With the latest I/O conference, Google has finally publicly announced its plans for its new runtime on Android. The Android RunTime, ART, is the successor and replacement for Dalvik, the virtual machine on which Android Java code is executed on. We’ve had traces and previews of it available with KitKat devices since last fall, but there wasn’t much information in terms of technical details and the direction Google was heading with it.

Contrary to other mobile platforms such as iOS, Windows or Tizen, which run software compiled natively to their specific hardware architecture, the majority of Android software is based around a generic code language which is transformed from “byte-code” into native instructions for the hardware on the device itself.

Over the years and from the earliest Android versions, Dalvik started as a simple VM with little complexity. With time, however, Google felt the need to address performance concerns and to be able to keep up with hardware advances of the industry. Google eventually added a JIT-compiler to Dalvik with Android’s 2.2 release, added multi-threading capabilities, and generally tried to improve piece by piece.

However, lately over the last few years the ecosystem had been outpacing Dalvik development, so Google sought to build something new to serve as a solid foundation for the future, where it could scale with the performance of today’s and the future’s 8-core devices, large storage capabilities, and large working memories.

Thus ART was born.

Architecture

First, ART is designed to be fully compatible with Dalvik’s existing byte-code format, “dex” (Dalvik executable). As such, from a developer’s perspective, there are no changes at all in terms of having to write applications for one or the other runtime and no need to worry about compatibilities.

The big paradigm-shift that ART brings, is that instead of being a Just-in-Time (JIT) compiler, it now compiles application code Ahead-of-Time (AOT). The runtime goes from having to compile from bytecode to native code each time you run an application, to having it to do it only once, and any subsequent execution from that point forward is done from the existing compiled native code.

Of course, these native translations of the applications take up space, and this new methodology is something that has been made possible today only due to the vast increases in available storage space on today’s devices, a big shift from the early beginnings of Android devices.

This shift opens up a large amount of optimizations which were not possible in the past; because code is optimized and compiled only once, it is worth to optimize it really well that one time. Google claims that it now is able to achieve higher level optimizations over the whole of an applications code-base, as the compiler has an overview of the totality of the code, as opposed to the current JIT compiler which only does optimizations in local/method chunks. Overhead such as exception checks in code are largely removed, and method and interface calls are vastly sped up. The process which does this is the new “dex2oat” component, replacing the “dexopt” Dalvik equivalent. Odex files (optimized dex) also disappear in ART, replaced by ELF files.

Because ART compiles an ELF executable, the kernel is now able to handle page handling of code pages - this results in possibly much better memory management, and less memory usage too. I’m curious what the effect of KSM (Kernel same-page merging) has on ART, it’s definitely something to keep an eye on.

The implications to battery life are also significant - since there is no more interpretation or JIT-work to be done during the runtime of an app, that results in direct savings of CPU cycles, and thus, power consumption.

The only downside to all of this, is that this one-time compilation takes more time to complete. A device’s first boot, and an application’s first start-up will be much increased compared to an equivalent Dalvik system. Google claims that this is not too dramatic, as they expect the finished shipping runtime to be equivalent or even faster than Dalvik in these aspects.

The performance gains over Dalvik are significant, as pictured above; the gains are roughly a 2x improvement in speed for code running on the VM. Google claimed that applications such as Chessbench that represent an almost 3x increase are a more representative projection of real-world gains that can be expected once the final release of Android L is made available.

Garbage Collection: Theory and Practice
POST A COMMENT

137 Comments

View All Comments

  • hahmed330 - Wednesday, July 2, 2014 - link

    One Stone... Three birds... Reply
  • Notmyusualid - Wednesday, July 2, 2014 - link

    Just switched my GS5 over to Art, from Dalvik, and Antutu result dropped by 8%...

    Yes, the choice is on the stock ROM, just goto developer options, and select runtime.
    Reply
  • tuxRoller - Thursday, July 3, 2014 - link

    Use ANY other benchmark. Who the hell knows how antutu works?
    For micro benchmarks try geekbench.
    If you're willing to do some compiling, linaro has a bunch of benchmarks it uses to determine progress.
    Reply
  • Notmyusualid - Thursday, July 3, 2014 - link

    Call me crazy, but I don't pay for apps....

    I take only the free ones.

    I see no free Geekbench on the Play Store.
    Reply
  • tuxRoller - Friday, July 4, 2014 - link

    I didn't realize you had to pay for it.
    Regardless, antutu is junk. Why? Because we don't know exactly what it does, or how it does it.
    The other option I mentioned is pick some of the linaro benchmark tools and compile them.
    I won't call you crazy for not buying apps because I don't know your situation. What I do, however, is try free versions and if they are good I buy them. They don't cost much and I don't waste battery with ads I'll ignore.
    Reply
  • Krysto - Thursday, July 3, 2014 - link

    I thought it was clear that the ART in L is NOT the one in KitKat, and has been revamped quite a bit. The final one, 5 months from now, will probably have big changes, too. Reply
  • Notmyusualid - Thursday, July 3, 2014 - link

    Will keep an eye out for it, but I'm expecting this to be no big deal now. Reply
  • ergo98 - Wednesday, July 2, 2014 - link

    Too much has been made regarding AOT and JIT. Note that Dalvik generally only JITs the DEX once, storing the result in /data/dalvik-cache.

    The big difference between Dalvik and ART is simply that ART was rewritten from the ground up based upon everything they learned from the Dalvik experience.
    Reply
  • errorr - Thursday, July 3, 2014 - link

    That and because of the Oracle lawsuit over Dalvik which is nicely mooted by ART. Reply
  • doubledeej - Wednesday, July 2, 2014 - link

    It never ceases to amaze me how many problems that were solved decades ago in computing are problems on modern computing platforms.

    Real compilation of code has been around forever -- the norm, in fact, for desktop and server computing with a few notable exceptions. Yet somehow taking what effectively amounts to interpreting code (just-in-time compilation is very similar to interpretation) and switching to compiling it ahead of execution is being touted as a new idea.

    The fact that Android has pretty much been completely reliant upon JIT running in a VM has always made me scratch my head. As clearly spelled out in the article, it cause huge performance issues, along with significant hits to battery life. And we're talking about mobile devices where we've got relatively low-power CPUs and GPUs, little memory, and finite battery capacity. But it has been the way that Android has worked from the beginning. Crazy that it hasn't really been addressed until now.

    And the idea that operating systems and development languages be in charge of garbage collection, and people being surprised that it causes performance hits, seems odd to me too. Managing your own memory isn't that hard to do. And it is a hell lot more efficient doing it yourself than making the language or OS figure out how to do it. It's a "clean up your own mess and put things back where you want them" vs. "make someone else do it and let them try to figure out where things go" situation. It might make development easier for entry-level developers, but it certainly isn't an efficient way to do things when performance and user experience are important.

    Because the developers that I work with aren't accustomed to managing memory, we're constantly running into issues. We've got scripts that allocate dozens or hundreds of megabytes of RAM and don't free it when they're done. They'll go through 3, 4, or 5 more of these processes within a single script, not freeing memory they're done with along the way, so by the time the script is done running hundreds of megabytes that aren't needed are still tied up. Because the language can't be sure if data is going to be used again it hangs around until the script has finished running.

    Create dozens or hundreds of instances of one of those scripts and you've got a performance nightmare. Relying on a language or OS to do garbage collection will have the same net result.
    Reply

Log in

Don't have an account? Sign up now