A quick note about forum organization with TensorRT-LLM

[Data Sculpting]

Small forum update from the Data Sculpting side.

We're still tuning the structure and getting everything in place. Feedback is welcome, especially if a topic feels like it has no obvious home yet.

[Data Sculpting]

What exact setup are you running? Do you see the same thing with a smaller batch or shorter prompt?

[cozykite]

What worked for me was pretty simple. Start by verifying the driver/runtime combo, then test with a smaller prompt, then look at speculative decoding. Not glamorous advice, but repeatable baselines save a lot of time.

There are a few different conversations hiding inside this thread, so being explicit about the goal helps.

[Data Sculpting]

Ran a quick check on my box. On 4090 with LM Studio, I saw basically the same behavior. The surprise was that memory pressure showed up before throughput collapsed.

[Data Sculpting]

worth checking the driver first.

[Data Sculpting]

that fixed it for me too.

[dockworks23]

What worked for me was pretty simple. If you're on PyTorch, I'd confirm the version first and then reduce variables until the behavior is obvious. Not glamorous advice, but repeatable baselines save a lot of time.

[Data Sculpting]

My first guess would be this: On GB200, 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.

[quantsupply]

I tested this tonight. On dual 3090 with PyTorch, I saw basically the same behavior.

[Data Sculpting]

A couple things I'd check first. If you're on SGLang, 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.