Expensive-looking oddities in CUDA profile

Running [this benchmark](https://gist.github.com/inducer/f836e9913a5c4121857f50814e5662e3) based on `wave-op-mpi.py` on 1c44c4b with the command
```
PYTHONHASHSEED=17 PYOPENCL_TEST=port:nvid setarch -R numactl -C 2 -m 0 nvprof -f -o yoink.nvvp python -O wave-op-mpi.py --dim=3 --order=4
```
on `dunkel` gives me the following profile in Nvidia's visual profiler:

There are at least two things wrong here (both circled):
- There are a bunch of big gaps where nothing seems to be happening. Why?
- Every now and then, a `cuLaunchKernel` seems to take a very long time. Why?

Curiously, there seem to be periods that don't suffer from this:

If we could fix these two types of stalls, I suspect our performance story would look quite a bit different.

cc @matthiasdiener @lukeolson

Other versions in use, for reproducibility:
- https://github.com/inducer/meshmode/commit/c7b78410fcf1d1f3e360d4a116ef190e450263dd
- https://github.com/inducer/arraycontext/commit/77b03e5e9eed69a9cfd51bebae68e0f746c5bacf
- https://github.com/inducer/loopy/commit/e232bb1361479f2e6fd3bd7efd7f4ecce270947f
- https://github.com/inducer/pyopencl/commit/a56281afc563e6be448cd8b2b2b69bc3afc59d78