06-14-2019 04:36 AM
Ubuntu 18.04 Kernel 4.15 with multiple monitors attached.
Vivado crashes when launching the gui with the following error:
java.lang.IllegalArgumentException: Window must not be zero
Same issue as with 2018.3. No error when only one display is enabled. 2018.2 and earlier versions work fine.
06-14-2019 06:43 AM
Hello @lematthias .
I have reported this again and have highlighted that there is still an issue in relation to unable to start vivado gui when Multiple monitors are used.
Please leave this with me and I will update you once there are any updates.
06-26-2019 05:18 PM
It isn't the most elegent, but if you ssh into your own box with X11 forwarding, you can launch the GUI and have it X11 forward to yourself. When you do it that way it thinks there is only one monitor. I haven't tried this yet with 2019.1, but that worked for me in 2018.3.
07-05-2019 06:44 AM
I'm facning the same problem 2018.3 and 2019.1. Is there any updates regarding this issue?
07-08-2019 08:40 AM
I'm having the same problem. Ubuntu 18.04, 3 monitors, NVidia 430 driver. Java errors on 19.1 won't start the gui.
07-17-2019 02:58 AM - edited 07-17-2019 02:58 AM
07-17-2019 03:03 AM
Im using this script to overcome this problem. I found this solution somewhere here in the Xilinx Forums!
Xephyr :1 -screen 1920x1024&
08-30-2019 12:27 AM
I workaround this by using autorandr (https://github.com/wertarbyte/autorandr). I've set up a configuration which temporarily turns all displays except one off. I start Vivado, then use another autorandr config to turn displays on again.
Relatively simple workaround for a really annoying bug.
10-13-2019 08:37 PM
I would also be keen to know when they fix this issue. It's a bit of a time sink every time vivado crashes and I need to restart it.
I'm using the workaround with xrandr to turn my monitors off. However, turning my monitors on with xrandr is problematic. I can't script the monitors turning on, as the xrandr can fail in several different ways, and applying fixes non-selectively just fixes a problem on one monitor and breaks a different monitor. It is a time costly exercise (5-15 minutes) of power cycling monitors and running different xrandr settings repeatively to get all the monitors back to normal.
Consequently, I would be very keen to know when this bug is fixed.
10-13-2019 10:37 PM
Well, autorandr is simple to use:
- on your DE set a mode where only one monitor is active
- run: autorandr --save startvivado
- set a mode where every display is active
- run: autorandr --save preferred
Then you can launch Vivado as:
autorandr startvivado; vivado & ; sleep 20; autorandr preferred
Workaround works fine. For the Vivado devs: this workaround indicates me that simply removing the failing code part would fix this really annoying bug.
10-14-2019 01:20 AM
Unfortunately, it's not a workaround that works for everyone.
I tried autorandr and it doesn't work on my system. I took a quick look at the shell debugging and it looks like there's output formatting issues with what autorandr expects and some of the command arguments aren't being accepted. I also looked at what autorandr does and it is doing the things that I'm doing manually in xrandr.
Fundamentally my xrandr doesn't run reliably. My xrandr doesn't reliably change settings, as it seems to start having a conflict with the window manager making additional changes in response to xrandr's changes and the monitor settings ending up corrupted. Just 10 minutes ago, my X11 session crashed from turn on the monitors again, and this is a normal thing for me while using vivado, to have to reopen all my documents/webpages/applications after starting vivado.
It would be good if they can fix the issue with vivado, as for some people, it's a major daily problem when using vivado.
10-14-2019 01:37 AM
Unfortunately, it's not a workaround for me as well !
As you said, hope they fix it with Vivado as it is a major daily problem for me.