For the complete documentation index, see llms.txt. Markdown versions of all pages are available by appending .md to any URL (e.g. /get-started.md).

Python class

MemoryStats

MemoryStats

class max.profiler.gpu.MemoryStats(total_bytes, free_bytes, used_bytes, reserved_bytes)

source

Bases: Struct

Detailed GPU memory usage statistics including total, free, used, and reserved memory.

This class provides comprehensive memory information for a GPU, allowing developers to monitor memory consumption and identify potential memory bottlenecks during model inference or training.

Parameters:

• total_bytes (int)
• free_bytes (int)
• used_bytes (int)
• reserved_bytes (int | None)

free_bytes

free_bytes: int

source

Currently available GPU memory in bytes.

reserved_bytes

reserved_bytes: int | None

source

Memory reserved by the driver, if available from the GPU vendor.

total_bytes

total_bytes: int

source

Total GPU memory capacity in bytes.

used_bytes

used_bytes: int

source

Currently allocated GPU memory in bytes.