Problem loading of Advanced hotspot results into GUI with 2015.4.0.410668

[Matevž_T_]

Hi,

Loading of Advanced hotspot results into GUI leads to an infinite loop of some sort and eventually pops up the dialog saying VTune is not responding. Most of the results seem to come up on the Summary page but the progress bar for "Collection and Platform Info" gets stuck at about 80%. amplxe-gui burns 100% CPU, apparently waiting for some mutex to be released, see strace below.

All other analysis types work ok. Displaying the results from the cli client works ok, e.g.:

amplxe-cl -R hotspots -r  ~/vtune-proj/trk-bld-host/r002ah
amplxe-cl -R hw-events -r  ~/vtune-proj/trk-bld-host/r002ah

Any idea what I could try to get around this?

Best, Matevz

[pid 16096] futex(0x28a9068, FUTEX_WAKE_PRIVATE, 1) = 0
[pid 16096] futex(0x28a9094, FUTEX_WAIT_BITSET_PRIVATE|FUTEX_CLOCK_REALTIME, 8459, {1433198657, 957790000}, ffffffff) = -1 ETIMEDOUT (Connection timed out)
[pid 16096] futex(0x28a9068, FUTEX_WAKE_PRIVATE, 1) = 0
[pid 16096] futex(0x28a9094, FUTEX_WAIT_BITSET_PRIVATE|FUTEX_CLOCK_REALTIME, 8461, {1433198658, 58034000}, ffffffff) = -1 ETIMEDOUT (Connection timed out)
[pid 16096] futex(0x28a9068, FUTEX_WAKE_PRIVATE, 1) = 0
[pid 16096] futex(0x28a9094, FUTEX_WAIT_BITSET_PRIVATE|FUTEX_CLOCK_REALTIME, 8463, {1433198658, 158400000}, ffffffff) = -1 ETIMEDOUT (Connection timed out)
[pid 16096] futex(0x28a9068, FUTEX_WAKE_PRIVATE, 1) = 0
[pid 16096] futex(0x28a9094, FUTEX_WAIT_BITSET_PRIVATE|FUTEX_CLOCK_REALTIME, 8465, {1433198658, 258828000}, ffffffff) = -1 ETIMEDOUT (Connection timed out)
[pid 16096] futex(0x28a9068, FUTEX_WAKE_PRIVATE, 1) = 0
[pid 16096] futex(0x28a9094, FUTEX_WAIT_BITSET_PRIVATE|FUTEX_CLOCK_REALTIME, 8467, {1433198658, 359052000}, ffffffff <unfinished ...>
[pid 16081] poll([{fd=3, events=POLLIN}, {fd=5, events=POLLIN}], 2, 0) = 0 (Timeout)
[pid 16081] poll([{fd=5, events=POLLIN|POLLOUT}], 1, -1) = 1 ([{fd=5, revents=POLLOUT}])
[pid 16081] writev(5, [{"&\2\2\0\271\5\300\6", 8}, {NULL, 0}, {"", 0}], 3) = 8
[pid 16081] poll([{fd=5, events=POLLIN}], 1, -1) = 1 ([{fd=5, revents=POLLIN}])
[pid 16081] recvfrom(5, "\1\1\36\251\0\0\0\0\335\1\0\0\0\0\0\0\325\0\330\1O\371\275\1\0\0\0\0\0\0\0\0", 4096, 0, NULL, NULL) = 32
[pid 16081] recvfrom(5, 0x25ff0c4, 4096, 0, 0, 0) = -1 EAGAIN (Resource temporarily unavailable)
[pid 16081] recvfrom(5, 0x25ff0c4, 4096, 0, 0, 0) = -1 EAGAIN (Resource temporarily unavailable)
[pid 16081] times({tms_utime=41698, tms_stime=42, tms_cutime=103, tms_cstime=30}) = 473453561
[pid 16096] <... futex resumed> )       = -1 ETIMEDOUT (Connection timed out)
[pid 16096] futex(0x28a9068, FUTEX_WAKE_PRIVATE, 1) = 0
[pid 16096] futex(0x28a9094, FUTEX_WAIT_BITSET_PRIVATE|FUTEX_CLOCK_REALTIME, 8469, {1433198658, 459420000}, ffffffff) = -1 ETIMEDOUT (Connection timed out)
[pid 16096] futex(0x28a9068, FUTEX_WAKE_PRIVATE, 1) = 0
[pid 16096] futex(0x28a9094, FUTEX_WAIT_BITSET_PRIVATE|FUTEX_CLOCK_REALTIME, 8471, {1433198658, 559772000}, ffffffff) = -1 ETIMEDOUT (Connection timed out)
[pid 16096] futex(0x28a9068, FUTEX_WAKE_PRIVATE, 1) = 0
[pid 16096] futex(0x28a9094, FUTEX_WAIT_BITSET_PRIVATE|FUTEX_CLOCK_REALTIME, 8473, {1433198658, 660034000}, ffffffff) = -1 ETIMEDOUT (Connection timed out)

[Robert_L_Intel1]

Hi Matevž T.

Could you upload the result set in a private message? 

BTW, are you also getting the option to send a report?  If so, it would be good to have that, too.

Thanks and regards,
-Bob

Intel® Premier Support
https://premier.intel.com/

If you prefer, for secure communications with the VTune Amplifier XE support staff, you can always submit your problem to Intel(R) Premier Support.

https://premier.intel.com/
 
It's free, secure technical support.

[Matevž_T_]

Hi Bob,

Sure, I put everything here: http://uaf-2.t2.ucsd.edu/~matevz/intel-forum-559128/

Thanks!

Matevz

[Matevž_T_]

Any ideas anyone? Can somebody confirm that loading of Advanced Hotspot results works ok for others? Matevz

[Dmitry_P_Intel1]

Hello Matevz,

I was able to download and got it working in VTune Amplifier XE 2015 Update 4.

What version do you use?

Thanks & Regards, Dmitry

[Matevž_T_]

Thanks Dmitry!

I also have 2015, update 4, see below ... we've been updating this since 2013. Would it make sense to nuke it and make a fresh install? Parallel studio is still on update 2.

Cheers, Matevz

[0850] root@phiphi ~# ls -l /opt/intel/
total 92
lrwxrwxrwx  1 root root    39 Mar 25 23:01 advisor_xe -> /opt/intel//advisor_xe_2015.1.10.380555
lrwxrwxrwx  1 root root    38 Nov 18  2014 advisor_xe_2015 -> /opt/intel//advisor_xe_2015.1.0.367266
drwxr-xr-x 15 root root  4096 Nov 18  2014 advisor_xe_2015.1.0.367266
drwxr-xr-x 16 root root  4096 Mar 25 23:01 advisor_xe_2015.1.10.380555
drwxr-xr-x  2 root root  4096 Mar 25 23:03 bin
drwxr-xr-x  3 root root  4096 Aug 16  2014 composer_xe_2013_sp1
drwxr-xr-x 13 root root  4096 Aug 16  2014 composer_xe_2013_sp1.3.174
drwxr-xr-x  3 root root  4096 Mar 25 23:03 composer_xe_2015
drwxr-xr-x 13 root root  4096 Jul 25  2014 composer_xe_2015.0.090
drwxr-xr-x 15 root root  4096 Mar 25 23:03 composer_xe_2015.2.164
lrwxrwxrwx  1 root root    16 Mar 25 23:03 composerxe -> composer_xe_2015
lrwxrwxrwx  1 root root    18 Mar 25 23:03 include -> composerxe/include
lrwxrwxrwx  1 root root    40 Mar 25 23:01 inspector_xe -> /opt/intel//inspector_xe_2015.1.2.379161
lrwxrwxrwx  1 root root    40 Nov 18  2014 inspector_xe_2015 -> /opt/intel//inspector_xe_2015.1.0.366509
drwxr-xr-x 15 root root  4096 Nov 18  2014 inspector_xe_2015.1.0.366509
drwxr-xr-x 16 root root  4096 Mar 25 23:01 inspector_xe_2015.1.2.379161
-rw-r--r--  1 root root 15022 Mar 25 23:03 intel_sdp_products.db
lrwxrwxrwx  1 root root    14 Mar 25 23:03 ipp -> composerxe/ipp
drwxr-xr-x  8 root root  4096 Apr 30 18:34 ism
lrwxrwxrwx  1 root root    14 Mar 25 23:03 lib -> composerxe/lib
drwxr-xr-x  2 root root  4096 Aug 16  2014 licenses
lrwxrwxrwx  1 root root    14 Mar 25 23:03 man -> composerxe/man
drwxr-xr-x  6 root root  4096 Apr 30 18:32 mic
lrwxrwxrwx  1 root root    14 Mar 25 23:03 mkl -> composerxe/mkl
drwxr-xr-x  4 root root  4096 Aug 16  2014 parallel_studio_xe_2013
drwxr-xr-x  6 root root  4096 Mar 25 23:03 parallel_studio_xe_2015
lrwxrwxrwx  1 root root    14 Mar 25 23:03 tbb -> composerxe/tbb
lrwxrwxrwx  1 root root    45 May 27 13:51 vtune_amplifier_xe -> /opt/intel/vtune_amplifier_xe_2015.4.0.410668
drwxr-xr-x  3 root root  4096 Nov 18  2014 vtune_amplifier_xe_2013
lrwxrwxrwx  1 root root    45 May 27 13:51 vtune_amplifier_xe_2015 -> /opt/intel/vtune_amplifier_xe_2015.4.0.410668
drwxr-xr-x 20 root root  4096 Mar 26 08:33 vtune_amplifier_xe_2015.2.0.393444
drwxr-xr-x  3 root root  4096 May 27 13:51 vtune_amplifier_xe_2015.3.0.403110
drwxr-xr-x 19 root root  4096 May 27 13:51 vtune_amplifier_xe_2015.4.0.410668

[Dmitry_P_Intel1]

Do you use VNC or X-forwarding?

Could you please try to launch standalone GUI and open the only this result?

If you see GUI freeze dialog there should be a button "collect stacks", is it possible to provide us the log?

Thanks & Regards, Dmitry

[Matevž_T_]

I use ssh X forwarding but the same happens with x2go.

I ran it as "amplxe-gui ~/vtune-proj/trk-bld-host/r006ah/" and after it complained that vtune is stuck I selected "Send report" and pasted what came out here: http://uaf-2.t2.ucsd.edu/~matevz/intel-forum-559128/report-2.txt

This is a strace: http://uaf-2.t2.ucsd.edu/~matevz/intel-forum-559128/vtune.strace.1

Cheers,
 Matevz

[Matevž_T_]

OK, it's indeed an X issue. I nfs mounted /opt/intel from the server (Centos6) to my desktop (F21) and the thing opens up ok.

It's funny opening of results works ok for all other analysis types :)

Cheers, Matevz

[Dmitry_P_Intel1]

What round trip time do you have from the host to remote machine?

I tried X-forwarding from RHEL 7 to Windows through MobaXterm_v7.2 and did not see the issue though the connection is pretty fast: 3ms.

Thanks & Regards, Dmitry

[Matevž_T_]

My RTT is only 0.4 ms but I have a colleague who connects from the east coast, probably about 90 ms. We both see the same effect, also when using X2Go remote desktop which probably hides the latency altogether.

One silly thing I noticed ... when I run from my desktop, a notification popup comes up during loading of results (titled Hotspots with a brief description) ... it does not show up when running via ssh or x2go.

And a thing that is most likely not related, /opt/intel/ism/ism doesn't come up, either ... it seems from its error log that it requires glsl through qt5 (oh joy) ... but both remote X and x2go only provide gl-1.5 support.

Best, Matevz

[Dmitry_P_Intel1]

Hello,

Just want to share good news from Matevz, that with VTune Amplifier XE 2016 Update 1 where we reimplemented summary pane the hang on summary working with GUI through X-forwarding vanished. 

So if somebody experienced with the same problem - it is worth to upgrade right away.

Thanks & Regards, Dmitry