diff --git a/doc/Plugins.md b/doc/Plugins.md
index 4a9c63c..547b75c 100644
--- a/doc/Plugins.md
+++ b/doc/Plugins.md
@@ -167,3 +167,7 @@ optimization.
 
 The `V8GetOptimizationStatus` plugin collects the V8 engine's optimization
 status for a given function after it has been benchmarked.
+
+### Class: `RecordMemorySpikePlugin`
+
+A plugin to record memory allocation spikes in your benchmark's run. It should help you understand the speed vs memory tradeoffs you're making.
diff --git a/lib/index.js b/lib/index.js
index 4ae5843..a0a90ad 100644
--- a/lib/index.js
+++ b/lib/index.js
@@ -15,6 +15,7 @@ const {
   V8NeverOptimizePlugin,
   V8GetOptimizationStatus,
   V8OptimizeOnNextCallPlugin,
+  RecordMemorySpikePlugin,
 } = require('./plugins');
 const {
   validateFunction,
@@ -205,6 +206,7 @@ module.exports = {
   V8NeverOptimizePlugin,
   V8GetOptimizationStatus,
   V8OptimizeOnNextCallPlugin,
+  RecordMemorySpikePlugin,
   chartReport,
   textReport,
   htmlReport,
diff --git a/lib/plugins.js b/lib/plugins.js
index b18a631..2990d22 100644
--- a/lib/plugins.js
+++ b/lib/plugins.js
@@ -12,6 +12,10 @@ const {
   MemoryPlugin,
 } = require('./plugins/memory');
 
+const {
+  RecordMemorySpikePlugin
+} = require('./plugins/memory-spike');
+
 const {
   validateFunction,
   validateArray,
@@ -53,5 +57,6 @@ module.exports = {
   V8NeverOptimizePlugin,
   V8GetOptimizationStatus,
   V8OptimizeOnNextCallPlugin,
+  RecordMemorySpikePlugin,
   validatePlugins,
 };
diff --git a/lib/plugins/memory-spike.js b/lib/plugins/memory-spike.js
new file mode 100644
index 0000000..9f6dde6
--- /dev/null
+++ b/lib/plugins/memory-spike.js
@@ -0,0 +1,109 @@
+const v8 = require("node:v8");
+
+const translateHeapStats = (stats = []) => {
+  const result = {};
+  for (const { space_name, space_used_size } of stats) {
+    result[space_name] = space_used_size;
+  }
+  return result;
+};
+
+const updateMaxEachKey = (current, update) => {
+  for (const key in current) {
+    current[key] = Math.max(current[key], update[key]);
+  }
+};
+
+const diffEachKey = (a, b, divBy = 1) => {
+  const result = {};
+  for (const key in a) {
+    result[key] = (b[key] - a[key]) / divBy;
+  }
+  return result;
+};
+
+const avgEachKey = (items) => {
+  const result = {};
+  for (const item of items) {
+    for (const key in item) {
+      result[key] = (result[key] || 0) + item[key];
+    }
+  }
+  for (const key in result) {
+    result[key] /= items.length;
+  }
+
+  return result;
+};
+
+const toHumanReadable = (obj) => {
+  const result = {};
+  for (const key in obj) {
+    if (obj[key] > 0) result[key] = `+${(obj[key] / 1024).toFixed(4)} KB`;
+  }
+  return result;
+};
+
+globalThis.__recordMemorySpike = (frequency = 2) => {
+  const initial = translateHeapStats(v8.getHeapSpaceStatistics());
+  const result = { ...initial };
+  const collect = () =>
+    updateMaxEachKey(result, translateHeapStats(v8.getHeapSpaceStatistics()));
+  const interval = setInterval(collect, frequency);
+  return {
+    collect,
+    getResult: () => {
+      clearInterval(interval);
+      collect();
+      return [initial, result];
+    },
+  };
+};
+
+class RecordMemorySpikePlugin {
+  #spikeSamples = {};
+  isSupported() {
+    try {
+      new Function(`gc()`)();
+      return true;
+    } catch (e) {
+      return false;
+    }
+  }
+
+  beforeClockTemplate() {
+    return [`const __mem_spike__ = __recordMemorySpike();`];
+  }
+  afterClockTemplate({ context, bench }) {
+    return [
+      `;
+      ${context}.benchName=${bench}.name;
+      ${context}.memSpike = __mem_spike__.getResult();
+      `,
+    ];
+  }
+
+  onCompleteBenchmark([_time, iterations, results]) {
+    gc();
+    const [initial, result] = results.memSpike;
+    const diff = diffEachKey(initial, result, iterations);
+    if (!this.#spikeSamples[results.benchName]) {
+      this.#spikeSamples[results.benchName] = [];
+    }
+    this.#spikeSamples[results.benchName].push(diff);
+  }
+
+  getResult(name) {
+    return toHumanReadable(avgEachKey(this.#spikeSamples[name]));
+  }
+
+  getReport() { 
+    process._rawDebug('grp',arguments);
+
+  }
+
+  toString() {
+    return "RecordMemorySpikePlugin";
+  }
+}
+exports.RecordMemorySpikePlugin = RecordMemorySpikePlugin;
diff --git a/package.json b/package.json
index fad2318..a1ed1ee 100644
--- a/package.json
+++ b/package.json
@@ -4,7 +4,7 @@
   "description": "",
   "main": "lib/index.js",
   "scripts": {
-    "test": "node --test --allow-natives-syntax"
+    "test": "node --test --allow-natives-syntax --expose-gc"
   },
   "repository": {
     "type": "git",
diff --git a/test/plugin-memspike.js b/test/plugin-memspike.js
new file mode 100644
index 0000000..5f40945
--- /dev/null
+++ b/test/plugin-memspike.js
@@ -0,0 +1,48 @@
+// @ts-check
+const { RecordMemorySpikePlugin, Suite
+} = require('../lib/index');
+
+const { test } = require("node:test");
+const assert = require("node:assert");
+const { setTimeout } = require("node:timers/promises");
+
+const wasteMemoryForAWhile = async () => {
+  const a = Buffer.alloc(1024 * 1024, "a");
+  await setTimeout(5);
+  a.at(1); // prevent optimization
+};
+function noop() {}
+
+test("RecordMemorySpikePlugin", async (t) => {
+  const bench = new Suite({
+    reporter: noop,
+    plugins: [new RecordMemorySpikePlugin()],
+  });
+  bench
+    .add("sequence", async () => {
+      for (let i = 0; i < 20; i++) {
+        await wasteMemoryForAWhile();
+      }
+    })
+    .add("concurent", async () => {
+      await Promise.all(
+        Array.from({ length: 20 }, () => wasteMemoryForAWhile()),
+      );
+    });
+
+  const [bench1, bench2] = await bench.run();
+  console.dir(
+    {
+      bench1,
+      bench2,
+    },
+    { depth: 100 },
+  );
+
+  const { plugins: [{ result: result1 }] } = bench1;
+  const { plugins: [{ result: result2 }] } = bench2;
+  const parseResult = (str) => parseFloat(str.replace(/[^\d.-]/g, ''));
+  assert.ok(parseResult(result1.new_space) > parseResult(result2.new_space), "Sequence new_space should be larger than concurrent new_space");
+  assert.ok(parseResult(result1.old_space) > parseResult(result2.old_space), "Sequence old_space should be larger than concurrent old_space");
+  
+});