[Bugfix] Fix DeepSeek V3.2 OOM during CG memory profiling (#36691)

Signed-off-by: Matthew Bonanni <mbonanni@redhat.com>

vllm/v1/worker/gpu_model_runner.py

@@ -5550,16 +5550,14 @@ class GPUModelRunner(
         kv_cache_spec = self.get_kv_cache_spec()
         kv_cache_groups = get_kv_cache_groups(self.vllm_config, kv_cache_spec)
         min_blocks = self.compilation_config.max_cudagraph_capture_size or 1
-        if kv_cache_groups:
-            page_size = kv_cache_groups[0].kv_cache_spec.page_size_bytes
-            group_size = max(len(g.layer_names) for g in kv_cache_groups)
-            available_memory = min_blocks * page_size * group_size
-        else:
-            available_memory = 1  # Attention-free model
 
+        # Temporarily change num_gpu_blocks_override to allocate a minimal KV cache
+        saved_override = self.cache_config.num_gpu_blocks_override
+        self.cache_config.num_gpu_blocks_override = min_blocks
         minimal_config = get_kv_cache_config_from_groups(
-            self.vllm_config, kv_cache_groups, available_memory=available_memory
+            self.vllm_config, kv_cache_groups, available_memory=0
         )
+        self.cache_config.num_gpu_blocks_override = saved_override
 
         self.initialize_kv_cache(minimal_config)
         self.cache_config.num_gpu_blocks = minimal_config.num_blocks