Model techniques map
Techniquesinference & servinginference quantization

implementation detail · filed under inference & serving

TRT-LLM inference optimization to NVFP4 on Blackwell

Optimizes TRT-LLM inference for NVFP4 on Blackwell hardware; the evidence does not specify a distinct quantization algorithm.

source
1
model
1
lab adopt it
1
strongest
used

How sources treat it

One count per evidence span, weakest treatment to strongest.

used 1

Documented in

Evidence

1 span quoted from the sources, strongest treatment first.

Poolside worked with NVIDIA to optimize inference from TRT-LLM serving to NVFP4 on Blackwell, down to a single DGX Spark.

usedinference servingin Laguna S 2.1Poolside and NVIDIA

Filed alongside

Other methods under inference & serving :: inference quantization.