In implementing https://github.com/jax-ml/jax/pull/25787, I realized that while we lower `tridiagonal_solve` to cuSPARSE on GPU, we were using an explicit implementation of the Thomas algorithm on CPU. We should instead lower to LAPACK's `gtsv` on CPU because it should be more numerically stable and faster.
PiperOrigin-RevId: 714069225
When it is possible to annotate an operation using both a `strided` and a
`splat` layout, we pick the `strided` layout. This is the correct choice when
propagating layouts down from parameters to the root; e.g.
```
? = add(strided, splat)
```
becomes
```
strided = add(strided, strided)
```
and requires a re-layout for the right-hand side argument.
The logic needs to be reversed to handle propagation in the opposite direction.
For example, code like
```
c : ?
use(c) : strided
use(c) : splat
```
should resolve to
```
c : splat
use(c) : strided
use(c) : splat
```
and incur a relayout in the `strided` use of `c`. This direction of propagation
is left as a `TODO` for now, to limit the amount of changes in a single commit.
PiperOrigin-RevId: 714056648
Fixes: #25140
Previously, the following code:
```
def f(i, x):
return lax.switch(i, [lambda x: dict(a=x),
lambda x: dict(a=(x, x))], x)
f(0, 42)
```
resulted in the error message:
```
TypeError: branch 0 and 1 outputs must have same type structure, got PyTreeDef({'a': *}) and PyTreeDef({'a': (*, *)}).
```
With this change the error message is more specific where the
difference is in the pytree structure:
```
TypeError: branch 0 output must have same type structure as branch 1 output, but there are differences:
* at output['a'], branch 0 output has pytree leaf and branch 1 output has <class 'tuple'>, so their Python types differ
```
Everything passes other than an io callback test due to the lowered `sdy.manual_computation` returning a token. Will be fixed in a follow-up.
PiperOrigin-RevId: 713780181
A pivoted QR factorization is possible in `scipy.linalg.qr`, thanks
to the `geqp3` routine of LAPACK. To provide the same functionality
in JAX, we implement a new primitive `geqp3_p` which calls the LAPACK
routine via the FFI on CPU devices.
Both `jax.scipy.linalg.qr` and `jax.lax.linalg.qr` now support the
use of column-pivoting on CPU devices.
To provide a GPU implementation of `geqp3` may require using MAGMA,
due to the lack of a `geqp3` implementation in `cuSolver` - see
ccb331707e80b16d89de6e5c9f2f89b87c1682ed (`jax.lax.linalg.eig`) for
an example of using MAGMA in GPU lowerings. Such a GPU implementation
can be considered in the future.
The Python warnings.catch_warnings() functionality is not thread-safe (https://py-free-threading.github.io/porting/#the-warnings-module-is-not-thread-safe), so we cannot use it during tests that use free-threading. This change introduces a private warnings test helper (test_warning_util.py), which hooks the CPython warning infrastructure and uses it to implement thread-safe warnings infrastructure.
This requires a handful of small modifications to tests to remove direct uses of the warnings module. We also sadly have to delete one TPU test that checks for a warning raised on another thread; there's no easy way for us to catch that in a thread-safe way, but that test seems like overkill anyway.
Boolean fields in the descriptor struct led to padding, which let random
bytes in the string representation of the struct and variance in HLO
from run to run.
