Skip to main content
Version: Devnet (v3.0.0-devnet.2)

Profiling and Optimizing Contracts

This guide shows you how to profile your Aztec transactions to identify bottlenecks and optimize gas usage.

Prerequisites​

  • aztec-nargo installed (see installation)
  • aztec-wallet installed (part of Sandbox)
  • Aztec contract deployed and ready to test
  • Basic understanding of proving and gate counts

Profile with aztec-wallet​

Step 1: Import test accounts​

aztec-wallet import-test-accounts

Step 2: Deploy your contract​

aztec-wallet deploy MyContractArtifact \
  --from accounts:test0 \
  --args <constructor_args> \
  -a mycontract

Step 3: Set up initial state​

aztec-wallet send setup_state \
  -ca mycontract \
  --args <setup_args> \
  -f test0

Step 4: Profile a transaction​

Instead of send, use profile with the same parameters:

aztec-wallet profile private_function \
  -ca mycontract \
  --args <function_args> \
  -f accounts:test0

Step 5: Analyze the output​

Gate count per circuit:
   SchnorrAccount:entrypoint                          Gates: 21,724     Acc: 21,724
   private_kernel_init                                Gates: 45,351     Acc: 67,075
   MyContract:private_function                        Gates: 31,559     Acc: 98,634
   private_kernel_inner                               Gates: 78,452     Acc: 177,086
   private_kernel_reset                               Gates: 91,444     Acc: 268,530
   private_kernel_tail                                Gates: 31,201     Acc: 299,731

Total gates: 299,731

The output shows:

  • Gate count per circuit component
  • Accumulated gate count
  • Total gates for the entire transaction

Profile with aztec.js​

Profile Modes
  • gates: Shows gate counts per circuit
  • execution-steps: Detailed execution trace
  • full: Complete profiling information

Step 1: Profile a transaction​

const result = await contract.methods
  .my_function(args)
  .profile({ 
    from: address,
    profileMode: 'gates',
    skipProofGeneration: false 
  });

console.log('Gate count:', result.gateCount);

Step 2: Profile deployment​

const deploy = await Contract.deploy(args).profile({ from: address, profileMode: 'full' });
Experimental

Flamegraph generation is experimental and may not be available in all versions.

Generate flamegraphs (if available)​

Generate and view​

# Compile first
aztec-nargo compile

# Generate flamegraph
aztec flamegraph target/contract.json function_name

# Serve locally
SERVE=1 aztec flamegraph target/contract.json function_name
Reading Flamegraphs
  • Width = Time in operation
  • Height = Call depth
  • Wide sections = Optimization targets

Common optimizations​

Key Metrics
  • Gate count: Circuit complexity
  • Kernel overhead: Per-function cost
  • Storage access: Read/write operations
Optimization Pattern

Batch operations to reduce kernel circuit overhead.

// ❌ Multiple kernel invocations
for i in 0..3 {
    transfer_single(amounts[i], recipients[i]);
}

// ✅ Single kernel invocation
for i in 0..3 {
    let note = Note::new(amounts[i], recipients[i]);
    storage.notes.at(recipients[i]).insert(note);
}
Storage Optimization

Group storage reads to reduce overhead.

// Read once, use multiple times
let values = [storage.v1.get(), storage.v2.get(), storage.v3.get()];
for v in values {
    assert(v > 0);
}

Minimize note operations​

Note Aggregation

Combine multiple small notes into fewer larger ones to reduce proving overhead.

// ❌ Many small notes = high overhead
for value in values {
    storage.notes.insert(Note::new(value, owner));
}

// ✅ Single aggregated note = lower overhead
let total = values.reduce(|a, b| a + b);
storage.notes.insert(Note::new(total, owner));

Profile different scenarios​

Profile with different inputs​

# Small values
aztec-wallet profile function -ca mycontract --args 10 -f test0

# Large values
aztec-wallet profile function -ca mycontract --args 1000000 -f test0

Profile execution modes​

// Profile gates only
await contract.methods.function().profile({ profileMode: 'gates' });

// Profile execution steps
await contract.methods.function().profile({ profileMode: 'execution-steps' });

// Full profile
await contract.methods.function().profile({ profileMode: 'full' });

Skip proof generation for faster iteration​

await contract.methods.function().profile({
  profileMode: 'gates',
  skipProofGeneration: true  // Faster but less accurate
});

Interpret profiling results​

Gate count guidelines​

  • < 50,000 gates: Excellent performance
  • 50,000 - 200,000 gates: Acceptable for most use cases
  • 200,000 - 500,000 gates: May cause delays, consider optimizing
  • > 500,000 gates: Requires optimization for production

Common optimization targets​

  1. private_kernel_inner - Reduce nested function calls
  2. private_kernel_reset - Minimize note nullifications
  3. Contract functions - Optimize computation logic
  4. private_kernel_tail - Reduce public function calls

Best practices​

Development workflow​

  1. Profile early - Establish baseline metrics
  2. Profile often - Check impact of changes
  3. Profile realistically - Use production-like data
  4. Document findings - Track optimization progress

Optimization priorities​

  1. User-facing functions - Optimize most-used features first
  2. Critical paths - Focus on transaction bottlenecks
  3. Batch operations - Combine related operations
  4. Cache calculations - Store reusable results

Next steps​