feat(benchmark): add --kv-types flag for KV cache quantization sweep

2026-03-27 12:29:19 +01:00
parent 38daf953bf
commit 7531f6fa74
4 changed files with 176 additions and 82 deletions
--- a/bin/benchmark
+++ b/bin/benchmark
@@ -23,10 +23,12 @@ case "${1:-help}" in
        echo "  --category LIST     Comma-separated: smoke,dense,moe"
        echo "  --skip-longctx      Skip long-context (32K) tests"
        echo "  --reps N            Standard test repetitions (default: 5)"
+        echo "  --kv-types LIST     KV cache sweep (e.g. f16,q8_0,q4_0 or q4_0:q8_0)"
        echo ""
        echo "Examples:"
        echo "  benchmark baseline --max-size 20 --skip-longctx"
        echo "  benchmark run --tag post-opt --category moe"
+        echo "  benchmark run --tag kv-sweep --kv-types f16,q8_0,q4_0 --context 131072"
        exit 1
        ;;
 esac