diff --git a/paddle/function/ConvOpTest.cpp b/paddle/function/ConvOpTest.cpp
index 59c7238d218a5adc4e19c9f6df627d7e203b965d..dfa2f784610b0dd60340e0ebc6a066437f3715eb 100644
--- a/paddle/function/ConvOpTest.cpp
+++ b/paddle/function/ConvOpTest.cpp
@@ -98,25 +98,112 @@ public:
   }
 };
 
+// Mainly used to test cases where the height and width (input, filter)
+// are not equal.
+template <DeviceType DType1, DeviceType DType2>
+class ConvolutionTest2 {
+public:
+  ConvolutionTest2(const std::string& conv1,
+                   const std::string& conv2,
+                   TestType type,
+                   std::string algo = "auto") {
+    for (size_t batchSize : {16}) {
+      for (size_t inputHeight : {7, 31}) {
+        for (size_t inputWidth : {10, 54}) {
+          for (size_t filterHeight : {1, 5}) {
+            for (size_t filterWidth : {3, 7}) {
+              for (size_t inputChannels : {7}) {
+                for (size_t outputChannels : {32}) {
+                  size_t stride = 1;
+                  size_t padding = 0;
+                  size_t outputHeight =
+                      (inputHeight - filterHeight + 2 * padding + stride) /
+                      stride;
+                  size_t outputWidth =
+                      (inputWidth - filterWidth + 2 * padding + stride) /
+                      stride;
+                  VLOG(3) << " batchSize=" << batchSize
+                          << " inputChannels=" << inputChannels
+                          << " inputHeight=" << inputHeight
+                          << " inputWidth=" << inputWidth
+                          << " outputChannels=" << outputChannels
+                          << " filterHeight=" << filterHeight
+                          << " filterWidth=" << filterWidth
+                          << " outputHeight=" << outputHeight
+                          << " outputWidth=" << outputWidth
+                          << " stride=" << stride << " padding=" << padding;
+
+                  std::vector<size_t> paddings = {padding, padding};
+                  std::vector<size_t> strides = {stride, stride};
+                  Compare2Function<DType1, DType2> test(
+                      conv1,
+                      conv2,
+                      FuncConfig()
+                          .set("paddings", paddings)
+                          .set("strides", strides)
+                          .set("groups", (size_t)1)
+                          .set("algo", algo));
+
+                  TensorShape input{
+                      batchSize, inputChannels, inputHeight, inputWidth};
+                  TensorShape filter{
+                      outputChannels, inputChannels, filterHeight, filterWidth};
+                  TensorShape output{
+                      batchSize, outputChannels, outputHeight, outputWidth};
+
+                  if (type == kForwardTest) {
+                    test.addInputs(BufferArg(VALUE_TYPE_FLOAT, input));
+                    test.addInputs(BufferArg(VALUE_TYPE_FLOAT, filter));
+                    test.addOutputs(BufferArg(VALUE_TYPE_FLOAT, output));
+                    test.run();
+                  } else if (type == kBackwardInputTest) {
+                    test.addInputs(BufferArg(VALUE_TYPE_FLOAT, output));
+                    test.addInputs(BufferArg(VALUE_TYPE_FLOAT, filter));
+                    test.addOutputs(BufferArg(VALUE_TYPE_FLOAT, input), ADD_TO);
+                    test.run();
+                  } else if (type == kBackwardFilterTest) {
+                    test.addInputs(BufferArg(VALUE_TYPE_FLOAT, output));
+                    test.addInputs(BufferArg(VALUE_TYPE_FLOAT, input));
+                    test.addOutputs(BufferArg(VALUE_TYPE_FLOAT, filter));
+                    test.run();
+                  }
+                }
+              }
+            }
+          }
+        }
+      }
+    }
+  }
+};
+
 TEST(Forward, GEMM) {
   ConvolutionTest<DEVICE_TYPE_CPU, DEVICE_TYPE_CPU> test(
       "NaiveConv-CPU", "GemmConv-CPU", kForwardTest);
+  ConvolutionTest2<DEVICE_TYPE_CPU, DEVICE_TYPE_CPU> test2(
+      "NaiveConv-CPU", "GemmConv-CPU", kForwardTest);
 }
 
 #ifndef PADDLE_ONLY_CPU
 TEST(Forward, GEMM2) {
   ConvolutionTest<DEVICE_TYPE_CPU, DEVICE_TYPE_GPU> test(
       "GemmConv-CPU", "GemmConv-GPU", kForwardTest);
+  ConvolutionTest2<DEVICE_TYPE_CPU, DEVICE_TYPE_GPU> test2(
+      "GemmConv-CPU", "GemmConv-GPU", kForwardTest);
 }
 
 TEST(BackwardInput, GEMM) {
   ConvolutionTest<DEVICE_TYPE_CPU, DEVICE_TYPE_GPU> test(
       "GemmConvGradInput-CPU", "GemmConvGradInput-GPU", kBackwardInputTest);
+  ConvolutionTest2<DEVICE_TYPE_CPU, DEVICE_TYPE_GPU> test2(
+      "GemmConvGradInput-CPU", "GemmConvGradInput-GPU", kBackwardInputTest);
 }
 
 TEST(BackwardFilter, GEMM) {
   ConvolutionTest<DEVICE_TYPE_CPU, DEVICE_TYPE_GPU> test(
       "GemmConvGradFilter-CPU", "GemmConvGradFilter-GPU", kBackwardFilterTest);
+  ConvolutionTest2<DEVICE_TYPE_CPU, DEVICE_TYPE_GPU> test2(
+      "GemmConvGradFilter-CPU", "GemmConvGradFilter-GPU", kBackwardFilterTest);
 }
 #endif
 
diff --git a/paddle/function/NaiveConvOp.cpp b/paddle/function/NaiveConvOp.cpp
index 94aba253e3e56d8a26429d64d539f29ed97ff30a..1d204f99e0e127688eeda28b46715a37c1100c4e 100644
--- a/paddle/function/NaiveConvOp.cpp
+++ b/paddle/function/NaiveConvOp.cpp
@@ -104,7 +104,7 @@ public:
     size_t inputHeight = inputs[0].shape()[2];
     size_t inputWidth = inputs[0].shape()[3];
     size_t filterHeight = inputs[1].shape()[2];
-    size_t filterWidth = inputs[1].shape()[2];
+    size_t filterWidth = inputs[1].shape()[3];
     size_t outputChannels = outputs[0].shape()[1];
     size_t outputHeight = outputs[0].shape()[2];
     size_t outputWidth = outputs[0].shape()[3];