[Misc]: Enable memory usage logging for vLLM GPU worker

vllm/v1/worker/gpu_worker.py

@@ -228,21 +228,25 @@ def determine_available_memory(self) -> int:
         non_torch_alloc_bytes = max(0, fwd_alloc_bytes - torch_allocated_bytes)
         # Total forward allocation (peak) is peak torch + non-torch
         peak_memory = peak_torch_memory + non_torch_alloc_bytes
-
-        available_kv_cache_memory = (
-            total_gpu_memory * self.cache_config.gpu_memory_utilization -
-            peak_memory)
-
-        GiB = lambda b: b / GiB_bytes
-        logger.debug(
-            "Initial free memory: %.2f GiB, free memory: %.2f GiB, "
-            "total GPU memory: %.2f GiB", GiB(self.init_gpu_memory),
-            GiB(free_gpu_memory), GiB(total_gpu_memory))
-        logger.debug(
-            "Peak torch memory: %.2f GiB, non-torch forward-pass memory: "
-            "%.2f GiB, available KVCache memory: %.2f GiB",
-            GiB(peak_torch_memory), GiB(non_torch_alloc_bytes),
-            GiB(available_kv_cache_memory))
+        memory_for_current_instance = total_gpu_memory * \
+            self.cache_config.gpu_memory_utilization
+        available_kv_cache_memory = memory_for_current_instance - peak_memory
+
+        msg = (
+            "The current vLLM instance can use "
+            f"Initial free memory: {self.init_gpu_memory / GiB_bytes:.2f} GiB"
+            " x gpu_memory_utilization "
+            f"({self.cache_config.gpu_memory_utilization:.2f})"
+            f" = {memory_for_current_instance / GiB_bytes:.2f} GiB\n"
+            "Non torch memory takes "
+            f"{non_torch_alloc_bytes / GiB_bytes:.2f} GiB"
+            " and pytorch activation peak memory takes "
+            f"{peak_memory / GiB_bytes:.2f} GiB"
+            "and available KVCache memory is "
+            f"{available_kv_cache_memory / GiB_bytes:.2f} GiB\n"
+            "The rest of the memory reserved for KV Cache is "
+            f"{available_kv_cache_memory / GiB_bytes:.2f} GiB.")
+        logger.info(msg)
 
         return int(available_kv_cache_memory)