Model techniques map
Techniquesinference & servinginference kernel

specific method · filed under inference & serving

Sparse pinned-host offload

Keeps vocabulary-parallel embedding shards in pinned host memory and gathers selected rows into a small GPU buffer.

source
1
model
1
lab adopt it
1
strongest
default

How sources treat it

One count per evidence span, weakest treatment to strongest.

default 1

Documented in

Evidence

1 span quoted from the sources, strongest treatment first.

SGLang keeps each rank's vocabulary-parallel table shard in pinned host memory and gathers the selected rows into a small BF16 GPU buffer with a Triton UVA kernel.

defaultunclearin Qwen3.8-Flash-NextSGLang

Filed alongside

Other methods under inference & serving :: inference kernel.