The GPUOutOfMemoryError interface of the WebGPU API describes an out-of-memory (oom) error indicating that there was not enough free memory to complete the requested operation.

It represents one of the types of errors surfaced by GPUDevice.popErrorScope and the uncapturederror event.

Out-of-memory errors should be relatively rare in a well-behaved app but are less predictable than GPUValidationErrors. This is because they are dependent on the device your app is running on as well as other apps that are using GPU resources at the time.

GPUOutOfMemoryError() Experimental

Creates a new GPUOutOfMemoryError object instance.

The message property is inherited from its parent, GPUError:

message Experimental Read only

A string providing a human-readable message that explains why the error occurred.

The following example uses an error scope to capture an out-of-memory error, logging it to the console.

device.pushErrorScope("out-of-memory");

let buffer = device.createBuffer({
  size: 100_000_000_000, // 100GB; far too big
  usage: GPUBufferUsage.COPY_SRC | GPUBufferUsage.MAP_WRITE,
});

device.popErrorScope().then((error) => {
  if (error) {
    // error is a GPUOutOfMemoryError object instance
    buffer = null;
    console.error(`Out of memory, buffer too large. Error: ${error.message}`);
  }
});

• The WebGPU API
• WebGPU Error Handling best practices