framework overhead in real deployments: anyone testing this?

[embershed99]

Did a small comparison this afternoon. On mini PC with TensorRT-LLM, I saw basically the same behavior. The surprise was that memory pressure showed up before throughput collapsed.

[llmpeak]

Interesting, I hadn't tried that. That matches what I saw on 7900 XTX.

[vectorpoint]

Appreciate the detail. The pattern around small local stacks has been pretty similar for me.

[audioriver]

Maybe, but I had the opposite result. The reported numbers can look compelling, but in ordinary usage I didn't see the same benefit. Could be a workload difference, but I wouldn't generalize too fast from one setup. Turns out 'just one more test' was a lie.

[marbletrack]

My first guess would be this: On 5090, the easiest win is often to back off the aggressive settings and see whether the problem survives.

Can you fill in a bit more context? Are you testing this on M3 Max or something else?

[promptfield]

Spent some time on this today and wanted to compare notes.

Two paths both seem viable, but they break down differently once reranking becomes important.

What would you pick if you had to live with it for a few months?

[localpatch]

My first guess would be this: If you're on MLX, I'd confirm the version first and then reduce variables until the behavior is obvious. Once you know the plain path works, then add the faster path back in piece by piece.

[plainline]

following this because I'm in the same boat.

[northcraft90]

I don't think that's the whole story. I wouldn't optimize around that first. On my side the bigger swing came from the runtime and not the model. Worth testing the simpler explanation before adopting the more interesting one.

[pixelcrate82]

Maybe, but I had the opposite result. I get why that seems right, but on Threadripper box the bottleneck felt elsewhere. Could be a workload difference, but I wouldn't generalize too fast from one setup.