tf.config.set_logical_device_configuration not working when attempting to create multiple virtual GPUs

NOTE: This is a cross-post from an issue filed here: https://github.com/tensorflow/tensorflow/issues/56837

I'm developing on a system with a single GPU (Apple M1 Pro) and trying to simulate multiple GPUs with virtual devices.

Using the examples found here: https://www.tensorflow.org/guide/gpu#using_multiple_gpus I get the following output:

systemMemory: 32.00 GB
maxCacheSize: 10.67 GB

1 Physical GPU, 1 Logical GPUs

where I was expecting the Logical GPUs to be greater than the Physical GPU count.

I have tested tf.config.set_logical_device_configuration for increasing Logical device count for CPUs, and this indeed does work for CPU virtualisation.

Standalone code to reproduce the issue

Taken from the code snippet found at: https://www.tensorflow.org/guide/gpu#using_multiple_gpus

gpus = tf.config.list_physical_devices("GPU")
if gpus:
    # Create 2 virtual GPUs with 1GB memory each
    try:
        tf.config.set_logical_device_configuration(
            gpus[0],
            [
                tf.config.LogicalDeviceConfiguration(memory_limit=1024),
                tf.config.LogicalDeviceConfiguration(memory_limit=1024),
            ],
        )
        logical_gpus = tf.config.list_logical_devices("GPU")
        print(len(gpus), "Physical GPU,", len(logical_gpus), "Logical GPUs")
    except RuntimeError as e:
        # Virtual devices must be set before GPUs have been initialized
        print(e)

Relevant log output

1 Physical GPU, 1 Logical GPUs

Tensorflow Version

tensorflow-macos v2.8.0 tensorflow-metal v0.5.0

OS Platform and Distribution

macOS-12.2.1

Python version

3.8

Followup

Unfortunately I do not have access to a Windows machine for testing, but I wonder if it's a problem with tensorflow-macos / tensorflow-metal specifically. Is this repo still the best place to log such an issue?

I tried again locally on macOS and bumped versions of tensorflow-macos to be:

tensorflow-macos        2.9.2
tensorflow-metal        0.5.0

and used Python 3.9, but still same problem as reported before, i.e.

1 Physical GPU, 1 Logical GPUs