[llvm] [LAA] Add initial support for non-power-of-2 store-load forwarding distance (PR #137873)

[Alexey Bataev via llvm-commits]

================
@@ -1769,24 +1770,61 @@ bool MemoryDepChecker::couldPreventStoreLoadForward(uint64_t Distance,
       break;
     }
   }
+  // RISCV VLA supports non-power-2 vector factor. So, we iterate in a
+  // backward order to find largest VF, which allows aligned stores-loads or
+  // the number of iterations between conflicting memory addresses is not less
+  // than 8 (NumItersForStoreLoadThroughMemory).
+  if (AllowNonPow2Deps) {
+    MaxVFWithoutSLForwardIssuesNonPowerOf2 =
+        std::min(8 * VectorizerParams::MaxVectorWidth / TypeByteSize,
+                 MaxNonPowerOf2StoreLoadForwardSafeDistanceInBits);
+
+    for (uint64_t VF = MaxVFWithoutSLForwardIssuesNonPowerOf2;
+         VF > MaxVFWithoutSLForwardIssuesPowerOf2; VF -= TypeByteSize) {
+      if (Distance % VF == 0 ||
+          Distance / VF >= NumItersForStoreLoadThroughMemory) {
+        uint64_t GCD =
+            isSafeForAnyStoreLoadForwardDistances()
+                ? VF
+                : std::gcd(MaxNonPowerOf2StoreLoadForwardSafeDistanceInBits,
+                           VF);
+        MaxVFWithoutSLForwardIssuesNonPowerOf2 = GCD;
+        break;
+      }
+    }
+  }
 
-  if (MaxVFWithoutSLForwardIssuesPowerOf2 < 2 * TypeByteSize) {
+  if (MaxVFWithoutSLForwardIssuesPowerOf2 < 2 * TypeByteSize &&
+      MaxVFWithoutSLForwardIssuesNonPowerOf2 < 2 * TypeByteSize) {
     LLVM_DEBUG(
         dbgs() << "LAA: Distance " << Distance
                << " that could cause a store-load forwarding conflict\n");
     return true;
   }
 
+  // Handle non-power-2 store-load forwarding distance, power-of-2 distance can
+  // be calculated.
+  if (AllowNonPow2Deps && CommonStride &&
+      MaxVFWithoutSLForwardIssuesNonPowerOf2 <
+          MaxNonPowerOf2StoreLoadForwardSafeDistanceInBits &&
+      MaxVFWithoutSLForwardIssuesNonPowerOf2 !=
+          8 * VectorizerParams::MaxVectorWidth / TypeByteSize) {
+    uint64_t MaxVF = MaxVFWithoutSLForwardIssuesNonPowerOf2 / CommonStride;
+    uint64_t MaxVFInBits = MaxVF * TypeByteSize * 8;
+    MaxNonPowerOf2StoreLoadForwardSafeDistanceInBits =
----------------
alexey-bataev wrote:

Unfortunately, it won't work. I tried this, but there might be different results because of CommonStride value

https://github.com/llvm/llvm-project/pull/137873