update llama.cpp API 20250719

JamePeng · JamePeng · commit e270eb15131c · 2025-07-19T22:40:00.000+08:00
diff --git a/llama_cpp/llama_cpp.py b/llama_cpp/llama_cpp.py
@@ -4041,6 +4041,7 @@ def llama_log_set(
 #
 #     int32_t n_p_eval;
 #     int32_t n_eval;
+#     int32_t n_reused; // number of times a ggml compute graph had been reused
 # };
 class llama_perf_context_data(ctypes.Structure):
     _fields_ = [
@@ -4050,6 +4051,7 @@ class llama_perf_context_data(ctypes.Structure):
         ("t_eval_ms", ctypes.c_double),
         ("n_p_eval", ctypes.c_int32),
         ("n_eval", ctypes.c_int32),
+        ("n_reused", ctypes.c_int32),
     ]
 
 

Original file line number	Diff line number	Diff line change
`@@ -4041,6 +4041,7 @@ def llama_log_set(`
`4041`	`4041`	`#`
`4042`	`4042`	`# int32_t n_p_eval;`
`4043`	`4043`	`# int32_t n_eval;`
	`4044`	`+# int32_t n_reused; // number of times a ggml compute graph had been reused`
`4044`	`4045`	`# };`
`4045`	`4046`	`class llama_perf_context_data(ctypes.Structure):`
`4046`	`4047`	`_fields_ = [`
`@@ -4050,6 +4051,7 @@ class llama_perf_context_data(ctypes.Structure):`
`4050`	`4051`	`("t_eval_ms", ctypes.c_double),`
`4051`	`4052`	`("n_p_eval", ctypes.c_int32),`
`4052`	`4053`	`("n_eval", ctypes.c_int32),`
	`4054`	`+ ("n_reused", ctypes.c_int32),`
`4053`	`4055`	`]`
`4054`	`4056`
`4055`	`4057`