[PATCH] D106033: [OpenMP] Folding threadLimit and numThreads when single value in kernels

[Johannes Doerfert via Phabricator via llvm-commits]

jdoerfert added inline comments.


================
Comment at: llvm/lib/Transforms/IPO/OpenMPOpt.cpp:1777-1778
 
+  /// Populate the Attributor with abstract attribute opportunities in the
+  /// function.
+  void registerFoldRuntimeCall(RuntimeFunction RF);
----------------
copy and paste.


================
Comment at: llvm/lib/Transforms/IPO/OpenMPOpt.cpp:3663
+        *this, IRPosition::function(*getAnchorScope()), DepClassTy::REQUIRED);
+
+    // What actual kernels reach this function
----------------
```
if (!CallerKernelInfoAA.ReachingKernelEntries.isValid())
  return indicatePessimisticFixpoint();
```
also below


================
Comment at: llvm/lib/Transforms/IPO/OpenMPOpt.cpp:3674
+      }
+      return indicatePessimisticFixpoint();
+    }
----------------
use early exit instead,
if (!...)
  return ...



================
Comment at: llvm/lib/Transforms/IPO/OpenMPOpt.cpp:3683
+    }
+    return indicatePessimisticFixpoint();
+  }
----------------
no reaching kernels is fine, keep it at none (which is the default) and just return UNCHANGED.


================
Comment at: llvm/lib/Transforms/IPO/OpenMPOpt.cpp:3684
+    return indicatePessimisticFixpoint();
+  }
+
----------------
make this function take the string attribute such that we can have a single one and not two.


================
Comment at: llvm/lib/Transforms/IPO/OpenMPOpt.cpp:3736
+    CallInst *CI = OpenMPOpt::getCallIfRegularCall(U, &RFI);
+    //errs() << IRPosition::callsite_function(*CI).getPositionKind() << " -- " << RF;
+    if (!CI)
----------------
leftover. (and you can also just pipe an IRPosition into errs())


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D106033/new/

https://reviews.llvm.org/D106033