AnggaSP

There absolutely does, Android did with low power audio. They even goes a step further by offloading bluetooth processing into DSP.

I’m not in this space anymore but as of Android 5-6 era aac and bt is offloaded to hexagon dsp on qualcomm device.

Yap it’s called GDC and I’ve set it up for customer that requires it because of regulatory compliance.

It’s very expensive but you got the box ready in your data center and is managed by both Google and partner.

Yet they’re lacking in the model spaces.

Things changed since Apple Intelligence but I was hoping there’ll be more things like live captions and what-not than chatbot use cases. I feel pixel is also moving towards that and abandoning the old way unfortunately.

TPUs are not that portable and easy for both inferencing and training. It has since improved a lot with their effort on the torch backend (XLA/TorchTPU) and JAX though.

But as far as i know it currently supports just that + tensorflow (which nobody uses it anymore, least here). And last we tried, so much of our kernels needs rework that it’s not worth the effort.

This may change since ironwood but we haven’t tried that generation.

Former custom ROM developer here who has reviewed and integrated valdikSS modification.

It’s not a bug that the patchset is doing, but enables dual-channel SBC to be negotiated in the source and sink connection. This enables the higher bit-rate without exceeding the maximum bitpool both Android and BT receiver impose.

There’s still negotiation occurs between the source and sink and if either one of them doesn’t support dual-channel SBc, it’ll fallback to whatever its supported. All the device that i had maintained supports it, while some cheap speaker that i test at that time doesn’t and able to negotiate joint stereo session.