search

erizmr / SPH_Taichi

A high-performance implementation of SPH in Taichi.
195 stars 39 forks source link

Taichi version #21

Closed DrChiZhang closed 1 year ago

DrChiZhang commented 1 year ago

Hi, thanks for sharing your interesting work. I have tried to run your code, while failed due to the restricted taichi version installed by PIP. Should I build the taichi from source other than with pip installation? Thanks

erizmr commented 1 year ago

Hi, there is no need to build from Taichi for this repo. Taichi >= 1.2.0 installed by PIP should work. BTW, which version of Taichi are you using? And it will be helpful if you can share some info about the failure? Thanks.

DrChiZhang commented 1 year ago

Hi I have installed taichi through pip install --upgrade taichi with taichi 1.4.0 installed. And the print error is the following by running the dragon_bath simulation:

warnings.warn( Traceback (most recent call last): File "run_simulation.py", line 36, in window = ti.ui.Window('SPH', (1024, 1024), show_window = True, vsync=False) File "/home/chi/.local/lib/python3.8/site-packages/taichi/ui/window.py", line 36, in init check_ggui_availability() File "/home/chi/.local/lib/python3.8/site-packages/taichi/ui/utils.py", line 85, in check_ggui_availability raise GGUINotAvailableException( taichi.ui.utils.GGUINotAvailableException: GGUI is not available since you have installed a restricted version of taichi. Please see yellow warning messages printed during startup for details.

thanks

erizmr commented 1 year ago

Hi, thanks for providing this. The GGUI supports x64 / CUDA / Vulkan backends. Would you mind provide more info about your OS/ GPU or CPU?

btw, pip installation from a manylinux wheel does not support GGUI as mentioned in https://github.com/taichi-dev/taichi/issues/5090

One solution is upgrade the pip and reinstall Taichi. If it is still not woking, you can try to follow https://docs.taichi-lang.org/docs/dev_install and build Taichi from source. The vulkan SDK shown in https://docs.taichi-lang.org/docs/dev_install#install-optional-dependencies is required for GGUI if building from source.

DrChiZhang commented 1 year ago

Hi My OS Distributor ID: Ubuntu Description: Ubuntu 20.04.5 LTS Release: 20.04 Codename: focal

GPU and CUDA +-----------------------------------------------------------------------------+ | NVIDIA-SMI 520.61.05 Driver Version: 520.61.05 CUDA Version: 11.8 | |-------------------------------+----------------------+----------------------+ | GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC | | Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. | | | | MIG M. | |===============================+======================+======================| | 0 NVIDIA RTX A5000 On | 00000000:01:00.0 Off | Off | | 30% 30C P8 16W / 230W | 216MiB / 24564MiB | 0% Default | | | | N/A | +-------------------------------+----------------------+----------------------+

Vulkan

========== VULKANINFO

Vulkan Instance Version: 1.2.131

Thanks for your help

erizmr commented 1 year ago

Thanks for providing this. Have you tried to upgrade PIP and re-install?

DrChiZhang commented 1 year ago

Thanks for providing this. Have you tried to upgrade PIP and re-install?

Tried while failed with the same error. Looks like this error is related with the Linux OS, I will try with another computer. Best

erizmr commented 1 year ago

Hi, I asked Taichi developers about this issue and got reply that pip >= 20.3, python 3.8.10+, 3.9.5+, 3.10.0+ is required for GGUI.

DrChiZhang commented 1 year ago

Hi, I asked Taichi developers about this issue and got reply that pip >= 20.3, python 3.8.10+, 3.9.5+, 3.10.0+ is required for GGUI.

Thanks, I have upgrade the pip to pip 22.3.1, now taichi works with the default installation.

DrChiZhang commented 1 year ago

I close this issue. Thanks a lot.