WASM Acceleration

CallData operations can be accelerated using WebAssembly for performance-critical applications. The Zig implementation compiles to WASM, providing near-native speed in browser and Node.js environments.

Build Modes

Voltaire provides two WASM build modes optimized for different use cases:

ReleaseSmall (Production)

Size-optimized build for production bundles: Characteristics:

Minimal bundle size (~50-100KB compressed)
Aggressive dead code elimination
Optimized for download speed
Suitable for web applications

ReleaseFast (Performance)

Performance-optimized build for compute-intensive workloads: Characteristics:

Maximum execution speed
Larger binary size (~150-300KB compressed)
Loop unrolling and inlining
Suitable for backend services, workers

Performance Comparison

Encoding
Decoding
Selector Matching

Encoding function call with parameters:

import { CallData, Abi, Address, TokenBalance } from '@tevm/voltaire';

const abi = Abi([{
  name: "transfer",
  type: "function",
  inputs: [
    { name: "to", type: "address" },
    { name: "amount", type: "uint256" }
  ]
}]);

const recipient = Address("0x70997970C51812dc3A010C7d01b50e0d17dc79C8");
const amount = TokenBalance.fromUnits("1000", 18);

// Benchmark
console.time("encode");
const calldata = abi.transfer.encode(recipient, amount);
console.timeEnd("encode");

Results (1M iterations):

Pure JS: ~850ms
WASM (ReleaseSmall): ~320ms (2.6x faster)
WASM (ReleaseFast): ~180ms (4.7x faster)

Decoding calldata to structured form:

const calldata = CallData("0xa9059cbb...");

console.time("decode");
const decoded = CallData.decode(calldata, abi);
console.timeEnd("decode");

Results (1M iterations):

Pure JS: ~920ms
WASM (ReleaseSmall): ~380ms (2.4x faster)
WASM (ReleaseFast): ~210ms (4.4x faster)

Extracting and matching function selectors:

console.time("selector");
const selector = CallData.getSelector(calldata);
const isTransfer = CallData.hasSelector(calldata, "0xa9059cbb");
console.timeEnd("selector");

Results (10M iterations):

Pure JS: ~450ms
WASM (ReleaseSmall): ~120ms (3.8x faster)
WASM (ReleaseFast): ~65ms (6.9x faster)

Usage

Automatic Selection

Voltaire automatically uses WASM when available:

import { CallData } from '@tevm/voltaire';

// Automatically uses WASM if loaded
const calldata = CallData("0xa9059cbb...");
const selector = CallData.getSelector(calldata);

No code changes needed - WASM acceleration is transparent.

Manual Loading

For advanced control, manually load WASM module:

Node.js
Browser
Worker

import { loadWasm } from '@tevm/voltaire/wasm-loader';
import { CallData } from '@tevm/voltaire';

// Load WASM module
await loadWasm('primitives');

// Now using WASM acceleration
const calldata = CallData.encode("transfer(address,uint256)", [addr, amount]);

import { loadWasm } from '@tevm/voltaire/wasm-loader';

// Load from CDN or local path
await loadWasm('primitives', {
  url: '/wasm/primitives.wasm'
});

// WASM now active for all CallData operations

// worker.ts
import { loadWasm } from '@tevm/voltaire/wasm-loader';
import { CallData } from '@tevm/voltaire';

// Load WASM in worker context
await loadWasm('primitives-fast'); // Use fast build

// Handle messages
self.onmessage = async (e) => {
  const { signature, params } = e.data;
  const calldata = CallData.encode(signature, params);
  self.postMessage({ calldata: CallData.toHex(calldata) });
};

Checking WASM Status

Verify if WASM is loaded:

import { isWasmLoaded } from '@tevm/voltaire/wasm-loader';

if (isWasmLoaded('primitives')) {
  console.log('Using WASM acceleration');
} else {
  console.log('Falling back to pure JS');
}

Memory Management

WASM module manages its own memory efficiently:

Linear Memory

WASM uses linear memory for all operations:

import { getWasmMemory } from '@tevm/voltaire/wasm-loader';

// Access WASM memory (advanced use only)
const memory = getWasmMemory('primitives');
console.log('Memory pages:', memory.buffer.byteLength / 65536);

Memory grows automatically when needed:

Initial: 16 pages (1MB)
Maximum: 256 pages (16MB)
Growth: Automatic on demand

Allocation Strategy

Zig’s allocator optimizes for CallData operations:

// Internal implementation
pub fn encodeCallData(
    allocator: std.mem.Allocator,
    selector: [4]u8,
    params: []const AbiValue,
) !CallData {
    // Use arena allocator for batch operations
    var arena = std.heap.ArenaAllocator.init(allocator);
    defer arena.deinit();

    // All allocations freed at once
    const data = try encodeParams(arena.allocator(), params);
    return CallData{ .data = data };
}

Benefits:

Minimal fragmentation
Batch deallocation
Zero-cost cleanup

Memory Limits

Set memory limits for safety:

import { loadWasm } from '@tevm/voltaire/wasm-loader';

await loadWasm('primitives', {
  memory: {
    initial: 32,  // 2MB
    maximum: 512, // 32MB
  }
});

Bundle Optimization

Tree-Shaking

Use tree-shakeable imports to minimize bundle size:

// Bundle includes only what you use
import { from, getSelector, toHex } from '@tevm/voltaire/CallData';

const calldata = from("0xa9059cbb...");
const selector = getSelector(calldata);
const hex = toHex(calldata);

// encode, decode, etc. excluded from bundle

Lazy Loading

Load WASM on demand to reduce initial bundle:

// Minimal initial bundle
import { CallData } from '@tevm/voltaire';

// Load WASM when needed
async function processCallData(data: string) {
  const { loadWasm } = await import('tevm/wasm-loader');
  await loadWasm('primitives');

  return CallData.decode(CallData(data), abi);
}

Code Splitting

Split WASM by route/feature:

// route-1.ts - Only loads when route accessed
export async function handleRoute1() {
  await import('tevm/wasm-loader').then(m => m.loadWasm('primitives'));
  // Use CallData operations
}

// route-2.ts - Independent WASM loading
export async function handleRoute2() {
  await import('tevm/wasm-loader').then(m => m.loadWasm('primitives'));
  // Use CallData operations
}

Platform Support

Compatibility

WASM module works across platforms:

Platform	Support	Notes
Chrome 57+	✅ Full	Native WASM support
Firefox 52+	✅ Full	Native WASM support
Safari 11+	✅ Full	Native WASM support
Edge 16+	✅ Full	Native WASM support
Node.js 12+	✅ Full	Built-in WASM runtime
Deno	✅ Full	Native WASM support
Bun	✅ Full	Optimized WASM JIT

Fallback

Automatic fallback to pure JS when WASM unavailable:

import { CallData } from '@tevm/voltaire';

// Works everywhere - WASM when available, JS fallback otherwise
const calldata = CallData.encode("transfer(address,uint256)", [addr, amount]);

No polyfills or configuration needed.

Benchmarking

Write benchmarks to verify performance in your environment:

import { bench, run } from 'mitata';
import { CallData, Abi } from '@tevm/voltaire';

const abi = Abi([{
  name: "transfer",
  type: "function",
  inputs: [
    { name: "to", type: "address" },
    { name: "amount", type: "uint256" }
  ]
}]);

bench('CallData.encode', () => {
  abi.transfer.encode(
    Address("0x70997970C51812dc3A010C7d01b50e0d17dc79C8"),
    TokenBalance.fromUnits("1", 18)
  );
});

bench('CallData.decode', () => {
  CallData.decode(calldata, abi);
});

await run();

Debugging WASM

Enable Debug Logging

import { setWasmDebug } from '@tevm/voltaire/wasm-loader';

setWasmDebug(true);

// Now logs WASM operations:
// [WASM] Loading primitives.wasm
// [WASM] Memory allocated: 65536 bytes
// [WASM] Function called: encode_calldata

Inspect Module

import { getWasmModule } from '@tevm/voltaire/wasm-loader';

const module = getWasmModule('primitives');
console.log('Exports:', WebAssembly.Module.exports(module));
console.log('Imports:', WebAssembly.Module.imports(module));

Production Recommendations

Use ReleaseSmall for web apps - Minimize download time
Use ReleaseFast for compute - Backend services, workers
Lazy load WASM - Don’t block initial page load
Monitor memory - Set limits for long-running processes
Test fallback - Ensure JS path works without WASM

Documentation Index

​Build Modes

​ReleaseSmall (Production)

​ReleaseFast (Performance)

​Performance Comparison

​Usage

​Automatic Selection

​Manual Loading

​Checking WASM Status

​Memory Management

​Linear Memory

​Allocation Strategy

​Memory Limits

​Bundle Optimization

​Tree-Shaking

​Lazy Loading

​Code Splitting

​Platform Support

​Compatibility

​Fallback

​Benchmarking

​Debugging WASM

​Enable Debug Logging

​Inspect Module

​Production Recommendations

​See Also