LabVIEW

cancel
Showing results for 
Search instead for 
Did you mean: 

cRIO 9074 run FPGA compiler on windows 10 with Labview 2018

Solved!
Go to solution

Hello Rolf,

 

thank you for the reply. I have got a Fedora 39 install running now in VirtualBox. I did install the 2018FPGA-LinuxISE.iso I had to install some rpm's and at least the install seems fine. I can connect labview to the FPGA compile server and the FPGA compile server is able to connect to the CompileWorker on the Fedora install. I can see there are logfiles created on the Linux machine and I can see the status is changing on the Compile worker window. So far so good.

 

Sadly the compilation status shows an error:

LabVIEW FPGA: Some of the compilation steps were not executed.

 

I did search for some more detailed logfiles on the Linux machine. I did find the install is located in:

/USR/Local/natinst/ with some logfiles in /USR/Local/natinst/nifpgacompileworker/

 

here I can find a logfile named "workerlog.txt" I did copy the part which describes the error:

 

2023-12-04 13:15:34Z|INFO|SessionId=vgEl0ly|WorkerId=SF5nL1l|JobId=bSsO7yj|Running process: /usr/local/natinst/NIFPGA/programs/xilinx14_7/ISE/bin/lin/coregen '-b' '/usr/local/natinst/NIFPGA/jobs/bSsO7yj_SF5nL1l/core_NiLvXipFloat32Add/coregen.cmd'
2023-12-04 13:15:34Z|INFO|SessionId=|WorkerId=|JobId=|Coregen attempt 4 of 4 for xco file:/usr/local/natinst/NIFPGA/jobs/bSsO7yj_SF5nL1l/core_NiLvXipFloat32Add/ReallyLongUniqueName_ReallyLongUniqueName.xco failed with CoregenExitCode=(255, , )
2023-12-04 13:15:34Z|INFO|SessionId=vgEl0ly|WorkerId=SF5nL1l|JobId=bSsO7yj||1|CoreGenerator|DbgStdErr|CoreGen process returned error code: (255, , )
2023-12-04 13:15:34Z|ERROR|SessionId=vgEl0ly|WorkerId=SF5nL1l|JobId=bSsO7yj|The Unit Of Work (path: /usr/local/natinst/NIFPGA/jobs/bSsO7yj_SF5nL1l) did not finish properly because of an exception: The step 'CoreGenerator' errored and we were asked to quit on error.
2023-12-04 13:15:34Z|ERROR|SessionId=vgEl0ly|WorkerId=SF5nL1l|JobId=bSsO7yj|NI.LV.FPGA.CompileWorkerUI.UowQuitException: The step 'CoreGenerator' errored and we were asked to quit on error. ---> NI.LV.FPGA.CompileWorkerUI.UowQuitException: CoreGen process returned error code: (255, , )

 

 

So there is something not okay at the compiler, but I cannot figure out what. I think I need another error log. because the error 255 seems is too general.. Is there any way I could debug this problem? Would you happen to know another logfile etc?

 

reading the logfile again I see the following text:

"core_NiLvXipFloat32Add/R"  the 32 does not correspond to the text in the Worker window which shows:

"Registered Capabilities: ( 'Xilinx 14.7' 'Xilinx 14.7 (64-bit)' ) "

maybe the 32 bit versus the 64 bit could be wrong?

 

 

Hope you can help! Thanx

 

 

 

 

 

 

 

 

0 Kudos
Message 11 of 26
(1,377 Views)

According to LabVIEW 2018 FPGA Module Xilinx Compilation Tools for Linux Readme, the compilation tool for Linux only supports RHEL or CentOS.

If you try to install it on an unsupported OS, it might not behave correctly as it has not been validated.

-------------------------------------------------------
Applications Engineer | TME Systems
https://tmesystems.net/
-------------------------------------------------------
https://github.com/ZhiYang-Ong
0 Kudos
Message 12 of 26
(1,369 Views)

Fedora being the upstream, community distro of Red Hat® Enterprise Linux, you could consider it a Red Hat distribution too in most aspects. It is in some ways more experimental and leading edge than Red Hat Enterprise, since Red Hat Enterprise is selective about what they take in from Fedora to be included in the next release. So there might be minor incompatibility hassles but the basic system including package management is the same.

Rolf Kalbermatter  My Blog
DEMO, Electronic and Mechanical Support department, room 36.LB00.390
0 Kudos
Message 13 of 26
(1,358 Views)

that was my thought also. The other reason for Fedora is that there is a lot of info online. The combination of FPGA compiling etc makes it a bit harder so I started a full-panic VirtualBox install of CentOS. So If anyone can explain the error it will be very welcome.  

0 Kudos
Message 14 of 26
(1,341 Views)

@rolfk wrote:

@~Its_Me~ wrote:

Thanx for the answer!

 

Still I wonder though what is meant with this text:

 

"

Hallo,

 

I have installed Embedded Control and Monitoring Suite on Windows 10, but I couldn't install ISE 14.7. So IP Builder is not working.

 

Solution ist, that you download ISE 14.7-iso and dont use the normal installer. Go to in the  downloaded iso to "...\Products\Xilinx14.7_Core\xilinx14" and now you can install ISE 14.7.

 

I can use IP Builder on Windows 10

"


I really wouldn't know. The sentence makes to me no sense at all, both grammatically as well as technically. If what he did is as confused as that text, I'm not sure he got anything at all working.


There is an oddball way to make ISE tools work on Win 10 - we used to do it this way.  It involves dumping all of the necessary files in a specific location on the system, and then modifying some obscure system settings.  It was all a bit dodgy, but it did work.  However, we found it was still a bit unstable and wouldn't always compile correctly.

 

We can't use the Cloud Compile Service because it's blocked by our IT policy.

 

We finally got fed up with the whole thing, and stole and old computer from IT and put Windows 7 on it.  To mitigate the security risk, we don't connect it to the outside world, just our internal test network.  It's been running great.

 

The only problem we've run into is that because it is an old computer that's not connected to the internet, it sometimes forgets the day and time, and defaults back to some random day in 2009.  When that happens, the licensing has a fit and it won't compile properly.  But the error code you get is really obscure and doesn't point to a date/time licensing issue in any way - it took quite a bit of time with NI Technical Support to figure this one out.  Now if a compile fails, the first thing we check is the date/time on this Win 7 compile server.

 

-Dave

Message 15 of 26
(1,330 Views)

thank you for the support,

I do not have a win7 package. I did just try CentOS but it was 7.9 and not 7.3 so it wont intstall. The strage thing is. The install did not complain about this when using Fedora 39. I did search for CentOS 7.3 it is hard to find. Think I did find it though but it will take hours to download ....this will be a late night... 😞

0 Kudos
Message 16 of 26
(1,329 Views)

As I was waiting for the download I did fake the CentOS 7.9 changed the version name to 7.3 but this also results in the same error. I did find something strange in the log though 

 

2023-12-04 10:39:08Z|INFO|SessionId=|WorkerId=|JobId=|Done Deserializing: /usr/local/natinst/nifpgacompileworker/configuration/workerPersistedState.xml
2023-12-04 10:39:08Z|INFO|SessionId=|WorkerId=|JobId=|PersistedState successfully parsed. Key=Linux64bitEnables32bit Value=True
2023-12-04 10:39:08Z|INFO|SessionId=|WorkerId=|JobId=|PersistedState successfully parsed. Key=CoregenCacheDirectory Value=corecache
2023-12-04 10:39:08Z|INFO|SessionId=|WorkerId=|JobId=|PersistedState successfully parsed. Key=MaxServerStatusSize Value=50000000
2023-12-04 10:39:08Z|INFO|SessionId=|WorkerId=|JobId=|Initialized coregen cache directory /usr/local/natinst/NIFPGA/corecache with size 10000000000B

 

Is the size correct or must I change this?

0 Kudos
Message 17 of 26
(1,320 Views)

I did just try CentOS 7.3 and I get the same results.

Strange thing was the install script told me I had the wrong version of the os wile it is CentOS 7.3.xxxx

Can connect compile starts same error though CoregenExitCode=(255, , )

I think I must be doing something else wrong. It would be helpfull to see more errors than only the CoregenExitCode=(255, , )

anyone a idea ?

0 Kudos
Message 18 of 26
(1,300 Views)

I think I did find what is going wrong.

 

Its_Me_1-1701735082530.png

There seems to be a lib missing. Called GLIBXX_3.4.11 I did try to yum install it. but I still get the message.

I did yum install libstdc++-4.8.5-44.el7.i686.rpm

I will have a look tomorrow it is too late now. 

 

0 Kudos
Message 19 of 26
(1,287 Views)

Hey Rolfk,

 

Think I did find what's going wrong actually I think there are multiple thing which go wrong. I did find this forum:

https://forums.ni.com/t5/LabVIEW/FPGA-Xilinx-Versions-supported-by-Linux-Version/td-p/2606435/page/4 

 

With someone with the same problem. There is a rpm missing:

sudo yum install ncurses-libs.i686

When I do install this this I get rid of the "error code: (255, , ) and the lib seems to include the "GLIBXX_3.4.11" But it gives me the following error:

 

2023-12-05 22:00:24Z|INFO|SessionId=Ugez4o8|WorkerId=Sr6xuqt|JobId=dF9StXv||1|CoreGenerator|DbgStdErr|CoreGen process returned error code: (127, , /usr/local/natinst/NIFPGA/programs/xilinx14_7/ISE/bin/lin/coregen: error while loading shared libraries: libstdc++.so.6: wrong ELF class: ELFCLASS64 )

Please correct me if I am wrong, I think I made a little bit progress and now the compiler has got a lib to work with, and it is almost the correct one, I think it did find a lib but it needs a 32 bit version and not the 64 bit version. Is this correct? I switched back to Fedora 39 btw because it is easier with VirtualBox, easier to copy and past text etc. Hints are welcome! Thank you all for the help received. 

 

 

0 Kudos
Message 20 of 26
(1,253 Views)