diff --git a/src/Futhark/CodeGen/ImpGen/GPU/Group.hs b/src/Futhark/CodeGen/ImpGen/GPU/Group.hs
index dbef2ec170..a337ab9732 100644
--- a/src/Futhark/CodeGen/ImpGen/GPU/Group.hs
+++ b/src/Futhark/CodeGen/ImpGen/GPU/Group.hs
@@ -684,19 +684,25 @@ data Precomputed = Precomputed
 segOpSizes :: Stms GPUMem -> SegOpSizes
 segOpSizes = onStms
   where
-    onStms = foldMap (onExp . stmExp)
-    onExp (Op (Inner (SegOp op))) =
+    onStms = foldMap onStm
+    onStm (Let _ _ (Op (Inner (SegOp op)))) =
       case segVirt $ segLevel op of
         SegNoVirtFull seq_dims ->
           S.singleton $ map snd $ snd $ partitionSeqDims seq_dims $ segSpace op
         _ -> S.singleton $ map snd $ unSegSpace $ segSpace op
-    onExp (BasicOp (Replicate shape _)) =
-      S.singleton $ shapeDims shape
-    onExp (Match _ cases defbody _) =
+    onStm (Let (Pat [pe]) _ (BasicOp (Replicate {}))) =
+      S.singleton $ arrayDims $ patElemType pe
+    onStm (Let (Pat [pe]) _ (BasicOp (Iota {}))) =
+      S.singleton $ arrayDims $ patElemType pe
+    onStm (Let (Pat [pe]) _ (BasicOp (Copy {}))) =
+      S.singleton $ arrayDims $ patElemType pe
+    onStm (Let (Pat [pe]) _ (BasicOp (Manifest {}))) =
+      S.singleton $ arrayDims $ patElemType pe
+    onStm (Let _ _ (Match _ cases defbody _)) =
       foldMap (onStms . bodyStms . caseBody) cases <> onStms (bodyStms defbody)
-    onExp (DoLoop _ _ body) =
+    onStm (Let _ _ (DoLoop _ _ body)) =
       onStms (bodyStms body)
-    onExp _ = mempty
+    onStm _ = mempty
 
 -- | Precompute various constants and useful information.
 precomputeConstants :: Count GroupSize (Imp.TExp Int64) -> Stms GPUMem -> CallKernelGen Precomputed