Enable autograd cache on inductor tests (#140890)

Summary:
This turns on AOTAutogradCache for all inductor tests. It clears AOTAutogradCache on each test as well, by virtue of the local cache using the same directory to store cache entries.

I've also tested with INDUCTOR_TEST_DISABLE_FRESH_CACHE=1, running all the tests. AOTAutogradCache successfully caches 99% of these. There are a few tests that use view_replay and therefore save functional tensors, which cause AOTAutogradCache to fail to pickle its result. Will look into next steps there, but for now, it seems okay if the cache just misses on those cases where it can't serialize the result. It would be better to check before pickling, though.

I've made the following small bugfixes to get this working:
- Inductor is sometimes used in a standalone mode without dynamo, which leads to attribute errors in check_can_cache. In general, we should *never* crash in cache checking, only bypass. So I change a try catch to check Exception instead of just a specific exception.
- Add extra structured logging for metadata on cache hits

X-link: pytorch/pytorch#140890
Approved by: https://github.com/bdhirsh

Reviewed By: atalman

Differential Revision: D66556085

Pulled By: jamesjwu

fbshipit-source-id: d1379a9946afca524e289459217845e29e97e142

userbenchmark/dynamo/dynamobench/_dynamo/utils.py

@@ -1130,6 +1130,14 @@ def __init__(self):
         # TODO: log to init/id tlparse after I add support for it
         log.info("ChromiumEventLogger initialized with id %s", self.id_)
 
+    def try_add_event_data(self, event_name: str, **kwargs) -> None:
+        """
+        Same as add_event_data, but will silently not log if the event isn't in the stack.
+        """
+        if event_name not in self.get_stack():
+            return
+        self.add_event_data(event_name, **kwargs)
+
     def add_event_data(
         self,
         event_name: str,

userbenchmark/dynamo/dynamobench/huggingface.py

@@ -37,6 +37,10 @@
 if "TORCHINDUCTOR_FX_GRAPH_CACHE" not in os.environ:
     torch._inductor.config.fx_graph_cache = True
 
+# Enable Autograd caching
+if "TORCHINDUCTOR_AUTOGRAD_CACHE" not in os.environ:
+    torch._functorch.config.enable_autograd_cache = True
+
 
 def pip_install(package):
     subprocess.check_call([sys.executable, "-m", "pip", "install", package])

userbenchmark/dynamo/dynamobench/torchbench.py

@@ -29,6 +29,10 @@
 if "TORCHINDUCTOR_FX_GRAPH_CACHE" not in os.environ:
     torch._inductor.config.fx_graph_cache = True
 
+# Enable Autograd caching
+if "TORCHINDUCTOR_AUTOGRAD_CACHE" not in os.environ:
+    torch._functorch.config.enable_autograd_cache = True
+
 
 def _reassign_parameters(model):
     # torch_geometric models register parameter as tensors due to