Tools

Data Hazards and Instruction Scheduling

📋 Simulation Instructions

This interactive pipeline simulator helps you visualize data hazards and understand how instruction scheduling and data forwarding affect pipeline performance.

Step 1: Build Your Instruction Sequence

  1. Select Instruction Type from the dropdown menu:

    • Compute Instructions: ADD, SUB, MUL, DIV
    • Memory Instructions: LOAD, STORE

  2. Choose Registers:

    • Destination (rd): Register where the result will be written (not applicable for STORE)
    • Source 1 (rs1): First source register
    • Source 2 (rs2): Second source register (not applicable for LOAD)
    • Offset: Memory offset for LOAD/STORE instructions

  3. Click "Add Instruction" to add it to your sequence

  4. Build a sequence of at least 2-3 instructions to observe data hazards

Example Sequence to Try:

1. ADD R1, R2, R3    (R1 = R2 + R3)
          2. SUB R4, R1, R5    (R4 = R1 - R5)  ← RAW hazard on R1
          3. MUL R6, R4, R7    (R6 = R4 * R7)  ← RAW hazard on R4

Step 2: Configure Instruction Latencies (Optional)

Each instruction type has configurable execution latencies:

  • Compute Instructions (ADD, SUB, MUL, DIV): Configure Execute stage cycles
  • Memory Instructions (LOAD, STORE): Configure Memory stage cycles

Default Latencies:

  • ADD/SUB: 1 cycle (Execute)
  • MUL: 3 cycles (Execute)
  • DIV: 4 cycles (Execute)
  • LOAD/STORE: 2 cycles (Memory)

To Modify:

  1. Locate the instruction type in the "Instruction Latencies" section
  2. Change the cycle count in the input field
  3. The simulation updates automatically

To Reset:

  • Click "Reset to Defaults" to restore original latency values

Step 3: Toggle Data Forwarding

Data forwarding (bypassing) allows results to be passed directly between pipeline stages without waiting for write-back.

  • Forwarding OFF (default): Instructions must wait for results to be written to the register file
  • Forwarding ON: Results can be forwarded from Execute/Memory stages, reducing stalls

To Toggle:

  1. Use the "Enable Data Forwarding" checkbox
  2. Observe how the pipeline visualization changes
  3. Compare performance metrics with and without forwarding

Step 4: Analyze the Pipeline Visualization

Once you add instructions, three visualization panels appear:

A. Pipeline Stage Diagram
  • Shows each instruction progressing through pipeline stages (F → D → E → M → W)
  • Colored cells: Active execution in that stage
  • Gray cells (stall): Pipeline stall due to data hazard
  • Hazard indicators: Red markers show where hazards occur
B. Stage Entry Timeline
  • Displays the cycle number when each instruction enters each stage
  • Helps identify delays and stall patterns
  • Shows the impact of hazards on instruction flow
C. Performance Metrics
  • Total Cycles: Total execution time for the instruction sequence
  • Stall Cycles: Number of cycles wasted due to hazards
  • CPI (Cycles Per Instruction): Average cycles per instruction
  • Throughput: Instructions completed per cycle
  • Comparison: Shows improvement when forwarding is enabled

Step 5: Reorder Instructions (Instruction Scheduling)

You can manually reorder instructions to reduce hazards:

  1. Drag and drop instructions in the "Instruction Sequence" panel
  2. Move independent instructions between dependent ones to fill stall slots
  3. Observe how reordering affects total cycles and stalls

Example Optimization:

Original (with stalls):
          1. ADD R1, R2, R3
          2. SUB R4, R1, R5    ← Stalls waiting for R1
          
          Optimized (insert independent instruction):
          1. ADD R1, R2, R3
          2. MUL R6, R7, R8    ← Independent, no stall
          3. SUB R4, R1, R5    ← R1 now ready

Step 6: Remove Instructions

  • Click the × button next to any instruction to remove it
  • Click "Clear All Instructions" to start over

🎯 Learning Objectives

By completing this simulation, you will:

  1. ✅ Identify RAW (Read After Write) data hazards in instruction sequences
  2. ✅ Understand how pipeline stalls impact performance
  3. ✅ Learn how data forwarding reduces stall cycles
  4. ✅ Practice instruction scheduling to minimize hazards
  5. ✅ Analyze performance metrics (CPI, throughput, stall cycles)
  6. ✅ Observe the effect of different instruction latencies on pipeline behavior

💡 Tips for Experimentation

  • Start with simple 2-3 instruction sequences to understand basic hazards
  • Try sequences with multiple dependencies to see cascading stalls
  • Compare the same sequence with forwarding ON vs OFF
  • Experiment with different latencies (e.g., increase MUL to 5 cycles)
  • Use drag-and-drop to find optimal instruction ordering
  • Pay attention to which registers are being read/written to identify dependencies