Open quintus0505 opened 3 years ago
Hi! Thank you for reporting this.
First of all, please make sure that you executed the setup_env
script in Vulkan SDK, i.e. following the instruction:
$ cd vulkansdk/1.2.162.0
$ source ./setup-env.sh
After you've done this, check that VULKAN_SDK
env variable is set by running this in your terminal:
$ echo $VULKAN_SDK
To make sure the SDK headers contain the correct macros (the build procedure seems to be complaining about VK_ERROR_UNKNOWN
), can you please search for it in the SDK folder, i.e. like this:
$ grep -r 'VK_ERROR_UNKNOWN' ./x86_64/include
Expected output is something like:
./x86_64/include/vulkan/vk_enum_string_helper.h: case VK_ERROR_UNKNOWN:
./x86_64/include/vulkan/vk_enum_string_helper.h: return "VK_ERROR_UNKNOWN";
./x86_64/include/vulkan/vulkan.hpp: eErrorUnknown = VK_ERROR_UNKNOWN,
./x86_64/include/vulkan/vulkan_core.h: VK_ERROR_UNKNOWN = -13,
Thank you very much! After following the new instruction, I got the same out put. But there is another thing I need to report. After finding I didn't miss any steps when setting my vk, I realized that it may be my conda-env's problem. I use the megaverse conda-env you provided this time and successfully passed this step. But when I changed to my old conda-env which I used conda env update -f environment.yml
to update, I get the same output by following your instruction but still magically get stuck at 97%.
Still, the problem is solved, thanks!
Oh, this is interesting! Thank you for letting us know! Looks like some package conflict then. We probably also shouldn't have precise versions of all packages specified in the conda env, it might mess with use cases like this (combination of multiple envs)
I have followed the instruction and get stuck at setup.py. It seems that my cuda and vk have some conflicts. After several cuda uninstall and install, I finally give up and come here to look for help.
Here is my code, I run this code with system ubuntu 18.04 and currently my cuda version is 10.2