[PATCH] D72714: [ARM][MVE] Tail-Predication: rematerialise iteration count in exit blocks

Dave Green via Phabricator via llvm-commits llvm-commits at lists.llvm.org
Wed Jan 15 08:49:34 PST 2020


dmgreen added inline comments.


================
Comment at: llvm/lib/Target/ARM/MVETailPredication.cpp:141
+
+  if (!L->isRecursivelyLCSSAForm(*DT, *LI)) {
+    LLVM_DEBUG(dbgs() << "ARM TP: Loop is not in LCSSA form\n");
----------------
Do you need to test this? Will formLCSSA either just work or not by calling it?

I just noticed an EXPENSIVE_CHECKS in formLCSSA that checks if the subloops are already in LCSSA form. Maybe use formLCSSARecursively instead?


================
Comment at: llvm/lib/Target/ARM/MVETailPredication.cpp:144-145
+    if (!formLCSSA(*L, *DT, LI, SE)) {
+      LLVM_DEBUG(dbgs() << "ARM TP: couldn't bring loop in LCCS form,"
+                        << "can't rematerialise iteration count in exit "
+                        << "blocks\n");
----------------
Space after the ","


================
Comment at: llvm/lib/Target/ARM/MVETailPredication.cpp:147
+                        << "blocks\n");
+      return;
+    } else
----------------
Does formLCSSA return false on failure, or nothing changed?


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D72714/new/

https://reviews.llvm.org/D72714





More information about the llvm-commits mailing list