy much improved performance of many mathematical and linear algebra functions, with up to ~100x speedup in some cases, comparing to the vanilla Python distributions. The technical preview is available for everybody at no cost and includes early access to profiling support in VTune Amplifier.

GUI OS X, Application Parameters with dash dash syntax

My application, on the command line uses these arguments:

./a.out --zones 32,32,32 --gset 1 --groups 32 --quad 96 --nest DGZ --niter 10

For the OS X gui, the Application parameters input does not accept dash-dash syntax.  It appears to try to format the double dashes into a single long dash, much like Word would do.  

How do I pass double-dash arguments to my application?

On a related question, in the GUI where do you set ENV variables, like OMP_NUM_THREADS=16 ?

Thanks

Ron

Installing VTune on a VM

I am trying to get VTune installed in a virtual environ. Specifically, an Amazon EC2 instance running ubuntu. I understand not all hardware counters are exposed, so functionality may be limited (Per Thomas Wilham's response). But all I need to get is a system wide hotspot (which process->function takes most time). I tried to install the drivers and have the following error throw up.

 

 

sudo ./insmod-sep3 -r -g users

ERROR: XEN dom0 is not supported ... exiting.

grep: /proc/xen/capabilities: No such file or directory

Hot spot analysis of MPI applications

Hi There,

I am using Intel's VTune Amplifier XE 2016 integrated in Visual Studio 2015 to do hotspot analysis. Can anybody tell me how to analyse an MPI application?

What I did is as follows.

In the "Application"  field of the Amplifier Project Properties, add "C:\Program Files (x86)\IntelSWTools\mpi\5.1.2.146\intel64\bin\mpiexec.exe"

in the "Application parameters" field of the Amplifier Project Properties, add "-n 4 C:\Users\wang.j\VS2015\C30\C30\x64\Release\C30.exe"

Hot spots

Hi There,

I am using Vtune Amplifier to find hotspots in my F90 code, in Visual Studio 2005 and Windows 7 environments.  The top 4 hot spots are 

__kmp_end_split_barrier

__kmp_x86_pause

ZwYieldExecution

KiFastSystemCallRet

These functions/subroutines are definitely not in my code. My questions are

1)Why they are listed there when I can do nothing to improve my code (or I am wrong, see Q3 below)?

2)How can I find out where (in my code) they are called?

Executing actions 14 % Loading 'systemcollector-28364.sc' file tbrw call

 

 

with the command: 

amplxe-cl -collect general-exploration -data-limit=1500 -target-tmp-dir=/home/vtune cufflinks -p 24 accepted_hits.bam

I have a problem with vtune2016 ample-cl, like the following log:

amplxe: Collection stopped.

amplxe: Using result path `/home/r003ge'

amplxe: Executing actions  0 %                                                 

amplxe: Warning: The result contains a lot of raw data. Finalization may take a long time to complete.

Dynamic code (JIT) is not shown with hardware-based performance counter profiling

I use the JIT profiling API to notify VTune of the dynamically generated code segment. The dynamic code portion appears in the profiler output when running the 'hotspot' collection but does not with the 'general-exploration' collection as well as with manually configured hardware events. I use exactly the same program invocation in both cases. Other parts of the code seem to be sampled correctly. I use amplxe-gui to analyse the results.

The system is 64-bit Linux with Vtune version:

订阅 Intel® VTune™ Amplifier XE