diff --git a/lib/Env.chpl b/lib/Env.chpl
index 2194ce1a..b818da7b 100644
--- a/lib/Env.chpl
+++ b/lib/Env.chpl
@@ -6,7 +6,7 @@ param developmentAndTesting = !releaseChAI;
 
 
 // Minimum needed rank of dynamicTensor for *any* build
-config param minRankNeeded = 6;
+config param minRankNeeded = 4;
 
 
 // Maximum needed rank of dynamicTensor for a release build
diff --git a/lib/NDArray.chpl b/lib/NDArray.chpl
index 86f034e9..bc396145 100644
--- a/lib/NDArray.chpl
+++ b/lib/NDArray.chpl
@@ -149,24 +149,39 @@ record ndarray : serializable {
         _domain = dom;
     }
 
-    proc reshape(const dom: ?t): ndarray(rank,eltType)
+    proc reshape(const dom: ?t): ndarray(dom.rank,eltType)
             where isDomainType(t)
-                && dom.rank == rank {
+                /* && dom.rank == rank */ {
         var arr = new ndarray(eltType,dom);
-        const arrDom  = arr.domain;
-        const selfDom = _domain;
- 
-        const inter = selfDom[arrDom];
-        arr.data[inter] = data[inter];
+
+        const arrDom = arr.domain;
+        const selfDom = this.domain;
+        
+        ref arrData = arr.data;
+        const ref selfData = this.data;
+
+        const arrShape = arrDom.shape;
+        const selfShape = selfDom.shape;
+        const selfShapeDivs = util.shapeDivisors((...selfShape));
+
+        const zero: eltType = 0;
+
+        forall (i,idx) in arrDom.everyZip() {
+            const selfIdx = util.indexAtHelperMultiples(i,(...selfShapeDivs));
+            const a = if util.shapeContains(selfShape,selfIdx)
+                        then selfData[selfIdx]
+                        else zero;
+            arrData[idx] = a;
+        }
         return arr;
     }
-
+/*
     proc reshape(const dom: ?t): ndarray(dom.rank,eltType)
             where isDomainType(t)
                 && dom.rank != rank {
 
         var arr: ndarray(dom.rank,eltType) = new ndarray(eltType,dom);
-
+        compilerError("Testing. Don't call me.");
         const selfDom = this.domain;
         const newDom  = arr.domain;
         const ref selfData = this.data;
@@ -200,7 +215,7 @@ record ndarray : serializable {
         
         // return arr;
     }
-
+*/
 
     // This can optimized such that it doesn't use two heavy utility functions...
     proc reshape(newShape: int ...?newRank): ndarray(newRank,eltType) {
diff --git a/lib/Utilities.chpl b/lib/Utilities.chpl
index cb5537d2..504e802e 100644
--- a/lib/Utilities.chpl
+++ b/lib/Utilities.chpl
@@ -193,6 +193,26 @@ module Utilities {
         return idxs;
     }
 
+    inline proc shapeDivisors(shape: int...?rank): rank*int {
+        var prod = 1;
+        var divs: rank * int;
+        for param j in 0..<rank {
+            param i = rank - j - 1;
+            divs(i) = prod;
+            prod *= shape(i);
+        }
+        return divs;
+    }
+
+    inline proc shapeProduct(shape: int...?rank): int {
+        if rank == 1 {
+            return 1;
+        } else {
+            const divs = shapeDivisors((...shape));
+            return divs(rank - 1) * shape(rank - 1);
+        }
+    }
+
     inline proc indexAtHelperProd(n: int, prod: int, shape: int ...?rank): rank * int where rank > 1 {
         var idx: rank * int;
         var order = n;
@@ -304,6 +324,14 @@ module Utilities {
         return arr;
     }
 
+    inline proc shapeContains(shape: ?rank*int, idx: ?idxRank*int): bool
+            where rank == idxRank {
+        var contained = true;
+        for param i in 0..<rank do
+            contained &= idx(i) < shape(i);
+        return contained;
+    }
+
     module Standard {
         private use ChplConfig;
 
diff --git a/test/correspondence/movement/reshape/reshape.chpl b/test/correspondence/movement/reshape/reshape.chpl
new file mode 100644
index 00000000..c751ed9c
--- /dev/null
+++ b/test/correspondence/movement/reshape/reshape.chpl
@@ -0,0 +1,19 @@
+use Tensor;
+
+var a = dynamicTensor.arange(2,3);
+
+Testing.numericPrint(a.reshape(3,2));
+
+Testing.numericPrint(a.reshape(6));
+
+Testing.numericPrint(a.reshape(1,2,3));
+
+Testing.numericPrint(a.reshape(1,1,6));
+
+Testing.numericPrint(a.reshape(6,1,1));
+
+Testing.numericPrint(a.reshape(2,1,1,3));
+
+Testing.numericPrint(a.reshape(3,1,1,2));
+
+Testing.numericPrint(a.reshape(1,3,2,1));
diff --git a/test/correspondence/movement/reshape/reshape.py b/test/correspondence/movement/reshape/reshape.py
new file mode 100644
index 00000000..104dbc29
--- /dev/null
+++ b/test/correspondence/movement/reshape/reshape.py
@@ -0,0 +1,23 @@
+import torch
+
+def test(imports):
+    print = imports['print_fn']
+
+    a = torch.arange(6,dtype=torch.float32).reshape(2,3)
+
+    print(a.reshape(3,2))
+    
+    print(a.reshape(6))
+
+    print(a.reshape(1,2,3))
+
+    print(a.reshape(1,1,6))
+
+    print(a.reshape(6,1,1))
+
+    print(a.reshape(2,1,1,3))
+
+    print(a.reshape(3,1,1,2))
+
+    print(a.reshape(1,3,2,1))
+
diff --git a/test/tiny/dynamic_shape.chpl b/test/tiny/dynamic_shape.chpl
index ec4f6834..2e1f56c1 100644
--- a/test/tiny/dynamic_shape.chpl
+++ b/test/tiny/dynamic_shape.chpl
@@ -1,6 +1,8 @@
 use Tensor;
 
 use List;
+import Utilities as util;
+
 /*
 var a = dynamicTensor.arange(2,3);
 
@@ -22,7 +24,7 @@ writeln(b.shape());
 
 writeln(b.shape().toList());
 */
-
+/*
 var c = dynamicTensor.arange(4) + 1;
 writeln(c);
 
@@ -36,4 +38,27 @@ writeln(d.squeeze());
 writeln(d.squeeze(1));
 
 writeln(d.reshape(new dynamicShape((2,2))));
-writeln(d.reshape(dShape=new dynamicShape((4,))));
\ No newline at end of file
+writeln(d.reshape(dShape=new dynamicShape((4,))));
+
+writeln(d.reshape(2,2));
+*/
+
+/*
+var a = dynamicTensor.arange(2,3);
+
+writeln(a);
+writeln(a.forceRank(2).array);
+
+writeln(a.reshape(3,2).forceRank(2).array);
+writeln(a.reshape(3,2).forceRank(2).array);
+
+*/
+
+var a = ndarray.arange(2,3);
+writeln(a);
+writeln(a.reshape(3,2));
+writeln(a.reshape(3,2));
+
+
+writeln(a.reshape(6));
+writeln(a.reshape(1,1,6));
diff --git a/test/tiny/tinygradNB.ipynb b/test/tiny/tinygradNB.ipynb
index c40b88c1..c467eaa4 100644
--- a/test/tiny/tinygradNB.ipynb
+++ b/test/tiny/tinygradNB.ipynb
@@ -122,6 +122,45 @@
     "print(t.unsqueeze(1).unsqueeze(1).squeeze(1).numpy().shape)"
    ]
   },
+  {
+   "cell_type": "code",
+   "execution_count": 22,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "[[0 1 2]\n",
+      " [3 4 5]]\n"
+     ]
+    }
+   ],
+   "source": [
+    "a = Tensor.arange(6).reshape(2,3)\n",
+    "print(a.numpy())"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 23,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "[[0 1]\n",
+      " [2 3]\n",
+      " [4 5]]\n"
+     ]
+    }
+   ],
+   "source": [
+    "b = a.reshape(3,2)\n",
+    "print(b.numpy())"
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": null,