Since installing Intel VTune Amplifier XE 2016 update1, I crash regularly when running the tool from within Visual Studio 2013.  I am running on WIndows 10.  I start the profiler, let it run for a while, and then stop it (doesn't matter whether I press the "stop" button or exit the program normally).  The event viewer logs have:

Faulting application name: amplxe-runss.exe, version: 0.0.0.0, time stamp: 0x561d8a1c
Faulting module name: ntdll.dll, version: 10.0.10240.16430, time stamp: 0x55c59f92
Exception code: 0xc0000374
Fault offset: 0x00000000000ea28c
Faulting process id: 0x1368
Faulting application start time: 0x01d10e849c7321c0
Faulting application path: C:\Program Files (x86)\IntelSWTools\VTune Amplifier XE 2016\bin64\amplxe-runss.exe
Faulting module path: C:\WINDOWS\SYSTEM32\ntdll.dll
Report Id: 25e07795-1769-43f7-b72d-07b350b86205
Faulting package full name: 
Faulting package-relative application ID:

I get a crash dump (about 10 MB), but I have little experience working with these to extract more information.  On the latest crash, I get the usual Windows dialog about the program having a problem and "searching for a solution".  When I dismiss the dialog, VTune continues to collect the trace information and finally displayed the results.  However, in other situations the trace is not collected, and in one case the UI where the collected information is displayed showed up as a solid black region.

Anyone else have similar problems?  Thanks.

You may receive the warning message "Windows requires a digitally signed driver" during installation on Windows* 7 and Windows* Server 2008 R2 systems.  

The VTune Amplifier hardware event-based sampling drivers (sepdrv.sys and vtss.sys) in version 2016 Update 1 are now signed with digital SHA-2 certificate key for compliance with Windows 10 requirements.  To install the drivers on Windows* 7 and Windows* Server 2008 R2 operating systems, you must add functionality for the SHA-2 hashing algorithm to the systems by applying Microsoft Security update 3033929: https://technet.microsoft.com/en-us/library/security/3033929

What it says in the topic. This appears to be essentially 100% reproducible, but it took me a while to figure out what's going on. It does not occur when the VTSS service is stopped (e.g. "net stop vtss"). 

I don't have a recent minidump (that last crash didn't produce one for some reason). Running WinDbg "!analyze -v" on an older minidump reports an "EXCEPTION_CODE: (NTSTATUS) 0xc0000094 - {EXCEPTION}  Integer division by zero" in vtss.sys.

Given that uptime seems to be a factor in this, I would hazard a guess that some time calculation inside vtss.sys is either overflowing or underflowing.

When using for android system (64bit system arm cpu) , error occured : 

C:\Program Files (x86)\IntelSWTools\VTune Amplifier 2016 for Systems\bin64>amplxe-cl -collect hotspots -target-system=android:VSEYVKW86SU8DAQK
amplxe: Using target: android:VSEYVKW86SU8DAQK
amplxe: Error: /data/local/tmp/com.intel.vtune//perfrun/bin/amplxe-runss.sh[24]: /data/local/tmp/com.intel.vtune//perfrun/bin/amplxe-runss: not executable: magic 7F45
amplxe: Error: Amplifier cannot detect Android device configuration.

I unzipped com.intel.vtune.arm.apk , extra "C:\Program Files (x86)\IntelSWTools\system_studio_2016.0.024\VTune Amplifier for Systems\target\android_arm\com.intel.vtune.arm.apk\res\raw\amplxeandroid.zip\bin32\" to another 32bit system android device, and then /data/amplxe_bin/amplxe-runss works fine.

It seems than Intel VTune Amplifier 2016 for Systems doesn't support arm 64bit android system?

Hi,

I am trying to open and view the results of general exploration analysis using VTune Amplifier 2015 for Systems, Update 3 (build 407750) on Windows 7 Enterprise 64-bit Service Pack 1. I had collected the data by running my application on a Linux machine, kernel version 3.13.0-57-generic. I had compiled my application with the debug information (-g flag). This was the command I had used to collect the data :

/opt/intel/vtune_amplifier_xe_2015.4.0.410668/bin64/amplxe-cl -collect general-exploration <full path to app> <full path to parameter files>

After collecting data using the general exploration option, I had copied the resulting r000ge/ (results) folder to my Windows machine to open and view the .amplxe file that was generated. 

I am able to view the results of the general exploration analysis in the "Summary" tab. However, I am not able to view the source code in the "Bottom-up" tab. Is there any trick that I am missing here?

Thanks in advance,

Raj

I am trying to do a system wide analysis on a Linux machine.  I get the following:

/proc/sys/kernel # /opt/intel/vtune_amplifier_xe/bin64/amplxe-cl  -collect general-exploration -knob enable-stack-collection=true  --duration unlimited --result-dir /tmp/woody
amplxe: Warning: Managed code profiling is not supported for the specified configuration.
amplxe: Warning: To profile kernel modules during the session, make sure they are available in the /lib/modules/kernel_version/ location.
amplxe: Error: Failed to generate map.events file. Data collection is interrupted.
amplxe: Internal Error
 

Any ideas what I am doing wrong?  Thanks!

Everything was working fine on 15.04 ubuntu but when I ugpraded to 15.10 vtune is broken. It won't do a simple analysis (says that 0 samples were collected)

 uname -a

Linux danielrh-desktop 4.2.0-16-generic #19-Ubuntu SMP Thu Oct 8 15:35:06 UTC 2015 x86_64 x86_64 x86_64 GNU/Linux

root@danielrh-desktop:/opt/intel/vtune_amplifier_xe/sepdk/src# lsb_release -a

No LSB modules are available.

Distributor ID:        Ubuntu

Description:        Ubuntu 15.10

Release:        15.10

Codename:        wily

and when I try to build the kernel module, it fails.

I tried to use the build command to build the kernel module and it failed with this error:

  AS [M]  /opt/intel/vtune_amplifier_xe_2016.1.0.424694/sepdk/src/sys64.o

/opt/intel/vtune_amplifier_xe_2016.1.0.424694/sepdk/src/sys64.S:28:24: fatal error: asm/dwarf2.h: No such file or directory

compilation terminated.

scripts/Makefile.build:294: recipe for target '/opt/intel/vtune_amplifier_xe_2016.1.0.424694/sepdk/src/sys64.o' failed

make[2]: *** [/opt/intel/vtune_amplifier_xe_2016.1.0.424694/sepdk/src/sys64.o] Error 1

Makefile:1398: recipe for target '_module_/opt/intel/vtune_amplifier_xe_2016.1.0.424694/sepdk/src' failed

make[1]: *** [_module_/opt/intel/vtune_amplifier_xe_2016.1.0.424694/sepdk/src] Error 2

make[1]: Leaving directory '/usr/src/linux-headers-4.2.0-16-generic'

Makefile:177: recipe for target 'default' failed

make: *** [default] Error 2

so I followed the directions to make sure packages were installed: they were

root@danielrh-desktop:/opt/intel/vtune_amplifier_xe/sepdk/src#         apt-get install build-essential

Reading package lists... Done

Building dependency tree       

Reading state information... Done

build-essential is already the newest version.

The following package was automatically installed and is no longer required:

  libdw1

Use 'apt-get autoremove' to remove it.

0 upgraded, 0 newly installed, 0 to remove and 3 not upgraded.

root@danielrh-desktop:/opt/intel/vtune_amplifier_xe/sepdk/src# apt-get install linux-headers-`uname -r`

Reading package lists... Done

Building dependency tree       

Reading state information... Done

linux-headers-4.2.0-16-generic is already the newest version.

The following package was automatically installed and is no longer required:

  libdw1

Use 'apt-get autoremove' to remove it.

Can someone help me get this operational on my standard ubuntu machine? Thanks!

Recently I need to know exactly the data communication(exactly how much and when) between the Host and MIC while running an offload application.

Unlike Nvidia, Intel doesn't provide a tool like NV Profiler.

So I guess maybe Vtune Amplifier XE can do this job. But unfortunately I get to know that when Amplxe only analyze MIC's performance while I analyse an offload application.

So I come to get help from the forum. Is there anyone who can help me? 

Hello,

while writing a simple test to figure out how to add user tasks and such I stumbled upon the following questions:

1. Which ITT functions are actually supported by VTune? Things I have found to work are tasks, domains, ids, string handles. Things that do not work for me are e.g. __itt_task_group, __itt_marker, __itt_metadata_add, __itt_relation_add . I digged a bit through the .trace files generated by VTune and all the calls I make to functions like _itt_task_group actually end up in the collected data, but they do not appear anywhere in the GUI. Also stuff like feeding the parent id to a task does not seem to make any difference. Relevant links that I found: 

  - Task Groups: https://software.intel.com/sites/products/documentation/gpa/13.4/win/Tas...

  - Markers: http://gpa.helpmax.net/en/intel-graphics-performance-analyzers-help/inte... 

  - Tracks and track groups: http://gpa.helpmax.net/en/intel-graphics-performance-analyzers-help/inte...

2. Is there really no way to add any data to a task besides putting all the information into the task name? Is there a way to group tasks just like task groups promise to do?

3. What's the difference between events (e.g. see https://software.intel.com/zh-cn/blogs/2011/03/24/400006948 ) and tasks (e.g. https://software.intel.com/en-us/articles/a-simplest-example-to-use-task... )? They seem to be pretty much the same.

4. Is there a way to export the custom data collected with VTune so that I can analyze it with some script later on? Of course I can just export myself, but it would be nice to have a coherent system.

Thank you,

-Matthias

Customer wants to build  vtune sampling driver built for their platform. And they don't use Windriver Workbench, but, use the command line “configure” script to configure our project. Cusotemr followed the guidelines listed in the installation guide and  included the layer definition as suggested, but, I am getting build errors while building in their project.

Part of the log build failure below shows that it fails on a unrelated component (openssh). Any feedback on why the customer is seeing this issue?

Error Log

Makefile:150: recipe for target 'ssh' failed

| make: *** [ssh] Error 1

| make: *** Waiting for unfinished jobs....

| x86_64-windriver-linux-ld: unrecognized option '-Wl,-O1'

| x86_64-windriver-linux-ld: use the --help option for usage information

| x86_64-windriver-linux-ld: unrecognized option '-Wl,-O1'

| x86_64-windriver-linux-ld: use the --help option for usage information

| Makefile:174: recipe for target 'ssh-keyscan' failed

| make: *** [ssh-keyscan] Error 1

| Makefile:168: recipe for target 'ssh-keysign' failed

| make: *** [ssh-keysign] Error 1

| x86_64-windriver-linux-ld: unrecognized option '-Wl,-O1'

| x86_64-windriver-linux-ld: unrecognized option 'x86_64-windriver-linux-ld-Wl,-O1: use the --help option for usage information

| '

| x86_64-windriver-linux-ld: use the --help option for usage information

| Makefile:171: recipe for target 'ssh-pkcs11-helper' failed

| make: *** [ssh-pkcs11-helper] Error 1

| Makefile:159: recipe for target 'ssh-add' failed

| make: *** [ssh-add] Error 1

| ERROR: oe_runmake failed

| WARNING: /nobackup/skreddy/cisco-xr-wr7-1102/target-n9000-gdb/bitbake_build/tmp/work/core2-64-wrs-linux/openssh/6.6p1-r0/temp/do_compile/run.do_compile.9605:1 exit 1 from

|   exit 1

| ERROR: Function failed: do_compile (log file is located at /nobackup/skreddy/cisco-xr-wr7-1102/target-n9000-gdb/bitbake_build/tmp/work/core2-64-wrs-linux/openssh/6.6p1-r0/temp/do_compile/log.do_compile.9605)

$<50>ERROR: Task 520 (/nobackup/skreddy/cisco-xr-wr7-1102/target-n9000-gdb/layers/oe-core/meta/recipes-connectivity/openssh/openssh_6.6p1.bb, do_compile) failed with exit code '1'

$<50>NOTE: Tasks Summary: Attempted 4559 tasks of which 94 didn't need to be rerun and 1 failed.

$<50>NOTE: Build completion summary:

$<50>NOTE:   From shared state: 0

$<50>NOTE:   From scratch: 344

Waiting for 0 running tasks to finish:

Summary: 1 task failed:

  /nobackup/skreddy/cisco-xr-wr7-1102/target-n9000-gdb/layers/oe-core/meta/recipes-connectivity/openssh/openssh_6.6p1.bb, do_compile

Summary: There were 12 WARNING messages shown.

Summary: There was 1 ERROR message shown, returning a non-zero exit code.

make: *** [fretta-image-base] Error 1

It seems that "VTune Amplifier XE 2016 Update1" consumes sample buffer even while it is paused

"The specified data limit of 500 MB is reached. Data collection is stopped." even before I Resume collection, and be able to start profiling.

I would like to delay sample collection untill the point my application is fully loaded and initialized, but have to specify too big sample data buffer over 1GB which is extremely slow at the later processing time.

Error when getting the report form the command line by filtering by column name,

(1)Profile my application "amplxe-cl -collect hotspots -target-pid <myapp_pid> -r <resultFile> "

(2) When i want to view the report listing only specific columns using "amplxe-cl -R top-down -r <resultFile> --column=Module", i get the error

Unknown option: --column=Module

The report has 'Module' as one of the column. 

After collecting results in the command line for a basic hotspots analysis,

"amplxe-cl -collect hotspots -r r000hs -knob enable-gpu-usage=true -knob enable-gpu-runtimes=true MyWorkload.exe",

where MyWorkload.exe contains openCL kernels running on Intel GPU, how do we generate a report that contains GPU statistics like EU array active, idle, stalled? The gpu-computing-tasks report is not showing this kind of data after the following command:

"amplxe-cl -report gpu-computing-tasks -r r000hs -report-output MyReport.csv -format csv -csv-delimiter comma".

I don't want to use the GUI since I'm looking into calling VTune from a batch file using the command line.

Any help is appreciated. Thanks in advance.

I remember that there once were analysis like snb-* for sandy-bridge. Now that is gone. What if I want to measure port saturation? Thank you so much.

Yuanfang

Trying to use Vtune with matrix.mic like described. With amplxe-gui and amplxe-cl -target-system=mic-native:0 I get:

amplxe: Warning: To enable hardware event-base[d] sampling, VTune Amplifier has disabled the NMI watchdog timer. The watchdog timer will be re-enabled after collection completes.
abstract_Reserve_PMU ret error
abstract_Release_PMU ret error
Invalid error code
amplxe: Collection failed.
amplxe: Internal Error

Regards

 Josef

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. Click here to register and download. For any questions, please jump onto the user forum. 

Hi,

I encountered a strange behavior on Xeon Phi, when I check times of execution of my program (simple program to generate mandelbrot image) I see that application written in MPI was about 10 times faster than application written in OpenMP. It's very strange for me so I checked a simple program only with one empty loop, and this program execution was 700 (OpenMP) seconds to 900 (MPI), but when I added any math calculation inside the loop OpenMP was as fast as MPI implementation or even slower. So right know I don't know exactly what to think about it, because OpenMP should be faster than MPI or at least as fast as MPI.

I think it's not a problem with transfer because transfer is also included in times..

Anybody has an idea what is wrong with this?

Hi

I'd like to restrict analysis is 3 ways. By this, I mean the collection of data. Nit the resulting report

1) I have a DLL that is loaded by an external process. I only want to analyse the code in the DLL. The overhead of the external program is so large that Amplifier grind to a halt. I'm not interested in what the main program does (yet).

2) I'd like to analyse only parts of the DLL. I've implemented the itt API. But with limited results. I've tried pause/resume, tasks and events.

3) I'd like to restrict analysis to only one thread and exclude analysis from everything else. Which itt API call is the appropriate one.

Any help gratefully received.

Steve 

Hi there

I am running Visual Studio 2013 on Windows 10, and Intel VTune Amplifier XE 2015.

I built my application in release mode (full optimization and debug symbols), I even enabled the Microsoft Debug Symbols server, but when I try to perform a basic hotspot analysis (or any analysis for the matter...) the exe is never launched, and eventually I get an error.

and at the end I get this

How can I resolve this? Any help is deeply appreciated.

[ubuntu@HomeNAS-ubuntu ~/Trash]$ amplxe-cl  --version
Intel(R) VTune(TM) Amplifier XE 2016 Update 1 (build 434111) Command Line Tool
Copyright (C) 2009-2015 Intel Corporation. All rights reserved.

[ubuntu@HomeNAS-ubuntu ~/Trash]$ amplxe-cl -collect-with runsa ./a.out
amplxe: Collection started. To stop the collection, either press CTRL-C or enter from another console window: amplxe-cl -r /home/ubuntu/Trash/r003runsa -command stop.
amplxe: Collection stopped.
amplxe: Using result path `/home/ubuntu/Trash/r003runsa'
amplxe: Executing actions 100 % done
amplxe: Error: 0x40000025 (Inapplicable report) -- The report 'summary' is not applicable to the result /home/ubuntu/Trash/r003runsa/r003runsa.amplxe.

What did I do wrong? Is it a bug? Thanks.