AutoJMH is a tool that takes as input one segment of code (belonging to an existing program) and one training execution (could be a unit test run for example) and builds full JMH microbenchmark for the segment. The tool generates the JMH class and then places the segment code inside the wrapper method, fully generating the microbenchmark. AutoJMH also initializes the variables of the microbenchmark and perform a series of analysis to automatically avoid common microbenchmarking pitfalls.
For example, suppose that given the following program we want to extract the int a = (int)Math.log(n * n / 2) * myBubbleSort(list) into a microbenchmark:
private static int gnomesort(List<Integer> ar) {
int i = 0; int k = 0;
while (i < ar.size()) {
if (i == 0 || ar.get(i-1) <= ar.get(i)) i++;
else { int tmp = ar.get(i); ar.set(i, ar.get(i-1)); ar.set(--i, tmp); }
k++;
}
return k;
}
//Sorts the list, add 'n' and returns the number
//of steps needed to sort the list and do some math with it
public int addAndSorted(List<Integer> list, int n) {
n = n < 1 ? 1 : n;
list.add(n);
/** @bench-this */
int a = (int)Math.log(n * n / 2) * myBubbleSort(list);
return a;
}
AutoJMH will generate the following microbenchmark for it:
@State(Scope.Thread)
public class org_packagename_YourClassName_348 {
int n;
List<Integer> list;
List<Integer> list___reset;
private static int gnomesort(List<Integer> ar) {
int i = 0; int k = 0;
while (i < ar.size()) {
if (i == 0 || ar.get(i-1) <= ar.get(i)) i++;
else { int tmp = ar.get(i); ar.set(i, ar.get(i-1)); ar.set(--i, tmp); }
k++;
}
return k;
}
@Setup
public void setup() {
Loader loader_348 = new Loader("usr/data/YourClassName_348");
list___reset = loader_348.readIntegerList();
list = loader_348.readEmptyList();
n = loader_348.readInt();
}
@Setup(Level.Invocation)
public void resetCode() {
list.clear(); list.addAll(list___reset);
}
@Benchmark
public int doBenchmark() {
int a = (int)Math.log(n * n / 2) * gnomesort(list);
return a;
}
}
The tool is able to extract the code and all its dependencies (method, variables), provide initialization values, avoid DCE and keep the microbenchmark in its one stable state.