diff --git a/scripts/start-server.sh b/scripts/start-server.sh index 2204ec4..60e4c37 100755 --- a/scripts/start-server.sh +++ b/scripts/start-server.sh @@ -24,7 +24,7 @@ exec /opt/llama.cpp/build-nccl/bin/llama-server \ --split-mode tensor \ --flash-attn on \ --batch-size 1024 \ - --ubatch-size 256 \ + --ubatch-size 512 \ --jinja \ --threads 12 \ --cache-type-k q4_0 \ diff --git a/tuning/cmp-tune.conf b/tuning/cmp-tune.conf new file mode 100644 index 0000000..dca5858 --- /dev/null +++ b/tuning/cmp-tune.conf @@ -0,0 +1,13 @@ +# /etc/cmp-tune.conf - Profile-driven tuning for CMP 50HX / CMP 90HX + +[efficient] +power_w = 170 +core_offset = 200 +mem_offset = 1000 +clock_max = 1950 + +[insane] +power_w = 220 +core_offset = 250 +mem_offset = 1200 +clock_max = 2050