llama.cpp

#1
by FrescoHF - opened

llama.cpp version b4970 has a performance improvement. How about updating llama.cpp more often? There are important changes coming out there all the time, especially lately

the version i use to make usually has no relevance to performance, you gain the uplift by updating your own engine

I just include that information for reference

The only time it matters is if there are optimizations for how a quant is made, not how it's run

Your need to confirm your account before you can post a new comment.

Sign up or log in to comment