GPU event notifications (thermal throttling, ECC errors, XID)

[pmady]

Add a tool or MCP notification mechanism that reports GPU hardware events.

Requirements:

• New tool: get_gpu_events — returns recent GPU events
• Track thermal throttling events via nvml.DeviceGetViolationStatus()
• Track ECC errors via nvml.DeviceGetTotalEccErrors()
• Track XID errors via nvml.DeviceGetXidErrors() (if available)
• Return events with timestamps, severity, and affected device

Why this matters:

• Agents can proactively alert on hardware degradation
• Thermal throttling explains sudden performance drops
• ECC errors indicate memory reliability issues

Difficulty: Hard — requires understanding NVML event APIs and potentially background polling