diff --git a/src/com/custom_computing_ic/dfe_snippets/kernels/PaddingKernel.java b/src/com/custom_computing_ic/dfe_snippets/kernels/PaddingKernel.java
new file mode 100644
index 0000000..ec756a0
--- /dev/null
+++ b/src/com/custom_computing_ic/dfe_snippets/kernels/PaddingKernel.java
@@ -0,0 +1,48 @@
+package com.custom_computing_ic.dfe_snippets.kernels;
+
+import com.maxeler.maxcompiler.v2.kernelcompiler.Kernel;
+import com.maxeler.maxcompiler.v2.kernelcompiler.KernelParameters;
+import com.maxeler.maxcompiler.v2.kernelcompiler.types.base.DFEVar;
+import com.maxeler.maxcompiler.v2.kernelcompiler.stdlib.core.CounterChain;
+
+/**
+ * Pads the inputs with 0 after nInputs have been processed.
+ *
+ * Scalars:
+ * NUM_INP - number of inputs that will be received from a previous kernel.
+ * TOTAL_CYCLES - number of cycles that this kernel should give an output.
+ *
+ * @since 18/05/2017
+ */
+public class PaddingKernel extends Kernel {
+  public static final String INP_NAME            = "PADDING_INP";
+  public static final String OUT_NAME            = "PADDING_OUT";
+  public static final String SCALAR_NUM_INP      = "NUM_INP";
+  public static final String SCALAR_TOTAL_CYCLES = "TOTAL_CYCLES";
+
+  /**
+   * Constructor of the Padding Kernel.
+   *
+   * @param parameters parameters for this kernel passed from a manager
+   * @param bitWidth number of bits to be processed in each cycle
+   * @param dbg flag to decide whether to output debug information
+   */
+  public PaddingKernel(KernelParameters parameters, int bitWidth, boolean dbg) {
+    super(parameters);
+
+    DFEVar nInputs = io.scalarInput(SCALAR_NUM_INP, dfeUInt(32));
+    DFEVar totalCycles = io.scalarInput(SCALAR_TOTAL_CYCLES, dfeUInt(32));
+
+    CounterChain chain = control.count.makeCounterChain();
+    DFEVar cycles = chain.addCounter(totalCycles, 1);
+    DFEVar paddingCycles = cycles >= nInputs;
+
+    DFEVar input = io.input(INP_NAME, dfeRawBits(bitWidth), ~paddingCycles);
+    DFEVar out = paddingCycles ? 0 : input;
+    io.output(OUT_NAME, out, dfeRawBits(bitWidth));
+
+    if (dbg) {
+      debug.simPrintf("PADDING: %d (%d) nInputs %d\n", cycles, totalCycles, nInputs);
+    }
+  }
+}
diff --git a/src/com/custom_computing_ic/dfe_snippets/kernels/README.md b/src/com/custom_computing_ic/dfe_snippets/kernels/README.md
new file mode 100644
index 0000000..1f2833d
--- /dev/null
+++ b/src/com/custom_computing_ic/dfe_snippets/kernels/README.md
@@ -0,0 +1,10 @@
+# Kernels
+
+This directory contains kernels that can be used for general MaxJ design development.
+
+## Resource Usage
+
+| Kernel Name           | LUTs | FFs | BRAM | DSP |
+|-----------------------|------|-----|------|-----|
+| `PaddingKernel(32)`   | 428  | 929 | 0    | 0   |
+| `UnpaddingKernel(32)` | 416  | 933 | 0    | 0   |
diff --git a/src/com/custom_computing_ic/dfe_snippets/kernels/UnpaddingKernel.java b/src/com/custom_computing_ic/dfe_snippets/kernels/UnpaddingKernel.java
new file mode 100644
index 0000000..d01ac0f
--- /dev/null
+++ b/src/com/custom_computing_ic/dfe_snippets/kernels/UnpaddingKernel.java
@@ -0,0 +1,48 @@
+package com.custom_computing_ic.dfe_snippets.kernels;
+
+import com.maxeler.maxcompiler.v2.kernelcompiler.Kernel;
+import com.maxeler.maxcompiler.v2.kernelcompiler.KernelParameters;
+import com.maxeler.maxcompiler.v2.kernelcompiler.types.base.DFEVar;
+import com.maxeler.maxcompiler.v2.kernelcompiler.stdlib.core.CounterChain;
+
+/**
+ * Un-pads the inputs and will not output zeros after
+ * nInputs have been processed.
+ *
+ * Scalars:
+ * NUM_INP - number of input data that are valid for output 
+ * TOTAL_CYCLES - number of cycles that this kernel should
+ * receive data from the previous kernel
+ *
+ * @since 18/05/2017
+ */
+public class UnpaddingKernel extends Kernel {
+  public static final String INP_NAME            = "UNPADDING_INP";
+  public static final String OUT_NAME            = "UNPADDING_OUT";
+  public static final String SCALAR_NUM_INP      = "NUM_INP";
+  public static final String SCALAR_TOTAL_CYCLES = "TOTAL_CYCLES";
+
+  /**
+   * Constructor of the Unpadding Kernel.
+   *
+   * @param parameters parameters for this kernel passed from a manager
+   * @param bitWidth number of bits to be processed in each cycle
+   * @param dbg flag to decide whether to output debug information
+   */
+  public UnpaddingKernel(KernelParameters parameters, int bitWidth, boolean dbg) {
+    super(parameters);
+
+    DFEVar nInputs = io.scalarInput(SCALAR_NUM_INP, dfeUInt(32));
+    DFEVar totalCycles = io.scalarInput(SCALAR_TOTAL_CYCLES, dfeUInt(32));
+    CounterChain chain = control.count.makeCounterChain();
+    DFEVar cycles = chain.addCounter(totalCycles, 1);
+    DFEVar unpaddingCycles = cycles >= nInputs;
+
+    DFEVar input = io.input(INP_NAME, dfeRawBits(bitWidth));
+    io.output(OUT_NAME, input, dfeRawBits(bitWidth), ~unpaddingCycles);
+
+    if (dbg) {
+      debug.simPrintf("UNPADDING: %d (%d) nInputs %d\n", cycles, totalCycles, nInputs);
+    }
+  }
+}
diff --git a/src/com/custom_computing_ic/dfe_snippets/manager/ManagerUtils.java b/src/com/custom_computing_ic/dfe_snippets/manager/ManagerUtils.java
index 98d5b7d..c304916 100644
--- a/src/com/custom_computing_ic/dfe_snippets/manager/ManagerUtils.java
+++ b/src/com/custom_computing_ic/dfe_snippets/manager/ManagerUtils.java
@@ -178,6 +178,10 @@ public static void addLinearStreamFromLMemToKernel(LMemInterface iface, KernelBl
         LMemCommandGroup.MemoryAccessPattern.LINEAR_1D);
   }
 
+  public static void addLinearStreamFromLMemToKernel(LMemCommandGroup group, KernelBlock kernel, String src, String dst) {
+    kernel.getInput(dst) <== group.addStreamFromLMem(src);
+  }
+
   @Deprecated
   public static void addLinearStreamFromKernelToLmem(CustomManager m, KernelBlock kernel, String name) {
     m.addStreamToOnCardMemory(
@@ -201,6 +205,10 @@ public static void addLinearStreamFromKernelToLMem(LMemInterface iface, KernelBl
         LMemCommandGroup.MemoryAccessPattern.LINEAR_1D) <== kernel.getOutput(src);
   }
 
+  public static void addLinearStreamFromKernelToLMem(LMemCommandGroup group, KernelBlock kernel, String src, String dst) {
+    group.addStreamToLMem(dst) <== kernel.getOutput(src);
+  }
+
   public static void ignoreLMemStreams(EngineInterface ei) {
     ei.ignoreLMem("cpu2lmem");
     ei.ignoreStream("fromcpu");