Kaushik Kulkarni

- git-blame shows was deprecated in 2016: https://github.com/inducer/loopy/commit/94361fd75f742fd5138b922f6f884e42a68e1f4e. - I will wait doing this until #222 lands in as none of our other routines use this API.

What to do about separate "loop entry roots" sharing a domain

9

Consider: ```python knl = lp.make_kernel( "{[i,j, l]: 0

[Enhancement]: Parallelizing a very long kernel is slow

6

```python3 import loopy as lp import vmprof def make_kernel(): n_insn = 2000 # 'ndomains' must be greater than 'n' as it also includes the iname domains # of sub-arrays refs....

[PyOpenCL execution]: Scalar argument always returned as pyopencl arrays

3

Minimal reproducer: ```python knl = lp.make_kernel( "{:}", """ y = 2*x """) evt, (out,) = knl(cq, x=np.asarray(3.0, dtype=float)) print(repr(out)) # prints: cl.Array(6.) ``` However if 'y' was anything other than...

Suggested here: https://github.com/inducer/loopy/pull/258#discussion_r602614785

`builtins.sum`: Spurious error for operands having `__add__` defined using `partialmethod`

2

Here's a minimal reproducer: ```python from __future__ import annotations from functools import partialmethod class A: def my_add(self, other: A, reverse: bool) -> A: return self __add__ = partialmethod(my_add, reverse=False) a...

Adding support for calling MLIR functions

2

- Added module `mlir.run` that contains helpers for calling MLIR functions. - Demonstrated some examples under `tests/test_run.py`.

It was suggested in https://github.com/inducer/pytato/pull/217 that we should preserve the structure of the expressions. This PR implements the points discussed over there. Draft because: - [ ] Fails CI

StringifyMapper: Incorrect result for side-by-side product nodes

7

```python >>> import pymbolic.primitives as p >>> a = p.Variable("a") >>> b = p.Variable("b") >>> c = p.Variable("c") >>> print(a * (b * c)) a*b*c # Should have been 'a*(b*c)'...