yekuang

[AOT] Do not re-JIT if the kernel has been added to the AOT module already

1

**Describe the bug** If a kernel has been added to an AOT module multiple times, it should only be JIT once. While I found this problem for taichi kernels, we...

**Concisely describe the proposed feature** The current AOT solution is good to use for dense Taichi fields. However, to support all kinds of Taichi fields, we probably need to systematically...

* [x] Make CPU backend work * [x] Support arg-less kernels * [x] Support scalar + Ndarray args * [x] Split LlvmProgramImpl into compile-time & runtime parts to form Runtime...

- [ ] GPU side implementation is messy, so is the code in unittests. - [x] A lot of the code should be shared between CPU/GPU. Kernel functions, random floating...

- [x] World size could be different in each dimension (Done) - [x] Support for the dynamic world bounding box, where we could see the effect of the fluid being...

For those platform that do not have CUDA, the code should still build and run the CPU version.

Support `SymlinkNode` in `ComputeMerkleTree()`?

9

IIUC, currently the file tree produced by `ComputeMerkeTree()` https://github.com/bazelbuild/remote-apis-sdks/blob/c00c035a74d44f1dc7d7ce8fe5dbca3b8956ed1a/go/pkg/tree/tree.go#L110 could only contain `FileNode` or `DirectoryNode`, but not `SymlinkNode`. This could be very useful for Swarming's isolate support, since the `.isolate`...

`to_pyarrow_table()` on a table in S3 kept getting "Generic S3 error: error decoding response body"

30

# Environment **Delta-rs version**: `deltalake==0.18.1` **Binding**: Python **Environment**: - **Cloud provider**: AWS S3 - **OS**: Linux - **Other**: - Python 3.10.4 - `pyarrow==16.1.0` - `pyarrow-hotfix==0.6` *** # Bug **What happened**:...