Ailing Zhang
Ailing Zhang
1. Change indicator label from `error-prone` to `full-ci` 2. Always run full ci test when dst branch is a rc branch
Currently Taichi only supports contiguous pytorch tensors as taichi kernel argument, https://github.com/taichi-dev/taichi/pull/4258 but it'll be convenient / efficient for users if they can directly pass in strided pytorch tensors without...
Copied from #4615 :P : This would significantly reduce the efforts to porting a demo written in Taichi to non-Python environments. In essence, the AOT module will save not only...
Currently we support `/rerun` command, `/rerun-failed` will be also very useful. (alternative is that we give everyone write permission to do this...
#1225 XLA has a optimized `einsum` implementation that we can use. Requires a change in upstream.
- [ ] @ghuau-innopeak noticed that increasing https://github.com/taichi-dev/taichi/blob/master/taichi/inc/constants.h#L11 from 16 to 32 broke implicit_fem demo on some Mali GPUs (I can repro this on a Find X5 Pro MTK version)....
I forgot to upgrade version number in `rc-v1.1.1` branch and had to yank this release due to the following two issues: - some of the built wheels have `1.1.0` in...
Since we moved opengl backend from glsl source codegen to spirv based, one important debugging feature we lost is printing support. #6075 added the support for vulkan and it shouldn't...
related: #5634 We'll use this issue to track progress in opengl aot support. To give some context, we removed opengl codegen and moved it on top of spirv codegen in...
Currently our build system supports multiple configs like `TI_WITH_LLVM, TI_WITH_VULKAN` etc, but we only test a few combos of them in CI pipeline. It'd be nice if we can add...