ovito 3.0.0-dev 466 Crashed when I rendered a complete animation

Quote from Zheng Wang on September 3, 2019, 7:44 amHi all.
I met many crashes of ovito 3.0.0-dev 466 when I rendered a complete animation.
The RAM usage increases continuously and will soon reached 96% when I rendered a complete animation, then the program will crashed. Not only for a video file but also for a series of images.
But there's no problem in ovito 2.9 and RAM usage alway kept under 16%.
If this is a bug? How to solve it?
Info of My Laptop:
██████████████████ ████████ OS: Manjaro Linux x86_64
██████████████████ ████████ Host: 81GD Lenovo XiaoXin Chao7000-15ARR
████████ ████████ Kernel: 5.3.0-1-MANJARO
████████ ████████ ████████ Uptime: 4 hours, 22 mins
████████ ████████ ████████ Packages: 2058 (pacman)
████████ ████████ ████████ Shell: bash 5.0.9
████████ ████████ ████████ Resolution: 1920x1080
████████ ████████ ████████ DE: KDE
████████ ████████ ████████ WM: KWin
████████ ████████ ████████ WM Theme: McMojave-light
████████ ████████ ████████ Theme: Breeze [KDE], Breath [GTK2/3]
████████ ████████ ████████ Icons: breath [KDE], deepin [GTK2/3]
Terminal: konsole
Terminal Font: Noto Mono 11
CPU: AMD Ryzen 5 2500U with Radeon Vega Mobile Gfx (8) @ 2.000GHz
GPU: AMD ATI Radeon RX 550/550X
GPU: AMD ATI Radeon Vega Series / Radeon Vega Mobile Series
Memory: 1873MiB / 18979MiB======= System info =======
Date: 周二 9月 3 13:29:58 2019
Application: Ovito 3.0.0-dev466
OS: Linux
uname output: 5.3.0-1-MANJARO #1 SMP Mon Aug 26 11:38:15 UTC 2019 x86_64 unknown GNU/Linux
LSB output: ManjaroLinux "Manjaro Linux" 18.1.0-rc7
Architecture: 64 bit
Floating-point size: 64 bit
Qt version: 5.12.4
Command line: /opt/ovito/bin/ovito
======= OpenGL info =======
Version: 4.5
Profile: core
Alpha: 0
Vendor: X.Org
Renderer: Radeon 500 Series (POLARIS12, DRM 3.33.0, 5.3.0-1-MANJARO, LLVM 8.0.1)
Version string: 4.5 (Core Profile) Mesa 19.1.5
Swap behavior: other
Depth buffer size: 24
Stencil buffer size: 8
Shading language: 4.50
Geometry shaders supported: yes
Using deprecated functions: no
Using point sprites: yes
Using geometry shaders: yes
Context sharing enabled: yes
Hi all.
I met many crashes of ovito 3.0.0-dev 466 when I rendered a complete animation.
The RAM usage increases continuously and will soon reached 96% when I rendered a complete animation, then the program will crashed. Not only for a video file but also for a series of images.
But there's no problem in ovito 2.9 and RAM usage alway kept under 16%.
If this is a bug? How to solve it?
Info of My Laptop:
██████████████████ ████████ OS: Manjaro Linux x86_64
██████████████████ ████████ Host: 81GD Lenovo XiaoXin Chao7000-15ARR
████████ ████████ Kernel: 5.3.0-1-MANJARO
████████ ████████ ████████ Uptime: 4 hours, 22 mins
████████ ████████ ████████ Packages: 2058 (pacman)
████████ ████████ ████████ Shell: bash 5.0.9
████████ ████████ ████████ Resolution: 1920x1080
████████ ████████ ████████ DE: KDE
████████ ████████ ████████ WM: KWin
████████ ████████ ████████ WM Theme: McMojave-light
████████ ████████ ████████ Theme: Breeze [KDE], Breath [GTK2/3]
████████ ████████ ████████ Icons: breath [KDE], deepin [GTK2/3]
Terminal: konsole
Terminal Font: Noto Mono 11
CPU: AMD Ryzen 5 2500U with Radeon Vega Mobile Gfx (8) @ 2.000GHz
GPU: AMD ATI Radeon RX 550/550X
GPU: AMD ATI Radeon Vega Series / Radeon Vega Mobile Series
Memory: 1873MiB / 18979MiB
======= System info =======
Date: 周二 9月 3 13:29:58 2019
Application: Ovito 3.0.0-dev466
OS: Linux
uname output: 5.3.0-1-MANJARO #1 SMP Mon Aug 26 11:38:15 UTC 2019 x86_64 unknown GNU/Linux
LSB output: ManjaroLinux "Manjaro Linux" 18.1.0-rc7
Architecture: 64 bit
Floating-point size: 64 bit
Qt version: 5.12.4
Command line: /opt/ovito/bin/ovito
======= OpenGL info =======
Version: 4.5
Profile: core
Alpha: 0
Vendor: X.Org
Renderer: Radeon 500 Series (POLARIS12, DRM 3.33.0, 5.3.0-1-MANJARO, LLVM 8.0.1)
Version string: 4.5 (Core Profile) Mesa 19.1.5
Swap behavior: other
Depth buffer size: 24
Stencil buffer size: 8
Shading language: 4.50
Geometry shaders supported: yes
Using deprecated functions: no
Using point sprites: yes
Using geometry shaders: yes
Context sharing enabled: yes

Quote from Alexander Stukowski on September 3, 2019, 9:42 amHi,
To further diagnose the problem, it would be helpful for me to know the following:
- Which renderer did you use (OpenGL or software-based)?
- Did the pipeline contain any modifiers? Do you have the impression that the memory problem occurs independent from the number and type of modifiers you have in the data pipeline?
-Alex
Hi,
To further diagnose the problem, it would be helpful for me to know the following:
- Which renderer did you use (OpenGL or software-based)?
- Did the pipeline contain any modifiers? Do you have the impression that the memory problem occurs independent from the number and type of modifiers you have in the data pipeline?
-Alex

Quote from Zheng Wang on September 3, 2019, 10:37 amQuote from Alexander Stukowski on September 3, 2019, 9:42 amHi,
To further diagnose the problem, it would be helpful for me to know the following:
- Which renderer did you use (OpenGL or software-based)?
- Did the pipeline contain any modifiers? Do you have the impression that the memory problem occurs independent from the number and type of modifiers you have in the data pipeline?
-Alex
Hi, Alex.
I tried all renderers. No one would finished rendering.
I tried with and without some modifiers, the problem exists in all situations.
The program may not release memory after finished rendering each frame?
I have noticed the increases of memory usage in each frame is nearly the same as the file size of a lammps dumped file of a single step.
Quote from Alexander Stukowski on September 3, 2019, 9:42 amHi,
To further diagnose the problem, it would be helpful for me to know the following:
- Which renderer did you use (OpenGL or software-based)?
- Did the pipeline contain any modifiers? Do you have the impression that the memory problem occurs independent from the number and type of modifiers you have in the data pipeline?
-Alex
Hi, Alex.
I tried all renderers. No one would finished rendering.
I tried with and without some modifiers, the problem exists in all situations.
The program may not release memory after finished rendering each frame?
I have noticed the increases of memory usage in each frame is nearly the same as the file size of a lammps dumped file of a single step.

Quote from Alexander Stukowski on September 3, 2019, 3:22 pmHi,
Thanks for providing the additional information. I was able to reproduce the problem and to fix a bug, which must have sneaked into the development version a while ago. Some data structures did not get released in a timely manner before the animation rendering process ends --eventually clogging memory when rendering large datasets or a large number of frames.
The issue should be resolved in Ovito build dev481. Let me know if you still notice any problems.
-Alex
Hi,
Thanks for providing the additional information. I was able to reproduce the problem and to fix a bug, which must have sneaked into the development version a while ago. Some data structures did not get released in a timely manner before the animation rendering process ends --eventually clogging memory when rendering large datasets or a large number of frames.
The issue should be resolved in Ovito build dev481. Let me know if you still notice any problems.
-Alex

Quote from Zheng Wang on September 3, 2019, 5:57 pmQuote from Alexander Stukowski on September 3, 2019, 3:22 pmHi,
Thanks for providing the additional information. I was able to reproduce the problem and to fix a bug, which must have sneaked into the development version a while ago. Some data structures did not get released in a timely manner before the animation rendering process ends --eventually clogging memory when rendering large datasets or a large number of frames.
The issue should be resolved in Ovito build dev481. Let me know if you still notice any problems.
-Alex
Hi,
It's OK now. Thank you very much. I'm very happy.
Quote from Alexander Stukowski on September 3, 2019, 3:22 pmHi,
Thanks for providing the additional information. I was able to reproduce the problem and to fix a bug, which must have sneaked into the development version a while ago. Some data structures did not get released in a timely manner before the animation rendering process ends --eventually clogging memory when rendering large datasets or a large number of frames.
The issue should be resolved in Ovito build dev481. Let me know if you still notice any problems.
-Alex
Hi,
It's OK now. Thank you very much. I'm very happy.
新的OVITO微信频道!
New for our users in China: OVITO on WeChat
Official OVITO WeChat channel operated by Foshan Diesi Technology Co., Ltd.
