Identify platform targets for future releases of TensorFlow Quantum
Post-0.7.5, we need to determine what platforms we can support. There is a matrix formed by operating system, OS version, CPU architecture, GPU/no GPU, and possibly other dimensions. This issue is meant as a place together notes on this topic.
Current matrix:
| OS | OS version(s) | CPU arch. | GPU/GPU interf. | TFQ support |
|---|---|---|---|---|
| Linux | Ubuntu 22.04, 24.04 | x86_64 | none | 0.7.5+ |
| MacOS | x86_64 | |||
| MacOS | Apple Silicon | |||
| Windows | x86_64 |
It looks like TensorFlow 2.16 was the last release that supported macOS on x86 (Intel) architectures. This means that after 0.7.5, we will not be able to support Macs unless we get TFQ running on Apple Silicon.
TensorFlow pages listing their tested build configurations:
Linux/Mac: https://www.tensorflow.org/install/source#linux Windows: https://www.tensorflow.org/install/source_windows