Community
Developer shows llama.cpp PR aiming to beat Strata decode performance on an R9700
A Reddit post links to a llama.cpp pull request on decode performance on AMD's R9700 GPU, compared against the Strata decode baseline.
Every outlet that ran it 1 article
r/LocalLLaMACommunityLead
Llama.cpp PR: How to beat strata decode performance with a R9700