nvprof shows error with TensorFlow - tensorflow

I am trying to run nvprof with cifar10_multigpu_train.py.
I am using following command
/home/ibm/tensorflow/third_party/gpus/cuda/bin/nvprof python cifar10_multi_gpu_train.py
It starts the application but after sometime it shows following errors and application exits.
==140659== Warning: Some profiling data are not recorded. Make sure cudaProfilerStop() or cuProfilerStop() is called before application exit to flush profile data.
======== Error: Application received signal 11
Any idea whats going wrong. File runs just fine without using nvprof.
Note: I have 0.12.0 version installed and I am on IBM PPC64le machine.

Related

WSLTTY/MINTTY Fails on Launch

After a machine restart last night my WSLTTY/MINTTY terminal crashes when I attempt to launch it. It displays an error message briefly before it crashes that reads:
note: backend error output: Assertion failed: connectRet == 0 (nix-sock.c: nix_local_connect: 67)
I'm on Windows 10, using WSL1, not WSL 2. I uninstalled WSLTTY and re-installed the most recent version, 3.5.1. The issue remained. I then uninstalled Ubuntu and re-installed the most recent version, 20.04.03 LTS. The issue still remains. I also tried installing an older version of WSLTTY/MINTTY (1.7.9) and received a slightly different error:
/bin/wslbridge: Exit 1 wslbridge error: failed to start backend process note: backend error output: wslbridge=backend.cc:44: int{anonymous}:: connectsocket(int const string &): Assertion 'connectRet==0' failed
So it seems the issue may have something to do with something called the wslbridge? Looks like there is an original version and a second version. wslbridge2.exe and wslbridge2-backend both live in wsltty/bin/ (with the newest version of WSLTTY/MINTTY).
I had WSLTTY/MINTTY configured pretty heavily and really enjoyed the way I had it set up. I'm really hoping I can get it working again and don't have to switch to a different terminal emulator or use the Windows one. Any help would be greatly appreciated.

Airflow Broken DAG error - The version of cryptography does not match the loaded shared object

Airflow 1.10.12 Seeing this error in the UI:
Broken DAG: [/home/airflow/dags/something.py] The version of cryptography does not match the loaded shared object. This can happen if you have multiple copies of cryptography installed in your Python path. Please try creating a new virtual environment to resolve this issue. Loaded python version: 2.9.2, shared object version: b'2.9'
The dags compile on the machine with no errors, but these messages appear for almost all the dags.
I have also recreated the virtualenv multiple times, but the error persists.
Anyone seen this before?
Turns out that a celery host had a scheduler running that was inserting the errors in the database. Stopped the extra scheduler and the messages went away

Creating network files in SUMO using NETCONVERT

Problem when calling netconvert in sumo:
I am trying to create my own scenario for simulation purposes.
I am using OpenStreetMaps for this.
python osmWebWizard.py
opens the browser and I select the area which I download.
netconvert --osm-files osm_bbox.osm.xml -o osm.net.xml
The error message I get is
Error: Cannot import network data without PROJ-Library. Please install packages proj before building sumo
Warning: Environment variable SUMO_HOME is not set, using built in type maps.
Quitting (on error).
My attempt to fix the problem is:
sudo apt-get install libproj*
But it seems like a dead end there and I am out of options.
Thank you.
EDIT
I have a gut feeling it has to do with libproj0 not being available anymore.

Raspberry Pi - Audio Fails After Adding RTC

I have a Raspberry Pi that I'm trying to hook-up to walkie-talkies to announce the current time every half hour plus different status updates automatically.
I had a CRON job running mpg123 that was announcing the time over the walkies perfectly, but then when I installed the drivers for this RasClock module as specified here (https://www.modmypi.com/blog/installing-the-rasclock-raspberry-pi-real-time-clock), all audio stopped working.
speaker-test says:
speaker-test 1.0.25
Playback device is default
Stream parameters are 48000Hz, S16_LE, 1 channels
Using 16 octaves of pink noise
Playback open error: -1,Operation not permitted
and mpg123 says:
[module.c:142] error: Failed to open module jack: file not found
[module.c:142] error: Failed to open module portaudio: file not found
[pulse.c:84] error: Failed to open pulse audio output: Connection refused
[nas.c:220] error: could not open default NAS server
[module.c:142] error: Failed to open module openal: file not found
[audio.c:180] error: Unable to find a working output module in this list: alsa,oss,jack,portaudio,pulse,nas,openal
[audio.c:532] error: Failed to open audio output module
[mpg123.c:897] error: Failed to initialize output, goodbye.
Now, the machine tends to freeze up a lot, too. When I tried suggestions I found online, such as adding "LD_LIBRARY_PATH=/usr/lib/mpg123" or "export LD_LIBRARY_PATH=/usr/lib:/usr/lib/mpg123" before the command, it made no difference.
What little hair I have left thanks you in advance for helping me through this.
I had the same error message with mpg123.
Before this message, I installed all these packages: mysql-server, build-essential, libmysqlclient-dev, libapache2-mod-wsgi.
I also changed group:
# usermod -G anothergroup pi
One of these two manipulations have caused my problem.
The solution in my case ?
Go in the /etc/group file and modify the line beginning with "audio" from this...
audio:x:NN:
to that...
audio:x:NN:pi
N.B.: NN is the GID. pi is the Raspberry Pi's default username.
To achieve the same result, there is also this command :
# usermod -a -G audio pi
Log out from your session and log in again.
P.S.: Could somebody add the mpg123 tag because I spent a lot of time without finding this topic, as I have exactly the same problem with mpg123 ?
I had the same issue run this command should fix it modprobe snd_bcm2835

Code Injection using mach_star

I was trying to inject code into the Finder process using mach_inject . I am currently using given the source given here
The code seems to get the PID of the finder process correctly. There is a bundle mach_inject_bundle which is to be injected.
The following error comes every time I run the injector.
mach_inject failing.. (os/kern) invalid address
error: (os/kern) invalid address
I am running the build using sudo from the terminal. I am using Snow Leopard MAC OS X 10.6.8.
I found some similar issues reported by others here . But no solution seems to work.
Can someone please help me out ??
Thanks
In the Architectures setting for the project just set x86_64.