EXAMPLES/DEVICE/EP/CSRC: Replaced torch::Event by cudaEvent_t wrapper. (#1822)

## What?
Replaced `torch::Event` by `nixl_ep::cuda::Event`.
Cleaned up includes in `nixl_ep.[c|h]pp`.
Minor refactoring of `nixl_ep::EventHandle`.

## Why?
This is a part of https://github.com/ai-dynamo/nixl/pull/1793.
Original PR is devided into parts to facilitate review.


<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->
## Summary by CodeRabbit

* **New Features**
* Added RAII-based CUDA event support with stream recording and safe
ownership transfer.
  * Added clearer CUDA warning messages for runtime failures.
* **Bug Fixes**
  * Improved event cleanup safety and centralized CUDA error reporting.
* Updated cross-stream synchronization to use direct CUDA event-based
waits.
* **Refactor**
  * Reworked event synchronization to manage CUDA events internally.
  * Updated CUDA integration to support the revised event flow.
<!-- end of auto-generated comment: release notes by coderabbit.ai -->

---------

Signed-off-by: Raul Akhmetshin <rakhmetshin@nvidia.com>