Maximum Java performance on a Windows 2K platform with JDK 1.4

Thread view:

Michael - 30 Nov 2004 05:27 GMT

Hi,

I was wondering if anyone knew what is the best performance (in ms) I
can get from a
Java program (assuming it has all the CPU time it can use)

From my experience, I am guessing it's threahold is about 10ms.

I have an application which needs to have near real-time performance for
one or more components in the architecture.

Reply to this Message

Grant Wagner - 30 Nov 2004 16:08 GMT

> Hi,
>
> I was wondering if anyone knew what is the best performance (in ms) I
> can get from a
> Java program (assuming it has all the CPU time it can use)

I'm not sure what this means. Even the simplest Java program will have
significant startup costs, resulting in an execution time from "java
YourProgram" to exit of several thousand milliseconds.

> From my experience, I am guessing it's threahold is about 10ms.

If you're talking about the accuracy of System.currentTimeMillis(), that's
platform dependant. In the case of Windows:

<url: http://mindprod.com/jgloss/time.html />

> I have an application which needs to have near real-time performance for
> one or more components in the architecture.

Java (where the garbage collector can kick in any time) is the wrong
technology to use if you need real-time performance. Although there is a
Real Time Specification for Java <url: https://rtsj.dev.java.net />, I don't
know if there's any production implementation of the spec.

--
Grant Wagner <gwagner@agricoreunited.com

Reply to this Message

Steve W. Jackson - 30 Nov 2004 16:41 GMT

>:Hi,
>:
[quoted text clipped - 6 lines]
>:I have an application which needs to have near real-time performance for
>:one or more components in the architecture.

Best performance doing what? It depends on details.

Signature

Steve W. Jackson
Montgomery, Alabama

Reply to this Message

Maximum Java performance on a Windows 2K platform with JDK 1.4

Free Magazines