[PATCH] D81515: [llvm] Release-mode ML InlineAdvisor

[Mircea Trofin via Phabricator via llvm-commits]

mtrofin added inline comments.


================
Comment at: llvm/include/llvm/Analysis/InlineModelRunner.h:21
+/// tensorflow "saved model".
+class InlineModelRunner {
+public:
----------------
davidxl wrote:
> Is this interface just for inliner or more general. Perhaps just name it MLModelRunner?
Renamed


================
Comment at: llvm/lib/Analysis/MLInlineAdvisor.cpp:37
+
+static cl::opt<float> SizeIncreaseThreshold(
+    "ml-advisor-size-increase-threshold", cl::Hidden,
----------------
davidxl wrote:
> Is this for controlling training time overhead?
Not only, it's also for controlling against misbehaving policies. Description was wrong, it's not native size increase, it's IR size.


================
Comment at: llvm/lib/Analysis/ReleaseModeModelRunner.cpp:31
+/// SavedModel for efficient execution.
+class ReleaseModeModelRunner final : public InlineModelRunner {
+public:
----------------
davidxl wrote:
> MLInferenceRunner?
What would we call the Development mode one then?


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D81515/new/

https://reviews.llvm.org/D81515