Rico Rodriguez

@oliver-batchelor interesting workaround, but would there be any performance impact from marshaling bits and deserializing structs like this inside the kernel, or would that get optimized away by the compiler?...

The other thing I considered was [argpack types](https://docs.taichi-lang.org/docs/master/argument_pack), but given the inability to pass a field/vector of argpacks, you end up in a situation that just recreates passing disjoint fields...

> Fields are global and just can't be "passed" into the kernel... The only way to do it is to use `@ti.template`. On the other hand I'm not sure whether...

Ah, nevermind - it appears that I have a race condition going on, as indicated by running with the CPU backend. I'm completely new to GPU programming - what is...

Here's what I'm doing now - is this a good approach? It seems to work, although obviously there's some loss of precision with float32. ```py def ti_volume_convolution(kernel_slab: NDArray, stimulus: NDArray):...

@jim19930609 So the exact warning was ``` UserWarning: The value of the smallest subnormal for type is zero. ``` It sounds like it might be the same issue that [this...

Sure, but this is a github submodule dependency that I'm not trying to fork in order to get rid of a syntax warning that shouldn't be there in the first...

yeah - you can see it's coloring the code correctly up until the end of that Namespace definition. So you can't get any *.libsonnet working? What version of the plugin...

Anyone monitoring this repo? @lihaoyi-databricks ?

@lihaoyi-databricks it's been years since I wrote anything in Java and I've never written an intellij/pycharm plugin, so I wouldn't know where to start. Whereabouts would the trivial fix take...