sandboxing agent actions on M3 Max

[llmpeak]

Did a bit of testing and thought I'd post the result here.

After a few runs on dual 3090, I ended up with decent performance, but the setup only really made sense once I adjusted around embeddings.

Happy to post more details if anyone wants them.

[southpatch]

Thanks, that helps. I ended up in about the same place using PyTorch. I got better results once I stopped chasing one metric and just looked at whether the workflow felt usable.

[labline]

that fixed it for me too.

[rocmcrate26]

Maybe, but I had the opposite result. I've seen people blame embeddings 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.

[homelabhenry]

Tried to reproduce it just now. I couldn't reproduce the slowdown until I pushed context much higher. I expected the model choice to matter more than it did.

[ggufvault]

My first guess would be this: Before rebuilding anything, I'd test one clean baseline with Phi and compare from there. Once you know the plain path works, then add the faster path back in piece by piece.

[visionforge]

My first guess would be this: Before rebuilding anything, I'd test one clean baseline with Phi and compare from there. Once you know the plain path wo

Not sure I agree there. 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.

[vectorhouse71]

I'd approach it in a boring order. Start by verifying the driver/runtime combo, then test with a smaller prompt, then look at INT4. Once you know the plain path works, then add the faster path back in piece by piece.

[sparrowghost]

What worked for me was pretty simple. On Threadripper box, the easiest win is often to back off the aggressive settings and see whether the problem survives. A lot of these threads end up being one bad config plus a pile of conclusions drawn from it.

[cudascope]

Interesting, I hadn't tried that. The pattern around NVFP4 has been pretty similar for me. Worth rechecking your defaults though, because one bad assumption can make the rest of the thread look confusing.