UEFI/BIOS -> Advanced -> AMD CBS -> NBIO -> GFX Configuration >
iGPU Advanced Control > Disabled
Dedicated Graphics Memory > Medium (16GB)
Remaining System Memory > 48GB
Different memory settings tried for Ollama, no significant difference.
Tested prompt: What is machine learning?
Model: deepseek-r1:1.5b
load duration: 12.277249ms
prompt eval count: 8 token(s)
prompt eval duration: 53ms
prompt eval rate: 150.94 tokens/s
eval count: 177 token(s)
eval duration: 3.808s
eval rate: 46.48 tokens/s
Model: deepseek-r1:7b
total duration: 35.692882833s
load duration: 24.075692ms
prompt eval count: 8 token(s)
prompt eval duration: 254ms
prompt eval rate: 31.50 tokens/s
eval count: 563 token(s)
eval duration: 35.412s
eval rate: 15.90 tokens/s
Model: deepseek-r1:8b
total duration: 19.19488017s
load duration: 19.389461ms
prompt eval count: 8 token(s)
prompt eval duration: 204ms
prompt eval rate: 39.22 tokens/s
eval count: 284 token(s)
eval duration: 18.969s
eval rate: 14.97 tokens/s
Model: deepseek-r1:14b
total duration: 1m29.657527203s
load duration: 19.170988ms
prompt eval count: 8 token(s)
prompt eval duration: 376ms
prompt eval rate: 21.28 tokens/s
eval count: 660 token(s)
eval duration: 1m29.261s
eval rate: 7.39 tokens/s
Model: deepseek-r1:32b
total duration: 5m53.281645538s
load duration: 18.26359ms
prompt eval count: 8 token(s)
prompt eval duration: 852ms
prompt eval rate: 9.39 tokens/s
eval count: 1308 token(s)
eval duration: 5m52.409s
eval rate: 3.71 tokens/s
Model: deepseek-r1:70b
total duration: 3m52.784148477s
load duration: 31.808434ms
prompt eval count: 8 token(s)
prompt eval duration: 2s
prompt eval rate: 4.00 tokens/s
eval count: 406 token(s)
eval duration: 3m50.749s
eval rate: 1.76 tokens/s