llm, add context param to ollama_threads.

tasks/workstation/linux/software/llm.yml

@@ -909,12 +909,20 @@
           alias model-sizes="ollama_sizes "
 
           function ollama_threads() {
+            # VAR1 - Model
             model="$1"
             if [[ -z $model ]]; then
               model="qwen3.5:4b"
             fi
 
-            #echo " Stopping all ollama usage..."
+            # VAR2 - Context
+            typeset -i num_ctx="$2"
+            context=""
+            if [[ -n $num_ctx ]]; then
+              context=',"num_ctx":'$num_ctx''
+            fi
+
+            echo " Stopping all ollama usage..."
             sudo sv restart ollama >/dev/null 2>&1
 
             # Benchmark Thread Count
@@ -922,7 +930,7 @@
               echo -n "`date` - Model $model w/ $t Thread(s), tokens/sec: "
 
               speed=$(curl -s http://localhost:11434/api/generate \
-                -d '{"model":"'$model'","prompt":"Hello","stream":false,"keep_alive":1,"options":{"num_thread":'$t'}}' \
+                -d '{"model":"'$model'","prompt":"Hello","stream":false,"keep_alive":1,"options":{"num_thread":'$t'$context}}' \
                 | jq '.eval_count * 1000000000 / .eval_duration'
               )
               ctx="`ollama ps | grep $model | awk '{print $7}'`"