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).

Mojo module

semaphore

This module provides a device-wide semaphore implementation for NVIDIA GPUs.

The Semaphore struct enables inter-CTA (Cooperative Thread Array) synchronization by providing atomic operations and memory barriers. It uses NVIDIA-specific intrinsics to implement efficient thread synchronization.

Examples:

```mojo
from max.gpu.sync import Semaphore

def semaphore(lock: Pointer[mut=True, Int32, ...]):
    var thread_id = 0
    var sem = Semaphore(lock, thread_id)

    # Wait for a specific state
    sem.wait(0)

    # Release the semaphore
    sem.release(1)
```

Structs

NamedBarrierSemaphore: A device-wide semaphore implementation for NVIDIA GPUs with named barriers.
Semaphore: A device-wide semaphore implementation for GPUs.

Structs​

Structs