HIP kernel compile issue

[phasetrack]

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

I'm helping a small team set up a local workflow. We don't need a giant stack, but we do care about INT4.

Would rather learn from your scars than collect my own all at once.

[tensorworks]

I'd approach it in a boring order. Before rebuilding anything, I'd test one clean baseline with FLUX and compare from there. Not glamorous advice, but repeatable baselines save a lot of time.

[rocmriley]

I'd approach it in a boring order. Before rebuilding anything, I'd test one clean baseline with FLUX and compare from there. Not glamorous advice, but

worth checking the driver first.

[labline]

[quote]I'd approach it in a boring order. Before rebuilding anything, I'd test one clean baseline with FLUX and compare from there. Not glamorous advi

Interesting, I hadn't tried that. I don't think your conclusion is far off. Worth rechecking your defaults though, because one bad assumption can make the rest of the thread look confusing.

[quietarch]

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

[dockworks23]

My first guess would be this: On M3 Max, 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.

[torchsmith25]

That matches what I was seeing. I ended up in about the same place using TensorRT-LLM.

[thetarunner]

That matches what I was seeing. I ended up in about the same place using TensorRT-LLM.

that fixed it for me too.

[rocmriley]

That matches what I was seeing. I ended up in about the same place using TensorRT-LLM.

A couple things I'd check first. On 4090, the easiest win is often to back off the aggressive settings and see whether the problem survives. Once you know the plain path works, then add the faster path back in piece by piece.

[rocmmixer]

What worked for me was pretty simple. Start by verifying the driver/runtime combo, then test with a smaller prompt, then look at VRAM headroom. Once you know the plain path works, then add the faster path back in piece by piece.