The Altair Community is migrating to a new platform to provide a better experience for you. In preparation for the migration, the Altair Community is on read-only mode from October 28 - November 6, 2024. Technical support via cases will continue to work as is. For any urgent requests from Students/Faculty members, please submit the form linked here
Major Graphics Glitches
Hi,
Whenever I reach the final screen in the import CSV wizard (or XML wizard) the graphics go crazy and start spamming extra parts of windows over the screen. (see attched pic) I'm using rapid miner 8.1.001 and my laptop is running windows ten with 24GB RAM and 8Gb Graphics .
Any ideas how to fix this?
Many Thanks,
Jamie
0
Answers
Hi Jamie,
This sounds indeed very weird.
Unfortunately I can't see the attached screenshot. Can you edit your post and add it again?
Best,
David
Hi @j_quigley yes agreed I would very much like to see your screenshots AND your studio log (go to .RapidMiner and attach "rapidminer-studio.log".
Thanks.
Scott
Hi,
Sorry, forgot the pic! I should correctly state that tearing is the proper term.
Some things I checked - strangely my avaiable ram was only at 1024 when I first encountered the tearing.
I reinstalled rapid miner and java and it is now showing at 20GB. The problem returned however, this time when the auto model window was open.
I discovered that a way to solve the problem, i.e make the tearing disappear, is to go to VIEW in the options and select restore deafult view. But this is only a temporary solution, the tearing appears again when opening the import window.
Log file
@j_quigley this looks like RM is losing focus and is 99% of the time related to lack of memory available for RM on your machine.
Hi,
is this always happening for a particular data you are trying to import? Would it be possible to send me (privately) a copy of the data?
The only time I have seen similar things are if something goes terribly wrong behind the scenes where the GUI is doing its work. But in all of those cases I have seen it in the log. Your log is perfectly clean. Therefore my only guess is that it may be related to the data, which is why I would like to try it myself.
If that is not possible, do you have another machine where you could install RapidMiner Studio and test it as well? Or just try to import something else?
Thank you,
Marco
Hi Marco,
is this always happening for a particular data you are trying to import?
It doesn't always happen for this data set - it also happens for other data sets. The data set in the video and in the pic I sent is an XML with 25,000 entities. Each having 50 attributes.
It also happens on the auto-model section...sometimes.
Would it be possible to send me (privately) a copy of the data?
I can't send you the data, sorry.
The only time I have seen similar things are if something goes terribly wrong behind the scenes where the GUI is doing its work. But in all of those cases I have seen it in the log. Your log is perfectly clean. Therefore my only guess is that it may be related to the data, which is why I would like to try it myself.If that is not possible, do you have another machine where you could install RapidMiner Studio and test it as well? Or just try to import something else?
I will try to recreate this on another machine I have,although the spec is much lower.
Here is a video of the error being created. Although the error only occured the 4th time I did the process. Could it be the cache..or equivalent, not sure how rapid miner works in the backend. The video caputre software only captured the parent window but you can see parts of import dialogue box appear when the tearing appears.
https://youtu.be/nmlBjydIX0c
Thanks for your help,
Jamie
wow. Thanks for that screenshare. What a mess.
What's your hardware setup? Do you have a lot of other stuff running as well? I will defer to @Marco_Boeck but my hunch is that your machine is overloaded and cannot keep up with the GUI. I would be very interested to see if you have this problem on another machine.
Scott
Hi,
thank you for the video! Normally, when screen tearing similar to this happens, it is because something exploded in the GUI working thread. But a characteristic for those errors is, that while elements become visible/invisible/partially inivisible, they do not move their position in the UI. In your video, the displayed elements also randomly move their location. This is impossible if something goes wrong inside the application itself.
I have never personally experienced behavior like in your video, but the only conclusion I consider possible with that video is that on your machine the hardware acceleration is broken. Java uses your graphics card/integrated graphics for rendering the UI by default. If that part is broken for whatever reason, then I imagine something like your case would be possible. This case does seem similar to yours: https://stackoverflow.com/questions/22737535/swing-rendering-appears-broken-in-jdk-1-8-correct-in-jdk-1-7
To check if this is indeed the case, you can do it by starting RapidMiner Studio via the "RapidMiner Studio.bat" file next to the usual .exe file. First however, you need to edit the .bat file and insert something in the line that reads
You need to add "-Dsun.java2d.d3d=false " before the "-cp [...]". The replacement should read
Then, just double-click the .bat file and see if the problem persists. If it does, you can try adding yet another command so that the line then reads
Now you can try again. If neither of those changes help, then I'm afraid we are out of options. Some hardware/driver combination on your system is causing it, and there is nothing else we can do about it..
Regards,
Marco
Hi Marco,
Thanks for the detailed help!
I tried the suggestion you gave, regarding editing the bat file but i couldn't get rapidminer to launch after editing the file. Instead I created a environmental variable (J2D_D3D) with the suggested change (=False) in system properties.
This seems to have worked as no tearing in the past 10 tests of importing/reimporting troublesome XML.
Many thanks!
Jamie
If it help's I'm working on a MSI Alpha 15 A3DDk.