diff --git a/packages/core/test/render3/perf/README.md b/packages/core/test/render3/perf/README.md
index ab79a0bb53..eb3dac46af 100644
--- a/packages/core/test/render3/perf/README.md
+++ b/packages/core/test/render3/perf/README.md
@@ -1,19 +1,19 @@
 ### Build
 
 ```
-yarn bazel build //packages/core/test/render3/perf:{name}.min_debug.es2015.js --define=compile=aot
+yarn bazel build //packages/core/test/render3/perf:${BENCHMARK}.min_debug.es2015.js --define=compile=aot
 ```
 
 ### Run 
 
 ```
-node dist/bin/packages/core/test/render3/perf/{name}.min_debug.es2015.js
+node dist/bin/packages/core/test/render3/perf/${BENCHMARK}.min_debug.es2015.js
 ```
 
 ### Profile
 
 ```
-node --no-turbo-inlining --inspect-brk dist/bin/packages/core/test/render3/perf/{name}.min_debug.es2015.js
+node --no-turbo-inlining --inspect-brk dist/bin/packages/core/test/render3/perf/${BENCHMARK}.min_debug.es2015.js
 ```
 
 then connect with a debugger (the `--inspect-brk` option will make sure that benchmark execution doesn't start until a debugger is connected and the code execution is manually resumed). 
@@ -64,7 +64,8 @@ The resulting output should look something like this:
 
 ### Notes
 
-In all the above commands {name} should be replaced with the actual benchmark (folder) name, ex.:
+In all the above commands `${BENCHMARK}` should be replaced with the actual benchmark (folder) name, ex.:
 - build: `yarn bazel build //packages/core/test/render3/perf:noop_change_detection.min_debug.es2015.js --define=compile=aot`
 - run: `time node dist/bin/packages/core/test/render3/perf/noop_change_detection.min_debug.es2015.js`
 - profile: `node --no-turbo-inlining --inspect-brk dist/bin/packages/core/test/render3/perf/noop_change_detection.min_debug.es2015.js profile`
+- experimenting `BENCHMARK=noop_change_detection; yarn bazel build //packages/core/test/render3/perf:${BENCHMARK}.min_debug.es2015.js --define=compile=aot && node dist/bin/packages/core/test/render3/perf/${BENCHMARK}.min_debug.es2015.js`
diff --git a/packages/core/test/render3/perf/directive_instantiate/index.ts b/packages/core/test/render3/perf/directive_instantiate/index.ts
index 25d65d2a64..ab60d58241 100644
--- a/packages/core/test/render3/perf/directive_instantiate/index.ts
+++ b/packages/core/test/render3/perf/directive_instantiate/index.ts
@@ -90,9 +90,7 @@ const createTime = directiveInstantiate('create');
 
 console.profile('directive_instantiate');
 while (createTime()) {
-  for (let i = 0; i < 50000; i++) {
-    createAndRenderLView(null, embeddedTView, viewTNode);
-  }
+  createAndRenderLView(null, embeddedTView, viewTNode);
 }
 console.profileEnd();
 
diff --git a/packages/core/test/render3/perf/element_text_create/index.ts b/packages/core/test/render3/perf/element_text_create/index.ts
index d9041a0cb3..43c2e870a5 100644
--- a/packages/core/test/render3/perf/element_text_create/index.ts
+++ b/packages/core/test/render3/perf/element_text_create/index.ts
@@ -81,9 +81,7 @@ const createTime = elementTextCreate('create');
 
 console.profile('element_text_create');
 while (createTime()) {
-  for (let i = 0; i < 100000; i++) {
-    createAndRenderLView(null, embeddedTView, viewTNode);
-  }
+  createAndRenderLView(null, embeddedTView, viewTNode);
 }
 console.profileEnd();
 
diff --git a/packages/core/test/render3/perf/micro_bench.ts b/packages/core/test/render3/perf/micro_bench.ts
index d727935eab..f3c7940079 100644
--- a/packages/core/test/render3/perf/micro_bench.ts
+++ b/packages/core/test/render3/perf/micro_bench.ts
@@ -7,8 +7,11 @@
  */
 const performance = require('perf_hooks').performance;
 
-const MIN_SAMPLE_COUNT_NO_IMPROVEMENT = 30;
-const MIN_SAMPLE_DURATION = 100;
+// Higher number here makes it more likely that we are more sure of the result.
+const MIN_SAMPLE_COUNT_NO_IMPROVEMENT = 100;
+// A smaller number here means that we are coming too close on timer resultion, but it also means
+// that it is less likely that we will be bothered by GC or preemptive multi tasking.
+const MIN_SAMPLE_DURATION = 10;
 
 const UNITS = ['ms', 'us', 'ns', 'ps'];
 export interface Benchmark {