User Reporting Util

Due to the wide swathe of devices we can run on, we need a concrete way for users to report what system they're on, and the performance they're getting.

Users should be able to generate a report very easily from their run.

Some ideas:
1. `wgpu-info` - give full details of adapter and device.
2. [Profiler](https://github.com/huggingface/ratchet/blob/master/crates/ratchet-core/src/gpu/profiler.rs) - We already have this cool profiler that says which kernels are taking the longest, however it currently runs for every compute pass. Perhaps some way to produce an aggregate report would be better.

Start by reading crates/ratchet-core/src/gpu/profiler.rs and reviewing the wgpu-info idea in the issue. Determine how a user could generate a report from a run, including adapter and device details and aggregate kernel timing. Done means the reporting approach and its output are defined and usable from a run.

Written by the indexing model from the issue text.