Crate mobench 

mobench

Command-line tool for building and running Rust benchmarks on mobile devices.

Overview

mobench is the CLI orchestrator for the mobench ecosystem. It handles:

• Building - Compiles Rust code for Android/iOS and packages mobile apps
• Running - Executes benchmarks locally or on BrowserStack devices
• Reporting - Collects and formats benchmark results

Installation

cargo install mobench

Quick Start

# Initialize a benchmark project
cargo mobench init --target android --output bench-config.toml

# Build for Android
cargo mobench build --target android

# Build for iOS
cargo mobench build --target ios

# Run locally (no device required)
cargo mobench run --target android --function my_benchmark --local-only

# Run on BrowserStack (use --release for smaller APK uploads)
cargo mobench run --target android --function my_benchmark \
    --iterations 100 --warmup 10 --devices "Google Pixel 7-13.0" --release

Commands

Command	Description
init	Initialize a new benchmark project
build	Build mobile artifacts (APK/xcframework)
run	Execute benchmarks locally or on devices
ci run	Run standardized CI orchestration (summary.json, summary.md, results.csv)
doctor	Validate local/CI prerequisites and configuration
config validate	Validate run config + matrix contract
devices resolve	Resolve deterministic device sets from matrix/profile
fixture ...	Fixture lifecycle helpers (init, build, verify, cache-key)
report ...	Render markdown and publish sticky PR comments
list	List discovered benchmark functions
fetch	Retrieve results from BrowserStack
package-ipa	Package iOS app as IPA
package-xcuitest	Package XCUITest runner

Output Directory

All build artifacts are written to target/mobench/ by default:

target/mobench/
├── android/           # Android project and APK
└── ios/               # iOS project, xcframework, and IPA

Use --output-dir to customize the output location.

Configuration

Benchmarks can be configured via command-line arguments or a TOML config file:

target = "android"
function = "my_crate::my_benchmark"
iterations = 100
warmup = 10

[browserstack]
app_automate_username = "${BROWSERSTACK_USERNAME}"
app_automate_access_key = "${BROWSERSTACK_ACCESS_KEY}"
project = "my-project"

BrowserStack Integration

The CLI integrates with BrowserStack App Automate for running benchmarks on real devices. Set credentials via environment variables:

export BROWSERSTACK_USERNAME="your_username"
export BROWSERSTACK_ACCESS_KEY="your_access_key"

Crate Ecosystem

This crate is part of the mobench ecosystem:

• mobench (this crate) - CLI tool
• mobench-sdk - Core SDK with timing harness and build automation
• mobench-macros - #[benchmark] proc macro

Note: The mobench-runner crate has been consolidated into mobench-sdk as its timing module.

CLI Flags

Global flags available on all commands:

• --dry-run - Preview what would be done without making changes
• --verbose / -v - Enable detailed output showing all commands

Modules

• config - Configuration file support for mobench.toml

Modules

config

Configuration file support for mobench.

summarize

Types and logic for the ci summarize command.

Structs

DeviceSelection

Device input sources used by RunRequest.

ExtractedBenchmarkResult

Extracted benchmark result for a single device.

Report

Standardized output file locations produced by a run.

RunRequest

Programmatic request payload for running a mobench benchmark flow.

RunResult

Result of a programmatic mobench run request.

Enums

MobileTarget

Mobile platform target for build/run operations.

Functions

extract_benchmark_summary

Extract a unified summary from per-device benchmark results.

run

run_request

Executes a RunRequest by invoking the current mobench binary and normalizing outputs.