I have 29 Simulink/Matlab Test. It has a lot of different reference models. Before running a 20 second simulation , it has to load all reference models and create a lot of simulation artifacts in a work folder. A lot of reference model are shared in-between test.
When running one test at a time, I have no issue, all simulation artifact are created and used to run the various simulation. Everything Passes.
When running it all via parallel processing. I have a issue.Some simulation artifact are not built or missing, hence my simulation fails even before running.But surprisingly, not all 29 of them fail. It actually random,last time it was 17, another time it was 22. And it even ran once with 0 fail.
Another note, I only have this issue when running it on a self-hosted computer on Azure-Pipelines for CI purposes.
I would like to fix this issue and reproduce stable test pass/fail results of one at a time run, but on parallel process run. How would I do that?
Error:
2020-11-03T03:16:27.1083996Z Making simulation target "Foo_src_sfun", ...
2020-11-03T03:16:27.1084227Z
2020-11-03T03:16:27.1084361Z
2020-11-03T03:16:27.1084502Z
2020-11-03T03:16:27.1084789Z Microsoft (R) Program Maintenance Utility Version 14.00.24210.0
2020-11-03T03:16:27.1085188Z Copyright (C) Microsoft Corporation. All rights reserved.
2020-11-03T03:16:27.1085441Z
2020-11-03T03:16:27.1085815Z NMAKE : fatal error U1052: file 'Foo_src_sfun.mak' not found
2020-11-03T03:16:27.1086175Z Stop.
2020-11-03T03:16:27.1089399Z ================================================================================
2020-11-03T03:16:27.1089936Z Error occurred in TestSim/testSim(File=test_FooTest1_slx) and it did not run to completion.
2020-11-03T03:16:27.1090308Z
2020-11-03T03:16:27.1090497Z ---------
2020-11-03T03:16:27.1090720Z Error ID:
2020-11-03T03:16:27.1090946Z ---------
2020-11-03T03:16:27.1091254Z 'Slvnv:simcoverage:SimulationFailed'
2020-11-03T03:16:27.1091481Z
2020-11-03T03:16:27.1091669Z --------------
2020-11-03T03:16:27.1091919Z Error Details:
2020-11-03T03:16:27.1092186Z --------------
2020-11-03T03:16:27.1092419Z Error using cvsim
2020-11-03T03:16:27.1092659Z Simulation failed
2020-11-03T03:16:27.1092864Z
2020-11-03T03:16:27.1093112Z Error in testRunner (line 145)
2020-11-03T03:16:27.1093477Z [cvdo, simOutRes] = cvsim(testObj,paramStruct) ;
2020-11-03T03:16:27.1093765Z
2020-11-03T03:16:27.1094034Z Error in TestSim/testSim (line 30)
2020-11-03T03:16:27.1094373Z [cvdo, simOutRes, ErrLog] = testRunner(File,20);
2020-11-03T03:16:27.1094638Z
2020-11-03T03:16:27.1094830Z Caused by:
2020-11-03T03:16:27.1095168Z Error using autobuild_kernel>autobuild_local (line 219)
2020-11-03T03:16:27.1095612Z Unable to create mex function 'Foo_src_sfun.mexw64'
2020-11-03T03:16:27.1096006Z required for simulation.
2020-11-03T03:16:27.1096427Z ================================================================================
Update:
I found that I have also another kind of error, leads pretty much to same result.
2020-11-03T03:18:36.1668328Z Making simulation target "Foo2_src_sfun", ...
2020-11-03T03:18:36.1668601Z
2020-11-03T03:18:36.1668735Z
2020-11-03T03:18:36.1669087Z 'Foo2_src_sfun.bat' is not recognized as an internal or external command,
2020-11-03T03:18:36.1669483Z operable program or batch file.
2020-11-03T03:18:36.1669685Z
2020-11-03T03:18:36.1669892Z >>Removing MiL paths...
2020-11-03T03:18:36.1670104Z >>Done
I made a runSingleTest() that I run before my parallel run. Before running it creates all required model reference mexw64 files in the **/work/sim_artifact folder.
Hence when the parallel run they don't need to create any new files, they either use whats already there or update the files.
I have been having no issue since that change. Just a longer run time because of that repetitive test.
Related
I have a Verilog project that makes use of a testbench written in SystemVerilog and a few imports/exports of functions through the DPI-C interface.
When attempting to simulate, I get an xsim error (as far as I can tell) and the simulation stops. I have struggled with this issue for a while, and there is no specific info given in the Vivado terminal together with the error.
The exact error received is:
ERROR: [USF-XSim-62] 'elaborate' step failed with error(s) while executing 'path/to/proj/<proj_name>/<proj_name>.sim/sim_1/behav/xsim/elaborate.bat' script. Please check that the file has the correct 'read/write/execute' permissions and the Tcl console output for any other possible errors or warnings.
ERROR: [Vivado 12-4473] Detected error while running simulation. Please correct the issue and retry this operation.
launch_simulation: Time (s): cpu = 00:00:01 ; elapsed = 00:00:06 . Memory (MB): peak = 1412.562 ; gain = 0.000
ERROR: [Common 17-39] 'launch_simulation' failed due to earlier errors.
Not only is the file created by Vivado (so it should have permissions to access it), but this error appears and disappears at random. For example, modifying the testbench file and then modifying it back to original (to force recompilation) sometimes allows the simulation to run, seemingly at random.
What is even more confusing is that there are some "safe states" of the testbench code that allow the simulation to always run. My initial hunch was that it was related to the DPI-C functions, but I tried altering the files in many ways and I didn't find any obvious correlation.
Mentions
I am using Vivado 2021.2
I have implemented an automated TCL script to compile the C files using xsc and insert them into the project directory
I am using xelab command line arguments to link the C code compiled with xsc: -sv_root path/to/xsc -sv_lib dpi.
Update
When I run xelab by itself (with -v), it gives the following output (after static elaboration and simulation data flow analysis passed):
SDG Object Count: 1284, SDG Object Memory Usage: 133 KB.
Time Resolution for simulation is 1ps
Compiling package std.std
ICR Memory Usage: 491KB, 8192KB
Compiling package xil_defaultlib.$unit_tb_sv
ICR Memory Usage: 491KB, 8192KB
Compiling module xil_defaultlib.decoder_interface
ICR Memory Usage: 500KB, 8192KB
Compiling module xil_defaultlib.lut_biases
ICR Memory Usage: 584KB, 8192KB
Compiling module xil_defaultlib.saturate_default
ICR Memory Usage: 758KB, 8192KB
Compiling module xil_defaultlib.variable_nodes_default
ICR Memory Usage: 1832KB, 8192KB
Compiling module xil_defaultlib.check_nodes_default
INFO: [XSIM 43-4009] "abs_prev_proc_elem", written at line 26 in file "D:/Projects/Matlab/NN_BP_BCH/nn-min-sum-decoding/hardware/check_nodes.v", has also been read in this always_comb/always_latch block and is not added to the sensitivity list.
INFO: [XSIM 43-4009] "reg_min", written at line 30 in file "D:/Projects/Matlab/NN_BP_BCH/nn-min-sum-decoding/hardware/check_nodes.v", has also been read in this always_comb/always_latch block and is not added to the sensitivity list.
INFO: [XSIM 43-4009] "reg_sign", written at line 31 in file "D:/Projects/Matlab/NN_BP_BCH/nn-min-sum-decoding/hardware/check_nodes.v", has also been read in this always_comb/always_latch block and is not added to the sensitivity list.
INFO: [XSIM 43-4009] "temp_reg", written at line 52 in file "D:/Projects/Matlab/NN_BP_BCH/nn-min-sum-decoding/hardware/check_nodes.v", has also been read in this always_comb/always_latch block and is not added to the sensitivity list.
ICR Memory Usage: 6896KB, 8192KB
Compiling module xil_defaultlib.interm_layer
ICR Memory Usage: 6902KB, 8192KB
Compiling module xil_defaultlib.llr_to_out_default
ICR Memory Usage: 6927KB, 8192KB
Compiling module xil_defaultlib.out_layer_default
ICR Memory Usage: 7674KB, 8192KB
Compiling module xil_defaultlib.decoder_top_default
ICR Memory Usage: 8281KB, 16384KB
Compiling module xil_defaultlib.tb
child killed: unknown signal
Are there any reasons why this is not a good idea? I ask because I constantly experience very, very inconsistent results. For example, while setting up my GH Actions over the last few days, I must have run at least 200 workflows. However, for the first time ever, I am now seeing this error:
Run ruby/setup-ruby#v1
with:
ruby-version: 3.0.2
bundler-cache: true
bundler: default
working-directory: .
cache-version: 0
env:
BUNDLE_GEMS__CONTRIBSYS__COM: ***
ImageOS: ubuntu20
Modifying PATH
Entries added to PATH to use selected Ruby:
/opt/hostedtoolcache/Ruby/3.0.2/x64/bin
Downloading Ruby
https://github.com/ruby/ruby-builder/releases/download/toolcache/ruby-3.0.2-ubuntu-20.04.tar.gz
Took 0.71 seconds
Extracting Ruby
/usr/bin/tar -xz -C /opt/hostedtoolcache/Ruby/3.0.2 -f /home/ubuntu/actions-runner-2/_work/_temp/7d0937cf-69b1-4c73-b1bd-7386fca820a2
/usr/bin/tar: x64/lib: Cannot utime: No such file or directory
/usr/bin/tar: Exiting with failure status due to previous errors
Took 0.52 seconds
Error: The process '/usr/bin/tar' failed with exit code 2
I have absolutely no clue whatsoever why this would be presenting itself. If I re-run the same workflow, the error goes away. I'm not sure if this is because one runner is conflicting with another while trying to access the /opt/hostedtoolcache/ directory or something else.
Here's the exact same job re-run without any issues:
I am trying to run my first UMAT subroutine with an ABAQUS job.
Fortran compiler seems to be linked to ABAQUS, since Abaqus command
abaqus verify -user_std
results to pass the check
Abaqus/Standard with user subroutines
...PASS
Continuing...
But trying to run the subroutine, an error occurs. The message file of the job reports:
Abaqus/Standard Analysis exited with an error - Please see the message file for possible error messages if the file exists.*
No message file (.msg) is produced by the software for the job.
The .log file reports the following:
Analysis initiated from SIMULIA established products
Abaqus JOB JOB NAME
Abaqus 2020
Abaqus License Manager checked out the following licenses:
Abaqus/Standard checked out 5 tokens from Flexnet server COMPUTER NAME
<1019 out of 1024 licenses remain available>.
Begin Compiling Abaqus/Standard User Subroutines
21/09/2021 12:15:52
Intel(R) Fortran Intel(R) 64 Compiler Classic for applications running on Intel(R) 64, Version 2021.3.0 Build ***************
Copyright (C) 1985-2021 Intel Corporation. All rights reserved.
End Compiling Abaqus/Standard User Subroutines
Begin Linking Abaqus/Standard User Subroutines
Creazione della libreria standardU.lib e dell'oggetto standardU.exp
libirc.lib(fast_mem_ops.obj) : warning LNK4210: .CRT section exists; there may be unhandled static initializers or terminators
End Linking Abaqus/Standard User Subroutines
21/09/2021 12:16:00
Begin Analysis Input File Processor
21/09/2021 12:16:00
Run pre.exe
21/09/2021 12:16:05
End Analysis Input File Processor
Begin Abaqus/Standard Analysis
21/09/2021 12:16:05
Run standard.exe
21/09/2021 12:16:06
Abaqus Error: Abaqus/Standard Analysis exited with an error - Please see the
message file for possible error messages if the file exists.
Begin SIM Wrap-up
21/09/2021 12:16:06
Run SMASimUtility.exe
21/09/2021 12:16:06
End SIM Wrap-up
Abaqus/Analysis exited with errors
I have checked some previous questions on the web, but it's hard to get further for me.
Any contribution is welcome.
Try adding the following to your environment file:
link_sl='LINK /NODEFAULTLIB:LIBCMT.LIB /dll /def:%E /out:%U %F %A %L %B'
Source
Our Jenkins server has been successfully building projects for years using Powershell to run a cake script. Out of the blue today, we started getting the following error on multiple projects:
12:10:46 Running build script...
12:10:46
12:10:46 Error: Unknown command 'Default'.
12:10:46
12:10:46 .\scripts\build.cake -target=Default -configuration=Release
12:10:46 -verbosity=Verbose -environment=dev -certificatePath=C:/Program Files
12:10:46 (x86)/xxxxx/Verisign Certificate -certificateName=xxxxxxxx_cert.pfx
12:10:46 -certificatePassword=xxxxxxxxxxx
12:10:46 ^^^^^^^ No such command
12:10:46
12:10:47 Build step 'Windows PowerShell' marked build as failure
The build.ps1 script has not changed. The build.cake script has not changed. Jenkins was not updated since the last successful builds nor any of the add-ins.
If I remove "-target=Default" from the argument list, I just get the same error on "-configuration=Release" indicating unknown command: "Release".
Has anyone else experienced this?
The possible root cause of this failure is the cake version update (0.38.x --> 1.0)
You have two ways to fix this issue
Change the Argument Syntex - arguments should always be called with multi-dash syntax (e.g. --target=Foo) instead of (e.g. -target=Foo)
Refer: https://cakebuild.net/docs/getting-started/upgrade
Pinning the cake version
Specified the version of Cake in the tools/packages.config file
Refer: https://cakebuild.net/docs/writing-builds/reproducible-builds/pinning-cake-version
I'm having a problem where jasmine-node silently fails if unhandled exceptions happen in a test.
If I run a single file, everything is OK and I get the expected jasmine output:
./node_modules/jasmine-node/bin/jasmine-node spec/unit/accessControlSpec.js
Finished in 0.011 seconds
4 tests, 6 assertions, 0 failures, 0 skipped
But, if I run all specs in a folder, it fails silently.
./node_modules/jasmine-node/bin/jasmine-node spec/unit
Tried --verbose and --captureExceptions but no luck.
In this specific case, some code inside a test was calling a method that didn't exist.
So, turns out the problem is that I'm not calling the correct command because I didn't install jasmine-node globally.
The correct way is:
node ./node_modules/jasmine-node/lib/jasmine-node/cli.js ./spec/unit
This is further described here: Command Line Usage