Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
buildxyz
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
4 ms
·
1.
▲
by
buildxyz
1y ago
Any speed up that is 2x is definitely worth fixing. Especially since someone has already figured out the issue and performance testing [1] shows that llamacpp* is lagging behind vLLM by 2x. This is a positive for all running LLMs locally us