I want to export TensorRT model(ex. yolov5s.engine) with Jetson AGX Xavier. But when I try below page, I got some error messages.
https://github.com/ultralytics/yolov5/issues/251
[TensorRT] WARNING: Skipping tactic 3 due to oom error on requested size of 2182 detected for tactic 4.
[TensorRT] ERROR: Tactic Device request: 2182MB Available: 1536MB. Device memory is insufficient to use tactic.
How can I fix it? Or can I ignore that error messages?
You can reduce the workspace size with this CLI flag in trtexec
--workspace=N Set workspace size in MiB.
TensorRT is trying different optimization tactics during the build phase. It looks like there is a tactic that tries to use more memory than your device has available. If the conversion finds a tactic that does not use too much memory your conversion should succeed. Please let me know if this worked for you.
Related
From the linux console I run a windows console tool using:
wine console_tool.exe ....
The console tool does not involve any windows. It's output is just textual.
Some output is added repeatedly after a given delay time.
However, besides the output of the console_tool.exe I get repeatedly the following error message also interleaved with the other output:
ERROR: ld.so: object 'libgtk3-nocsd.so.0' from LD_PRELOAD cannot be preloaded (cannot open shared object file): ignored.
I already tried to export LD_PRELOAD=/usr/lib/x86_64-linux-gnu/libgtk3-nocsd.so.0 but then the only change is that the error message changes:
ERROR: ld.so: object '/usr/lib/x86_64-linux-gnu/libgtk3-nocsd.so.0' from LD_PRELOAD cannot be preloaded (wrong ELF class: ELFCLASS64): ignored.
I also attempted to apt install the :i386 version of the libgtk3-nocsd.so.0 but it doesn't seem to exists.
I don't know why a console application may trigger a gtk error message. This is beyond my knowledge.
My preferred goal would be to tell wine that the .exe does not use windows and it does not need to not use gtk for windows emulation at all. However, as this may not be possible, my second preferred goal would be to solve the gtk issue.
Can you help me achieve at least one of those goals?
I do not know what the error messages mean. However, the wineconsole command runs console executable using wine. Example:
wineconsole console_tool.exe
Reaching out to the wider group as I am totally stumped trying to install Intel MPSS 4.x for my Xeon Phi 7220P.
I followed the precise steps in this link: Intel MPSS Linux User Guide Rev 4.4.1, and did it 3 times to make sure I wasn't missing any steps or making mistakes, but keep getting the following error readout:
modprobe: WARNING: Module mic_x200_dma not found.
modprobe: WARNING: Module scif_bus not found.
modprobe: WARNING: Module vop_bus not found.
modprobe: WARNING: Module cosm_bus not found.
modprobe: WARNING: Module scif not found.
modprobe: WARNING: Module vop not found.
modprobe: WARNING: Module mic_cosm not found.
modprobe: WARNING: Module mic_x200 not found.
As a result of this, I cant do the basic MPSS commands such as micctrl -s etc nor use the Xeon Phi whatsoever.
I am running Centos 7 (862 kernel), and know its not listed in the Intel pdf, but did not think this should be causing an issue as it seems that the above kernel modules are simply not being installed seemingly from Intel MPSS - but not sure if this diagnosis is correct.
Would appreciate your help - many thanks in advance!
It is complaining about that because at one point your kernel got updated from 3.10.0-514.el7 to a later version, (it happens automatically when you do a yum update, annoying I know)
Check your kernel version by running
uname -r
When you installed/compiled all the modules they were placed into /lib/modules/3.10.0-514.el7.x86_64 which is where the source code you have exported to.
You have 2 options:
Recompile the source code to work for your current kernel version (which is a pain and has it's own problems)
Revert your host kernel back to 3.10.0-514.el7 via grub config (example here) and everything will work nicely
I too struggled with this very much in the beginning and I had to read pretty much every line of source code and spend countless hours debugging until I found out. At this point there is nothing I do not know about the Xeon Phi x100/x200.
The documentation is not bad, but it didn't cover this bare essential, which is frustrating.
I'm trying to use I2S (not I2C) with the Teensy LC and a SPH0645 digital microphone. The audio library that is included with the Teensyduino (v1.6.7 and ArduinoIDE 1.8.5) code is sufficient for my needs, but I get the following assembler error when I try to compile the Recorder example from the audio library:
/var/folders/4y/x458y3013g5d1xsgl7grxgwr0000gn/T//ccvznZoL.s: Assembler messages:
/var/folders/4y/x458y3013g5d1xsgl7grxgwr0000gn/T//ccvznZoL.s:231: Error: selected processor does not support `smull r0,ip,r3,r5' in Thumb mode
/var/folders/4y/x458y3013g5d1xsgl7grxgwr0000gn/T//ccvznZoL.s:232: Error: shifts in CMP/MOV instructions are only supported in unified syntax -- `mov ip,ip,asl r6'
/var/folders/4y/x458y3013g5d1xsgl7grxgwr0000gn/T//ccvznZoL.s:233: Error: unshifted register required -- `orr r0,ip,r0,lsr r7'
My problem apart from the error is that the message doesn't indicate which file is the culprit. So I was hoping someone maybe has some more experience with this kind of problem ;)
I have tried running the Openmdao paraboloid tutorial as well as benchmarks and I consistently receive the same error which reads as following:
[0]PETSC ERROR: ------------------------------------------------------------------------
[0]PETSC ERROR: Caught signal number 11 SEGV: Segmentation Violation, probably memory access out of range
[0]PETSC ERROR: Try option -start_in_debugger or -on_error_attach_debugger
[0]PETSC ERROR: or see http://www.mcs.anl.gov/petsc/documentation/faq.html#valgrind
[0]PETSC ERROR: or try http://valgrind.org on GNU/linux and Apple Mac OS X to find memory corruption errors
[0]PETSC ERROR: configure using --with-debugging=yes, recompile, link, and run
[0]PETSC ERROR: to get more information on the crash
---------------------------------------------------------------------
MPI_abort was invoked on rank 0 in communicator MPI_COMM_WORLD with errorcode 59.
NOTE: invoking MPI_ABORT causes MPI to kill all MPI processes.
you may or may not see output from other processes, depending on exactly when Open MPI kills them.
I don't understand why this error is occurring and what I can do to be able to run OpenMDAO without getting this error. Can you please help me with this?
Something has not gone well with your PETSc install. Its hard to debug that from afar though. It could be in your MPI install, or your PETSc install, or your petsc4py install. I suggest not install PETSc or PETSc4Py through pip though. I've had mixed success with that. Both can be installed from source without tremendous difficulty.
However, to run the tutorials you don't need to have PETSc installed. You could remove those packages and the tutorials will run correctly in serial for you
I'm trying to clone a reasonably big svn repository with git-svn and at a certain point I get a error message:
Failure loading plugin: APR: Can't create a character converter from 'UTF-8' to native encoding: Cannot allocate memory at /usr/libexec/git-core/git-svn line 5061
And sometimes a
Cannot allocate memory: zlib (compress2): out of memory: Compression of svndiff data failed at /usr/libexec/git-core/git-svn line 5061
error message. I still have ~3GB RAM free. What should I do so git-svn can utilize it?
(I'm doing this on RedHat Enterprise Linux 6.5 if that makes any difference)
From:
This error message is about the memory git is trying to allocate --
it's more than what is free. This is most likely caused by a large
file having been checked into SVN. Unfortunately, there's no easy way
to fix it (apart from buying more memory) -- you would have to remove
the large file and the commit adding it from SVN.
However try following:
Increase swap memory
Increase ulimit