Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.lang.java.programmer > #35635 > unrolled thread

Re: Java is faster!

Back to article view | Back to comp.lang.java.programmer

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.

Contents

  Re: Java is faster! Askfor <askfor@nowhere.net> - 2016-09-23 20:34 +0200

#35635 — Re: Java is faster!

Stefan Ram wrote:
>        "Java is faster!
>        In terms of throughput in a long running system ...
>        For specific applications ...
>        On a well tuned JVM ...
>        With near-optimal GC strategy selected ..."
>
>    says Chandler Carruth, one of Google's C++ experts. (JFTR)
>

Long ago I have been using Borland C compiler on DOS. One day I wanted 
to see how it generates x86 assembly code. I have written a simple C 
program and instructed compiler to generate assembly source.

When it was done I was impressed. The code was barely readable, but 
efficient. Jumps were made in totally wild ways, but it made sense.

If human being tried to write the same code in assembly, he or she would 
have had to split it in some logically structured and readable  sections 
because it was relatively large piece of assembly code. Because of it, 
the code probably wouldn't have been optimal. Compiler, on the other 
hand, did not bother with it. It simply organized code in most efficient 
way, ignoring readability.

So, it is possible for JVM to generate large chunks of machine code when 
it needs that, and it might be more optimized because it has "bigger 
picture". Higher leve instructions generate larger chunks of code, which 
might get better optimization.

[toc] | [standalone]

Back to top | Article view | comp.lang.java.programmer

csiph-web