Click to auto-format PR. [Please do not click if the PR is not owned/reviewed by you.]
#2338, [Refactor] Add a class Callable to unify Kernel and Function, by xumingkuan
#2337, [opt] Slightly simplify algebraic simplification, by xumingkuan
#2334, [Example] cornell_box: Remove unused sphere, by jammm
#2327, [Lang] [ir] Consecutive indices for all snodes, by KLozes
#2274, [misc] Support Python 3.9, by hooyuser
#2273, [cuda] Add memfence to dynamic allocate and append, by k-ye
#2268, [metal] Add 3-stage GC Metal kernels, by k-ye
#2261, [Doc] Update install.rst, Change installer for archlinux from yaourt to yay, by liberix
#2255, [Mac] Set up Apple M1 GitHub actions, by yuanming-hu
#2248, [Bug] [llvm] [ir] Complex multilevel structs. Consecutive indices for all snodes Draft, by KLozes
#2172, [doc] Troubleshooting CUDA out of memory (taichi_blend#8), by archibate
#2076, [cc] [refactor] Re-open automatically detect available compiler (clang or gcc), by archibate
#2052, [export] [cc] Record kernel group as C string array, by archibate
#2050, [Bug] [lang] Fix is_global on 0-D vectors, by archibate
#2025, [opencl] Setup basis of OpenCL backend, by archibate
#1861, [Misc] (WIP) Automatic release pipeline, by rexwangcc
#1830, [refactor] [lang] Restore ODOP support, by archibate
#1826, [Lang] Support x.fill(n) and x.copy_from(y) in top-level of Taichi-scope, by archibate
#1823, [doc] Document the Python coding style, by archibate
#1814, [workflow] Add release test workflow, by archibate
#1719, [Perf] Make Python-scope accessors 8x faster, by archibate
#1586, [Doc] Add section explaining backgrounds of Taichi to overview.rst, by archibate