-
Notifications
You must be signed in to change notification settings - Fork 2.3k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
meet some problems of Voxel Challenge on taichi 1.0.1(Received signal 22) #4900
Comments
Same here, |
same here, and I check and meets the same problem for cpu,vulkan,gpu arch. |
I got this:
Vulkan api Version: 1.3.194 |
It looks like this wasn't actually fixed, I compiled the latest commit and still got the same error:
Vulkan API: 1.3.203 It could possibly have something to do with me having an AMD GPU, but given that its the same error the folks above me got on NVIDIA GPUs I highly doubt it |
@Smethan that is quite baffling... Can you try to add |
I have no idea what I have done in the week since I posted that to anger my windows machine, but I cannot for the life of me get this to build, even just pulling from master wont compile, so I can't really test it at the moment. In the mean time, if you can compile it with that line in there and send it my way I'd be happy to run it. |
Scratch that last comment, after a few hours of finagling I was able to get it to compile and this is the output I got:
with the only real different line there being
which doesn't really tell us much as far as I can tell. I'm personally a bit lost on what to do here, this low level stuff isn't exactly my area of expertise 😅 |
This actually tells us quite a lot... Basically we can not acquire a function pointer that is supposed to be supported by Vulkan 1.3 Core, which means the driver isn't reporting its version capability correctly... Can you try to get a device profile using this tool here? https://vulkan.gpuinfo.org/download.php |
Sure thing! Here you go: |
Also let's move this to #4951 |
Interestingly I got the same error message GPU is Radeon RX 5500. Edit: a quick hack for a similar setup is to change VK_API_VERSION_1_2 to VK_API_VERSION_1_3 in VulkanEnvSettings. |
This change fixes issue taichi-dev#4900. On certain AMD GPUs (such as RX 5500), the Vullkan version needs to be 1.3 in order for Volk to correctly populate the "vkGetDeviceBufferMemoryRequirements" function pointer. If the version is 1.2, the pointer will be NULL and cause an assertion failure. Tested by running the voxel challenge program and see the program working as expected on a Windows PC.
Describe the bug
Running example1.py by using the pycharm with taichi 1.0.1,python 3.10.2,arch=cuda,windows and got many errors .
(I have tried to switch python version,it still didn't work)
After I switch the taichi to 1.0.0 it did't report errors and program runs correctly.
So,what is going on here?
I have seen https://docs.taichi-lang.org/lang/articles/install
But it didn't mentioned "Received signal 22 (SIGABRT)"
Log/Screenshots
Please post the full log of the program (instead of just a few lines around the error message, unless the log is > 1000 lines). This will help us diagnose what's happening. For example:
btw,on my report log here is full of unreadable marks
A s s e r t i o n f a i l e d : m _ V u l k a n F u n c t i o n s . v k G e t D e v i c e B u f f e r M e m o r y R e q u i r e m e n t s ! = n u l l p t r , f i l e D : / a / t a i c h i / t a i c h i / e x t e r n a l / V u l k a n M e m o r y A l l o c a t o r / i n c l u d e \ v k _ m e m _ a l l o c . h , l i n e 1 4 4 4 0
The text was updated successfully, but these errors were encountered: