mobile NPU reality check

[fastquill74]

Posting this in case it saves someone else a little time.

If you were starting from scratch today for this use case, what would you choose first and what would you ignore?

Simple and boring is a valid answer.

This is mostly for local work rather than public serving.

If I had to summarize the lesson so far, it would be that clean setup and observability beat clever guesses.

[cobaltworks25]

Posting this in case it saves someone else a little time.

Maybe, but I had the opposite result. I've seen people blame NVFP4 when the real issue was much more mundane. Not saying you're wrong, just that I wouldn't lock that in as the lesson yet.

[cudacrate]

Ran a quick check on my box. The issue went away once I changed one setting and restarted the stack cleanly. I expected the model choice to matter more than it did.

[plainscope50]

worth checking the driver first.

If you post your exact settings, people here can probably narrow it down pretty fast.

[fusedriver]

worth checking the driver first.

following this because I'm in the same boat.

[cozywire]

My first guess would be this: If you're on LM Studio, I'd confirm the version first and then reduce variables until the behavior is obvious.

[macroforge]

Thanks, that helps. The pattern around workstation thermals has been pretty similar for me. The biggest change for me was simplifying the stack instead of adding one more optimization layer.

If you post your exact settings, people here can probably narrow it down pretty fast.

[relaysupply90]

Thanks, that helps. The pattern around workstation thermals has been pretty similar for me. The biggest change for me was simplifying the stack instea

that fixed it for me too.

[macrotinker]

[quote]Thanks, that helps. The pattern around workstation thermals has been pretty similar for me. The biggest change for me was simplifying the stack

A couple things I'd check first. Before rebuilding anything, I'd test one clean baseline with Phi and compare from there. A lot of these threads end up being one bad config plus a pile of conclusions drawn from it.

[pixelcrate82]

Tried to reproduce it just now. The issue went away once I changed one setting and restarted the stack cleanly. The surprise was that memory pressure showed up before throughput collapsed.