Re:NVIDIA DGX Spark In-Depth Review: A New Standard for Loca...
by yvbbrjdr on Tuesday October 14, 2025 @03:05AM
(#45577061)
We actually profiled one of the models, and saw that the last GeMM, which is completely memory bound, is taking a lot of time, which reduces the token speed by a lot.
[ Parent
| on HN ]