Re: Serious issue



"Karl Uppiano" <Karl_Uppiano@xxxxxxx> wrote in message news:g1wil.1407$eK2.30@xxxxxxxxxxxxxxxxxxxxxxx


"EricF" <efriedNoSpam@xxxxxxxxx> wrote in message news:HIPhl.228$Qc.169@xxxxxxxxxxxxxxxxxxxxxxxxx
In article <0v4eo4d0i6ra315hcd8nto57id1343rvul@xxxxxxx>, Tilly<nospam@xxxxxxxxxx> wrote:
On Fri, 30 Jan 2009 14:54:31 -0500, John W Kennedy <jwkenne@xxxxxxxxxxxxx>
wrote:

On 1/30/09 9:16 AM, Tilly wrote:
Why is java so slow compared to other languages performing the same tasks on
the
same hardware?

In general, it isn't. It is typically slow to /start/, because the Java
VM and its library must be loaded. But it generally runs about as fast
as other languages -- in some cases, demonstrably faster.

Except for Java ME. That tends to run slowly because it's generally
interpreted.


This hasn't been our experience can you provide links to examples of timings
for comparisons please? We looked but couldnt find any.

In general we find java running 10 times slower than C++ and this is a problem
for us. We need to decide whether to keep using it or go elsewhere. Startup is
(as you say) also a serious issue.

Do you know of any serious language comparison articles. Again we cant find
anything other that the usual PC World "Wow we love this" - (cause its new and
we can write about it ad nauseum without actually telling you anything useful)

Anything that does a serious comparison would be welcome
cheers.

Tilly
-------------------------
Not the Hun

I recently did some benchmarks of Java and C++ merging pdf files. The Java
code was a bit faster, but the pdf libraries differed so this was not a true
comparison.

Which version of Java are you using? 1.6 seems to be quite a bit faster than
older versions but it all depends on what you are doing.

If you are using a modern version of Java, C++ is 10 faster and this is
important to you, then stick with the C++.

Back in 1995 (IIRC) we wrote a Java 3280 terminal emulator applet to accompany an ActiveX version. We were told to port the ActiveX design. After about three days of that, we abandoned the porting idea, and wrote it in Java, using Java idioms, style and design patterns. The resulting Java applet ran faster than the ActiveX controls.

I think our final Java design was actually superior to the ActiveX controls, and that made the real difference -- More than the choice of languages. And we were using JDK 1.2.

P.S., That should read "3270 terminal". TN3270E to be exact...

.

Relevant Pages

  • Re: Can Java Programmer Learn C++ Quickly?
    ... I'd argue that this isn't true if he has been writing good Java. ... issues at the design level, ... decisions in these libraries. ... investing a bit of time learning the STL. ...
    (comp.lang.java.programmer)
  • Re: C++ or Java ?
    ... benefit of an overall design will be a mess, ... Is C++ better than Java? ... C++ is a more complex language than Java and that makes it more ... libraries and framworks one ends up having to learn. ...
    (comp.os.linux.development.apps)
  • Re: Serious issue
    ... Except for Java ME. ... but the pdf libraries differed so this was not a true ... I think our final Java design was actually superior to the ActiveX controls, and that made the real difference -- More than the choice of languages. ...
    (comp.lang.java)
  • Re: Pretty obvious. Isnt it?
    ... Java1.2 and on) Java APIs cause advertising that it supports "Java1.1.4" ... Easier portability of an app between Sun Java and .NET means big ... Plugs VCL and ActiveX controls ...
    (microsoft.public.dotnet.vjsharp)
  • Re: java doesnt work
    ... > I had tried changing betoween the JVM and Sun java before, ... >> Supports Java Applets = True ... >> Supports ActiveX Controls = True ... >>> Oh and the Dancing Duke was dancing ...
    (microsoft.public.windows.inetexplorer.ie6.browser)