/var/opinion - Is Hardware Catching Up to Java?
I had a wonderful experience chatting with the folks at RapidMind. The company provides C/C++ programmers with a very clever tool to help them exploit the parallel processing capability of multiple core CPUs. The idea is to make parallel processing at least somewhat transparent to the programmer. Programmers can use RapidMind to improve multicore performance of existing applications with minimal modifications to their code and build new multicore applications without having to deal with all of the complexities of things like thread synchronization.
RapidMind and any other solutions like it are invaluable and will remain so for a long, long time. C/C++ is an extremely well-entrenched legacy programming platform. I can't help but wonder, however, if the trend toward increased parallel processing in hardware will create a slow but steady shift toward inherently multithreaded, managed programming platforms like Java.
The inherent trade-off to adding cores to a CPU poses an interesting problem for any programming platform. As you add cores, you increase power consumption, which generates more heat. One way to reduce heat is to lower the processing speed of the individual cores, or at the very least, keep them from advancing in speed as quickly as they might have if history could have continued along its current path.
As a result, any single thread of execution would run faster on a speedy single-core CPU than it would on a slower core of a multicore CPU. The seemingly obvious answer is to split up the task into multiple threads. Java is multithreaded by nature, right? Therefore, all other things being equal, a Java application should be a natural for multicore CPU platforms, right?
Not necessarily. If Java was a slam-dunk superior way to exploit parallel processing, I would have posed this shift to Java as a prediction, not a question. It's not that simple. For example, don't let anyone tell you that Java was built from the ground up to exploit multiple processors and cores. It ain't so. Java's famous garbage collection got in the way of parallel programming at first. Java versions as late as 1.4 use a single-threaded garbage collector that stalls your Java program when it runs, no matter how many CPUs you may have.
But Java's multithreaded design more easily exploits parallel processing than many other languages, and it ended up lending itself to improvements in garbage collection. JDK 5.0 includes various tuning parameters that may minimize the impact of garbage collection on multiple cores or CPUs. It's not perfect, and you have to take into account the way your application is designed, but it is a vast improvement, and the improvement is made possible by the fact that Java was well conceived from the start.
Obviously, these features aren't enough. IBM builds a lot of additional parallelism into its WebSphere application server. In addition, IBM and other researchers are working on a Java-related language X10, which is designed specifically to exploit parallel processing (see x10.sourceforge.net/x10home.shtml).
It is also interesting to note that when Intel boasts about its quad-core performance on Java, its numbers are based on using the BEA jRockit JVM, not the Sun JVM. See www.intel.com/performance/server/xeon/java.htm for the chart and more information. I suspect Intel used this JVM because the BEA JVM employs garbage collection algorithms that are more efficient for use on multiple cores.
The fact that Intel used a non-Sun JVM makes this whole question of the future of Java on multicore CPUs interesting. I won't discount the possibility that Intel may have tuned its code to work best with the BEA JVM. But it is a significant plus for Java that you can choose a JVM best suited for the hardware you have. The big plus is that you still have to learn only one language, Java, and this does not limit your choice of architecture or platforms. If you run your application on a multicore platform, you can choose between JVMs and JVM-specific tuning parameters to squeeze extra performance out of your application.
Now, think a bit further into the future. Imagine parallelism showing up in a future generation of cell phones or other small devices. What better language than a platform-neutral one to take advantage of this future?
Some disclaimers are in order, though. First, I don't think any tool or language will soon make optimal programming for multiple cores or CPUs totally transparent. Java has the potential to exploit multiple cores with less impact to the programmer than many other languages, however, which is why I think Java's future is getting even more promising with this new hardware trend. Second, although I think Java has the edge, other managed, multithreaded languages will no doubt ride this same wave. Python, Ruby and your other favorite language in this genre probably have an equally bright future.
Nicholas Petreley is Editor in Chief of Linux Journal and a former programmer, teacher, analyst and consultant who has been working with and writing about Linux for more than ten years.
Getting Started with DevOps - Including New Data on IT Performance from Puppet Labs 2015 State of DevOps Report
August 27, 2015
12:00 PM CDT
DevOps represents a profound change from the way most IT departments have traditionally worked: from siloed teams and high-anxiety releases to everyone collaborating on uneventful and more frequent releases of higher-quality code. It doesn't matter how large or small an organization is, or even whether it's historically slow moving or risk averse — there are ways to adopt DevOps sanely, and get measurable results in just weeks.
Free to Linux Journal readers.Register Now!
- Hacking a Safe with Bash
- Django Models and Migrations
- Secure Server Deployments in Hostile Territory, Part II
- The Controversy Behind Canonical's Intellectual Property Policy
- Home Automation with Raspberry Pi
- Huge Package Overhaul for Debian and Ubuntu
- Shashlik - a Tasty New Android Simulator
- Embed Linux in Monitoring and Control Systems
- KDE Reveals Plasma Mobile
- diff -u: What's New in Kernel Development