diff --git a/Dockerfile b/Dockerfile
index 2103e1c..29e5f24 100644
--- a/Dockerfile
+++ b/Dockerfile
@@ -23,5 +23,5 @@ WORKDIR /app
 COPY . /app/
 
 # Compile ONLY task4 during the build to avoid unrelated compile failures
-RUN mpic++ -O2 -std=c++17 -o task4 task4.cpp
+RUN mpic++ -O2 -std=c++17 -o task5 task5.cpp
 RUN chmod +x /app/task4 || true
\ No newline at end of file
diff --git a/task5.cpp b/task5.cpp
new file mode 100644
index 0000000..5c32f61
--- /dev/null
+++ b/task5.cpp
@@ -0,0 +1,62 @@
+#include <mpi.h>
+#include <cstdio>
+#include <vector>
+
+int main(int argc, char** argv) {
+  MPI_Init(&argc, &argv);
+
+  int worldSize = 0;
+  int worldRank = 0;
+  MPI_Comm_size(MPI_COMM_WORLD, &worldSize);
+  MPI_Comm_rank(MPI_COMM_WORLD, &worldRank);
+
+  // Enforce exactly 5 processes (hard-coded layout below assumes 5 ranks)
+  if (worldSize != 5) {
+    if (worldRank == 0) {
+      std::printf("This program must be run with exactly 5 MPI processes.\n");
+    }
+    MPI_Abort(MPI_COMM_WORLD, 1);
+  }
+
+  // Hard-coded total elements and source data on root
+  const int totalElements = 17;
+  std::vector<int> sendBuffer;
+  if (worldRank == 0) {
+    sendBuffer.resize(totalElements);
+    for (int i = 0; i < totalElements; ++i) {
+      sendBuffer[i] = i;  // 0..16
+    }
+  }
+
+  // Hard-coded sendCounts and displacements for 5 ranks and 17 elements
+  // Rank 0 -> 4 elems [0..3]
+  // Rank 1 -> 4 elems [4..7]
+  // Rank 2 -> 3 elems [8..10]
+  // Rank 3 -> 3 elems [11..13]
+  // Rank 4 -> 3 elems [14..16]
+  const int sendCounts[5]     = {4, 4, 3, 3, 3};
+  const int displacements[5]  = {0, 4, 8, 11, 14};
+
+  const int myReceiveCount = sendCounts[worldRank];
+  std::vector<int> receiveBuffer(myReceiveCount, -1);
+
+  MPI_Scatterv(
+    worldRank == 0 ? sendBuffer.data() : nullptr, // sendbuf (root only)
+    sendCounts,                                    // sendcounts (root only)
+    displacements,                                 // displs (root only)
+    MPI_INT,                                       // send type
+    receiveBuffer.data(),                          // recvbuf (all ranks)
+    myReceiveCount,                                // recvcount for this rank
+    MPI_INT,                                       // recv type
+    0,                                             // root
+    MPI_COMM_WORLD                                 // communicator
+  );
+
+  for (int i = 0; i < myReceiveCount; ++i) {
+    std::printf("Rank %d received element %d: %d\n", worldRank, i, receiveBuffer[i]);
+  }
+
+  std::printf("Bye from %d\n", worldRank);
+  MPI_Finalize();
+  return 0;
+}
\ No newline at end of file