perf: accelerate longhaul prompt loading

This commit is contained in:
Owen Qwen
2026-07-29 12:44:50 -05:00
parent c2baa66534
commit fdd9fcf85e
7 changed files with 525 additions and 55 deletions
+1
View File
@@ -56,6 +56,7 @@ test parameters:
-ub, --ubatch-size <n> (default: 512)
-ctk, --cache-type-k <t> (default: f16)
-ctv, --cache-type-v <t> (default: f16)
--longhaul-cache <GiB> expert cache size for longhaul mode
-t, --threads <n> (default: system dependent)
-C, --cpu-mask <hex,hex> (default: 0x0)
--cpu-strict <0|1> (default: 0)