Yotam Gingold
Yotam Gingold
Have you tried the `2to3` tool that comes with Python? > On Jul 3, 2022, at 9:16 PM, Omar Irfan Khan ***@***.***> wrote: > > > I converted the first...
If I remember correctly, numpy used to let missing dimensions broadcast, so a (2,3) array and a (3,) array were compatible. Later, they started requiring that you explicitly add a...
I haven't looked into this, but it may be easy to add. Are you saying that there are CUDA-enabled SuiteSparse builds? I would assume that they expose a C function...
The library is opened dynamically. It may be possible to check if the function exists and then import pycuda to allocate GPU memory. That would be a relatively small task.
My students have this same problem on Windows. I don't have this issue on macOS 15.6.1). These macros were removed in the latest commit to glfw3webgpu: . It works again...
How could that be part of the spec? In JavaScript, I didn't think it was possible to manually release a RenderPass.
Where are things like this RenderPass lifetime rule specified? My understanding was that the JS API and the C API should be implementable on top of each other. I don't...
Yes, this is exactly the same. The extra scope for the render pass shown there serves as a workaround for a ref-counted language (including C++ with RAII). I guess a...
My understanding is that this is a deviation from the spec.
This is a wgpu bug ("deviation from the spec") for render and compute passes. Is there a bug tracker elsewhere that we should use to report this?