[Internship] Benchmarking remote execution

java/com/engflow/binaryinput/BUILD

@@ -0,0 +1,32 @@
+load("@rules_java//java:defs.bzl", "java_binary", "java_library")
+
+NUM_FILES = 100
+
+# Generates a number of java files based on the value of NUM_FILES
+# Each file is named HelloX.java where X is the number of the file
+# Each file contains a class with a greetNum method that prints "Hello" + the number of the file
+[genrule(
+    name = "Hello" + str(x),
+    outs = ["Hello" + str(x) + ".java"],
+    cmd_bash = "echo 'package com.engflow.binaryinput;" + "\n" +
+    "public class Hello" + str(x) +
+    " { public static void greetNum() { System.out.println(\"Hello " + str(x) + "\"); } }' > $@",
+) for x in range(1,NUM_FILES+1)]
+
+# Generates a java library that contains all the generated java files
+[java_library(
+    name = "genbinary" + str(x),
+    srcs = [":Hello" + str(x) + ".java" for x in range(1,NUM_FILES+1)],
+    visibility = ["//visibility:public"],
+) for x in range(1,NUM_FILES+1)]
+
+# Main class
+java_binary(
+    name = "main",
+    srcs = ["Main.java"],
+    main_class = "com.engflow.binaryinput.Main",
+    deps = [
+        ":genbinary" + str(x) for x in range(1,NUM_FILES+1)
+    ],
+    args = [str(NUM_FILES)],
+)

java/com/engflow/binaryinput/BenchmarkScript.py

@@ -0,0 +1,104 @@
+import subprocess
+import uuid
+import os
+import json
+
+# This script is used to benchmark the performance of Bazel builds by modifying the BUILD file to define the number of files,
+# running Bazel commands to build targets, and analyzing the Bazel profile data to extract critical time and total run time.
+
+# Usage:
+# 1. Run the script and provide the number of files, execution type (local/remote), and number of iterations when prompted.
+# 2. The script will modify the BUILD file, run Bazel commands, and analyze the profile data.
+# 3. The results will be printed, showing the highest critical time and total run time.
+
+# Modify the BUILD file to define the number of files
+def modify_build_file(num_files):
+    with open('BUILD', 'r') as file:
+        lines = file.readlines()
+
+    with open('BUILD', 'w') as file:
+        for line in lines:
+            if line.startswith('NUM_FILES'):
+                file.write(f'NUM_FILES = {num_files}\n')
+            else:
+                file.write(line)
+
+def run_bazel_command(command):
+    subprocess.run(command, check=True)
+
+# Read the Bazel profile data
+def analyze_bazel_profile(profile_path):
+    with open(profile_path, 'r') as file:
+        profile_data = json.load(file)
+    return profile_data
+
+# Extract the critical time and total run time from the Bazel profile data
+def extract_times(profile_data):
+    critical_time = 0
+    start_time = None
+    end_time = None
+
+    # Iterate through the events in the profile data to extract the critical time and total run time
+    for event in profile_data['traceEvents']:
+        if event.get('cat') == 'critical path component':
+            critical_time += event['dur'] / 1000000.0  # Convert microseconds to seconds
+        if event.get('cat') == 'build phase marker' and event['name'] == 'Launch Blaze':
+            start_time = event['ts'] / 1000000.0  # Convert microseconds to seconds
+        if event.get('cat') == 'build phase marker' and event['name'] == 'Complete build':
+            end_time = event['ts'] / 1000000.0  # Convert microseconds to seconds
+
+    # Calculate the total run time
+    total_run_time = end_time - start_time if start_time and end_time else None
+
+    return critical_time, total_run_time
+
+
+def main():
+    num_files = int(input("Enter the number of files: "))
+    execution_type = input("Enter the execution type (local/remote): ")
+    iterations = int(input("Enter the number of iterations: "))
+
+    modify_build_file(num_files)
+
+    results = []
+
+    # Path to the Bazel profile data
+    # Using an absolute path to avoid issues with the Bazel workspace
+    profile_path = os.path.abspath('profile.json')
+
+    for i in range(iterations):
+        if execution_type == 'local':
+            # Clear the Bazel cache
+            run_bazel_command(['bazel', 'clean', '--expunge'])
+            # Generate the input files
+            targets = [f':genbinary{j}' for j in range(1, num_files + 1)]
+            run_bazel_command(['bazel', 'build'] + targets)
+            # Build the main target and generate the Bazel profile data
+            run_bazel_command(['bazel', 'build', f'--profile={profile_path}', ':main'])
+        elif execution_type == 'remote':
+            # Generate a unique key for the cache silo
+            key = str(uuid.uuid4())
+            # Generate the input files
+            targets = [f':genbinary{j}' for j in range(1, num_files + 1)]
+            run_bazel_command(['bazel', 'build', '--config=engflow', f'--remote_default_exec_properties=cache-silo-key={key}'] + targets)
+            # Build the main target and generate the Bazel profile data
+            run_bazel_command(['bazel', 'build', '--config=engflow', f'--profile={profile_path}', f'--remote_default_exec_properties=cache-silo-key={key}', ':main'])
+
+        profile_output = analyze_bazel_profile(profile_path)
+        critical_time, total_run_time = extract_times(profile_output)
+        results.append((critical_time, total_run_time))
+
+    critical_times = [result[0] for result in results]
+    total_run_times = [result[1] for result in results]
+
+    print(results)
+
+    # Calculate the highest critical time and total run time
+    critical_time_max = max(critical_times)
+    total_run_time_max = max(total_run_times)
+
+    print(f'Highest Critical Time: {critical_time_max}')
+    print(f'Highest Total Run Time: {total_run_time_max}')
+
+if __name__ == '__main__':
+    main()

java/com/engflow/binaryinput/Main.java

@@ -0,0 +1,21 @@
+package com.engflow.binaryinput;
+
+import java.lang.reflect.InvocationTargetException;
+
+public class Main {
+    public static void main(String[] args) {
+        try {
+            // args[0] is the number of files to read
+            int numFiles = Integer.parseInt(args[0]);
+
+            // Load and run the greetNum method from each class
+            for(int i = 1; i <= numFiles; i++){
+                Class<?> clazz = Class.forName("com.engflow.binaryinput.Hello" + i);
+                clazz.getMethod("greetNum").invoke(null);
+            }
+
+        } catch (ClassNotFoundException | InvocationTargetException | IllegalAccessException | NoSuchMethodException e) {
+            throw new RuntimeException(e);
+        }
+    }
+}

java/com/engflow/binaryinput/README.md

@@ -0,0 +1,52 @@
+# Multiple Binary Input Example
+
+## Overview
+
+The goal of this example project is to test the performance of Engflow's remote execution and caching service based on the number of input binary files in the dependency graph. The project contains a `genrule` that generates a specified number of Java binaries for the `genbinary` Java library, which are then listed as dependencies in the main binary. The `Main.java` file loops through each generated class and calls its `greetNum` method.
+
+## Project Structure
+
+- `java/com/engflow/binaryinput/Main.java`: Main class that dynamically loads and invokes methods from generated classes.
+- `java/com/engflow/binaryinput/BUILD`: Bazel build file for the `main` java binary and the `genbinary` library.
+
+## Usage
+
+To generate the test files, build the `genbinary` library using the `genrule`:
+```sh
+bazel build //java/com/engflow/binaryinput:genbinary
+```
+
+Then, the program can be run with the following command:
+```sh
+bazel run //java/com/engflow/binaryinput:main
+```
+
+## How It Works
+
+1. **Generation of Java Binaries:**
+    - The `genrule` in the `BUILD` file generates a specified number of Java classes (`Hello1.java`, `Hello2.java`, ..., `HelloN.java`).
+    - Each generated class contains a `greetNum` method that prints a unique message.
+
+2. **Main Class Execution:**
+    - The `Main.java` file in `binaryinput` dynamically loads each generated class using reflection.
+    - It then invokes the `greetNum` method of each class, printing the corresponding message.
+
+## Configuration
+
+The number of generated files is controlled by the `NUM_FILES` variable in the `BUILD` file of the `binaryinput` package. Modify this variable to change the number of generated classes and observe the performance impact on Engflow's remote execution and caching service.
+
+## Example
+
+To generate and run the program with 10 input binary files:
+
+1. Set `NUM_FILES` to 10 in `java/com/engflow/binaryinput/BUILD`.
+2. Build the `genbinary` library:
+   ```sh
+   bazel build //java/com/engflow/binaryinput:genbinary
+   ```
+3. Run the `main` binary:
+   ```sh
+   bazel run //java/com/engflow/binaryinput:main
+   ```
+
+This will generate 10 Java classes, build the `genbinary` library, and run the `main` binary, which will print messages from each generated class.

java/com/engflow/cycleexample/README.md

@@ -0,0 +1,36 @@
+# Cyclical Dependency Example
+
+## Overview
+
+This project is designed to evaluate and benchmark the performance of EngFlow's remote execution and caching services. It specifically focuses on testing scenarios involving cyclical-like structures in the dependency graph, which are handled through interfaces and constructor injection.
+
+## Purpose
+
+The primary objectives of this project are to:
+
+1. **Generate Performance Data**: Create examples with cyclical-like dependencies to test and gather performance data for EngFlow’s remote caching and execution services.
+2. **Benchmark Analysis**: Compare the performance of local versus remote execution and caching to evaluate the efficiency and effectiveness of the service.
+3. **Support Automation Development**: Contribute to the development of automation algorithms for resource assignment by providing valuable data on how cyclical dependencies impact performance.
+
+## Project Structure
+
+The project is organized into several packages, each representing different components of the cyclical dependency example:
+
+- `class_a`: Contains `ClassA` which depends on `ClassB` through an interface.
+- `class_b`: Contains `ClassB` which implements `InterfaceB` and depends on `ClassC`.
+- `class_c`: Contains `ClassC` which implements `InterfaceA` and can be initialized with a reference to `ClassA`.
+- `interface_a`: Defines the interface `InterfaceA` implemented by `ClassA` and `ClassC`.
+- `interface_b`: Defines the interface `InterfaceB` implemented by `ClassB`.
+- `main`: Contains the `Main` class which processes the input file.
+- `input`: Contains the input text file used by the `Main` class.
+
+## How the Program Works
+
+The program takes a text input file and recursively prints each word with each class (`ClassA` prints a word, then `ClassB`, and so on) until the string is empty. The input file should be specified in the `data` attribute of the `java_binary` rule in the `BUILD` file.
+
+## How to Run Tests
+
+To run the tests and gather performance data, use the following Bazel command:
+
+```sh
+bazel test //java/com/engflow/cycleexample/class_a:class_a_test

java/com/engflow/cycleexample/class_a/BUILD

@@ -0,0 +1,22 @@
+package(default_visibility = ["//visibility:public"])
+
+java_library(
+    name = "class_a",
+    srcs = ["ClassA.java"],
+    deps = [
+        "//java/com/engflow/cycleexample/class_b:class_b",
+        "//java/com/engflow/cycleexample/interface_a:interface_a",
+    ],
+)
+
+java_test(
+    name = "class_a_test",
+    srcs = ["ClassATest.java"],
+    test_class = "com.engflow.cycleexample.class_a.ClassATest",
+    deps = [
+        ":class_a",
+        "//java/com/engflow/cycleexample/class_b:class_b",
+        "//java/com/engflow/cycleexample/class_c:class_c",
+        "//java/com/engflow/cycleexample/interface_a:interface_a",
+    ],
+)

java/com/engflow/cycleexample/class_a/ClassA.java

@@ -0,0 +1,31 @@
+package com.engflow.cycleexample.class_a;
+
+import com.engflow.cycleexample.class_b.ClassB;
+import com.engflow.cycleexample.interface_a.InterfaceA;
+
+public class ClassA implements InterfaceA {
+    private ClassB classB;
+
+    public ClassA(ClassB classB) {
+        this.classB = classB;
+    }
+
+    @Override
+    public void methodA(String input) {
+        // If the input is null or empty, return immediately
+        if (input == null || input.isEmpty()) {
+            return;
+        }
+
+        //Find the index of the first space character in the input string.
+        int spaceIndex = input.indexOf(' ');
+        //Extract the word from the beginning of the input string up to the space character.
+        String word = (spaceIndex == -1) ? input : input.substring(0, spaceIndex);
+        //Extract the remaining part of the input string after the space character.
+        String remaining = (spaceIndex == -1) ? "" : input.substring(spaceIndex + 1);
+
+        //Print the word extracted from the input string.
+        System.out.println("ClassA: " + word);
+        classB.methodB(remaining);
+    }
+}

java/com/engflow/cycleexample/class_a/ClassATest.java

@@ -0,0 +1,42 @@
+package com.engflow.cycleexample.class_a;
+
+import com.engflow.cycleexample.class_b.ClassB;
+import com.engflow.cycleexample.class_c.ClassC;
+import org.junit.Test;
+import static org.junit.Assert.assertTrue;
+
+public class ClassATest {
+    private static class TestClassB extends ClassB {
+        boolean methodBCalled = false;
+
+        public TestClassB(ClassC classC) {
+            super(classC);
+        }
+
+        @Override
+        public void methodB(String input) {
+            methodBCalled = true;
+        }
+    }
+
+    @Test
+    public void testMethodA() {
+        // Create a ClassC instance with a null ClassA object
+        ClassC classC = new ClassC(null);
+
+        // Create a TestClassB instance with the ClassC object
+        TestClassB testClassB = new TestClassB(classC);
+
+        // Create a new ClassA instance with the TestClassB object
+        ClassA classA = new ClassA(testClassB);
+
+        // Properly initialize classC with classA
+        classC.setClassA(classA);
+
+        // Call methodA on classA with a sample input
+        classA.methodA("sample input");
+
+        // Verify that methodB on the TestClassB was called
+        assertTrue(testClassB.methodBCalled);
+    }
+}

java/com/engflow/cycleexample/class_b/BUILD

@@ -0,0 +1,21 @@
+package(default_visibility = ["//visibility:public"])
+
+java_library(
+    name = "class_b",
+    srcs = ["ClassB.java"],
+    deps = [
+        "//java/com/engflow/cycleexample/class_c:class_c",
+        "//java/com/engflow/cycleexample/interface_b:interface_b",
+    ],
+)
+
+java_test(
+    name = "class_b_test",
+    srcs = ["ClassBTest.java"],
+    test_class = "com.engflow.cycleexample.class_b.ClassBTest",
+    deps = [
+        ":class_b",
+        "//java/com/engflow/cycleexample/class_c:class_c",
+        "//java/com/engflow/cycleexample/interface_b:interface_b",
+    ],
+)