when TensorRT-LLM clearly wins with SGLang

[kernelmike]

What worked for me was pretty simple. Before rebuilding anything, I'd test one clean baseline with Qwen and compare from there. Once you know the plain path works, then add the faster path back in piece by piece.

[batchcraft]

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

[kernelmike]

My first guess would be this: On 7900 XTX, the easiest win is often to back off the aggressive settings and see whether the problem survives. Not glamorous advice, but repeatable baselines save a lot of time.

[kernelgrid]

Tried to reproduce it just now. On B200 with SGLang, I saw basically the same behavior. The main thing that moved was latency, not quality.

[vectorport]

following this because I'm in the same boat.

[syshouse]

following this because I'm in the same boat.