prutschman-iv

cupy.sign(NaN) == 0, unlike numpy

9

### Description numpy.sign(numpy.nan) returns NaN, but cupy.sign(cupy.nan) returns 0. I expected either the behavior to be the same, or the difference to be documented. (Edit: I see now that the...

Match numpy sign semantics

3

Here is my first pass at modifying the `sign` function to match `numpy`'s NaN semantics. It includes a test case using `check_unary_inf_nan`. This is to fix #8327 I've confirmed that...

Build error on main under Windows w/ CUDA 12.2

3

### Description I can successfully build from the v12 branch. When I try to build from main, the build fails. I've attached a complete output from the build as a...

Catastrohpic accuracy loss in large float32 array for nanmean and nanstd

2

**Describe the bug** Starting somewhere between 10 million and 50 million elements, the `bn.nanmean` and `bn.nanstd` functions appear to experience a catastrophic loss of accuracy with float32 data. **To Reproduce**...

High CPU usage when idle, possibly DNS related

6

**Describe the issue**: I'm seeing high CPU usage (4-6% on an 18 core/36 thread CPU) by "Service Host: DNS Client" associated with dask workers, even when they're not running tasks....