data may be subject to informational limitations due to government restrictions. Includes only Android, iPhone & iPad users. Figures may understate true ChatGPT user base (e.g., desktop or mobile webpage Ongoing AI Product Rollouts Meta Users* 3.4B+ Apple Devices** 2.35B Google 4.9B Search Users, 3B+ Android Users, 1.5B AI Overviews Users & 1B+ Assistant Devices TikTok Users*** 1B+ Amazon 600MM+ Alexa Smart Replies • Jules • Imagen 4 • Veo 3 • Lyria 2 • Flow TV • Project Moohan • Glasses with Android XR • Magentic-UI • Copilot Studio multi-agent orchestration • GitHub Copilot asynchronous functioning

Negligible overhead (a few nanoseconds per zone) Cross-platform (Windows, Linux, macOS, iOS, Android, WASM*) Hybrid profiling capabilities (sampling and/or instrumentation) (CPU and GPU instrumentation) Negligible overhead (a few nanoseconds per zone) Cross-platform (Windows, Linux, macOS, iOS, Android, WASM*) Hybrid profiling capabilities (sampling and/or instrumentation) Tracing capabilities Negligible overhead (a few nanoseconds per zone) Cross-platform (Windows, Linux, macOS, iOS, Android, WASM*) Hybrid profiling capabilities (sampling and/or instrumentation) Tracing capabilities

Negligible overhead (a few nanoseconds per zone) Cross-platform (Windows, Linux, macOS, iOS, Android, WASM*) Hybrid profiling capabilities (sampling and/or instrumentation) (CPU and GPU instrumentation) Negligible overhead (a few nanoseconds per zone) Cross-platform (Windows, Linux, macOS, iOS, Android, WASM*) Hybrid profiling capabilities (sampling and/or instrumentation) Tracing capabilities Negligible overhead (a few nanoseconds per zone) Cross-platform (Windows, Linux, macOS, iOS, Android, WASM*) Hybrid profiling capabilities (sampling and/or instrumentation) Tracing capabilities

project restricted to Linaro members ● Three sub-projects: ○ Arm Compute Library ○ Arm NN ○ Android NN Driver ● Arm Compute Library has been integrated by: ○ MATLAB Coder ○ ONNX RuntimeArm platform pixel2 (snapdragon 835), mate10/mate10pro (kirin 970), p20/p20pro (kirin 970) -target=arm64-linux-android -mattr=+neon llvm firefly rk3399, rock960, ultra96 -target=aarch64-linux-gnu -mattr=+neon rasp3b

projects targeting Android or embedded • Improve IntelliSense tooltip for make_shared/make_unique/emplace_back/etc • Implement the equivalent of Wmissing-braces • Update the bundled Android NDK to Release

large: - Toolchains, NDKs - The space is very continuous and large (Android NDK api level) - Not easy to generate a ton of android_ndk_apiXX_archYY.cps - Avoid as much as possible - Regular libraries

Mountain Pass Sports Bar Anyone welcome!Help improve RealtimeSanitizer ● Adding support ○ Windows ○ Android ○ Other architectures (than x86-64 and arm64) ● Adding new interceptors ● Testing out experimental Suppression lists ● Later ○ Detection of nondeterministic loops ● Someday? ○ Windows support ○ Android/iOS support ○ Other architectures beyond Arm64 and x86_64Optimizer Compiler backend Linker Assembler

and newer - Android smartphone x86-64 (AMD64) System V ABI - Linux server - old Mac x86-64 (AMD64) Microsoft ABI - Windows device armv7-a System V ABI - ancient iPhone - low-end Android smartphone

框架优化的要点 • 共享inode cache • 共享data cache的映射 • GETATTR流程 • ⽂件读取流程 • 相关⼯作 • extFUSE • google android12 passthrough什么是eBPF • ebpf是不同环境下内核配置， 调试，监控⼯具 • map映射 • 验证器 • Hook • Helper api配置TCP Initial

to 13.1.0 • Clang: 11.1.0 to 15.0.7 • Apple Clang: 14.0.0.14000029 to 15.0.0.15000040 • Android NDK: r18 (Clang 7.0) to r26 (Clang 17.0.2) • Emscripten: 3.1.20 (Clang 16.0.0) to latest (3.1