Linked by Thom Holwerda on Mon 6th Oct 2008 10:37 UTC, submitted by John Mills
Thread beginning with comment 332714
To view parent comment, click here.
To read all comments associated with this story, please click here.
To view parent comment, click here.
To read all comments associated with this story, please click here.
Mono has improved in performance quite a lot since 1.2.6 and is quickly gaining ground on Java6-server, especially with the new Linear IR branch which you can read about at http://www.mono-project.com/Linear_IL
The Linear IR branch is poised to make it into Mono 2.2 and is already as much as 30% faster than Mono 2.0 and there are lots more optimizations that can be done (and indeed, just last week SIMD optimizations were implemented which increase the performance even more).
That is excellent news. Do you have a link to any benchmarks ? (especially against JRE 1.6u10).
Mono is also fully open source.
As far as waivers from Sun, I am unaware of these. Could you provide a link to these waivers?
I have never seen nor heard of any such waivers before.
http://java.sun.com/javase/6/docs/legal/license.html
See section 2. That is a patent grant for all implementations of the Java specification (and you don't have to ask Sun for permission to use it). However you don't have the right to call it "Java" unless you pass Sun's compatibility tests (since it wants to avoid fragmentation). Unfortunately it doesn't look like Microsoft will ever do this for .NET
This is all subjective, and let's not forget that Mono can continue even if Microsoft moves to something else next week, next month, next year, or next decade.
Mono is, after-all, a Free Software project that anyone can continue to maintain. It is not in any way dependent upon the life span of Microsoft's .NET just as C in Linux has been in use long after much of the Windows world switched to C++.
Completely true. What you're missing is that while Mono would soldier on as soon as there is fragmentation then adoption would be split. While it is not absolutely necessary to avoid this (after all, part of the point of Free Software is the 'right to fork'), this really dissipates developer effort and confuses users. It should be avoided for practical reasons.
Despite my preference for Java I'd like to thank you once again all your efforts with Mono (the community still needs it). It's nice to have choice.
Edited 2008-10-06 21:17 UTC
"
Mono has improved in performance quite a lot since 1.2.6 and is quickly gaining ground on Java6-server, especially with the new Linear IR branch which you can read about at http://www.mono-project.com/Linear_IL
The Linear IR branch is poised to make it into Mono 2.2 and is already as much as 30% faster than Mono 2.0 and there are lots more optimizations that can be done (and indeed, just last week SIMD optimizations were implemented which increase the performance even more).
Mono has improved in performance quite a lot since 1.2.6 and is quickly gaining ground on Java6-server, especially with the new Linear IR branch which you can read about at http://www.mono-project.com/Linear_IL
The Linear IR branch is poised to make it into Mono 2.2 and is already as much as 30% faster than Mono 2.0 and there are lots more optimizations that can be done (and indeed, just last week SIMD optimizations were implemented which increase the performance even more).
That is excellent news. Do you have a link to any benchmarks ? (especially against JRE 1.6u10). "
There were a couple of benchmarks that the JIT developers have been using for comparison purposes. They've been using the Debian Language Shootout benchmark programs, the Portable.NET benchmark and some others. The "Benchmark Results" section towards the bottom of the page I linked to earlier names the benchmarks used and what the performance improvements are for those different tests.
"
Mono is also fully open source.
As far as waivers from Sun, I am unaware of these. Could you provide a link to these waivers?
I have never seen nor heard of any such waivers before.
Mono is also fully open source.
As far as waivers from Sun, I am unaware of these. Could you provide a link to these waivers?
I have never seen nor heard of any such waivers before.
http://java.sun.com/javase/6/docs/legal/license.html
See section 2. That is a patent grant for all implementations of the Java specification (and you don't have to ask Sun for permission to use it). However you don't have the right to call it "Java" unless you pass Sun's compatibility tests (since it wants to avoid fragmentation). Unfortunately it doesn't look like Microsoft will ever do this for .NET
" This is very interesting, I had not known about this before.
Thanks for the link, I'll have to read it more carefully later.
"
This is all subjective, and let's not forget that Mono can continue even if Microsoft moves to something else next week, next month, next year, or next decade.
Mono is, after-all, a Free Software project that anyone can continue to maintain. It is not in any way dependent upon the life span of Microsoft's .NET just as C in Linux has been in use long after much of the Windows world switched to C++.
This is all subjective, and let's not forget that Mono can continue even if Microsoft moves to something else next week, next month, next year, or next decade.
Mono is, after-all, a Free Software project that anyone can continue to maintain. It is not in any way dependent upon the life span of Microsoft's .NET just as C in Linux has been in use long after much of the Windows world switched to C++.
Completely true. What you're missing is that while Mono would soldier on as soon as there is fragmentation then adoption would be split. While it is not absolutely necessary to avoid this (after all, part of the point of Free Software is the 'right to fork'), this really dissipates developer effort and confuses users. It should be avoided for practical reasons. "
That all assumes that things fork, though, which wouldn't necessarily happen. (hopefully it wouldn't, actually).
Ideally, the continuation of Mono would remain compatible with the specified portions of .NET and Mono would just be extended (via extension methods for the core and new libraries/compiler features that do not break existing code).
Despite my preference for Java I'd like to thank you once again all your efforts with Mono (the community still needs it). It's nice to have choice.
Choice is indeed good to have
And thank you for your gratitude for our work, it is very much appreciated!
Mono has improved in performance quite a lot since 1.2.6 and is quickly gaining ground on Java6-server, especially with the new Linear IR branch which you can read about at http://www.mono-project.com/Linear_IL
Mono still has a long way to go, from the looks of things. http://shootout.alioth.debian.org/u32q/benchmark.php?test=all&lang=...
Java is up to 6.9 times faster and is on average twice as fast. Granted, they use the -server flag which turns on much more aggressive optimizations hence the much higher performance and larger memory consumption. Nevertheless, there is still a huge delta between Mono and Java.
"
Mono has improved in performance quite a lot since 1.2.6 and is quickly gaining ground on Java6-server, especially with the new Linear IR branch which you can read about at http://www.mono-project.com/Linear_IL
Mono has improved in performance quite a lot since 1.2.6 and is quickly gaining ground on Java6-server, especially with the new Linear IR branch which you can read about at http://www.mono-project.com/Linear_IL
Mono still has a long way to go, from the looks of things. http://shootout.alioth.debian.org/u32q/benchmark.php?test=allâ©...
Java is up to 6.9 times faster and is on average twice as fast. Granted, they use the -server flag which turns on much more aggressive optimizations hence the much higher performance and larger memory consumption. Nevertheless, there is still a huge delta between Mono and Java. "
You are comparing Java6 -server with, at best, Mono 2.0.0 (hard to tell because Mono 1.9 also reported 2.0.0 afaik), the optimizations are happening in the post-2.0 code base (as I mentioned earlier). If you compare Java6 -client (which is the default) against Mono, they are actually neck-and-neck, possibly with Mono even coming out ahead.
Note also that raw speed isn't the only factor, Mono consistently uses a lot less memory than Java. For one of those tests, Java uses more than 10x the amount of memory than Mono uses.
But yes, you are correct in saying that Mono still has a ways to go if you want to compare Java6 -server with Mono (which doesn't have any special "server" optimizations).





Member since:
2007-06-21
1) It is very, very much faster (especially Java 1.6u10 which accelerates all graphics operations)
Example:
http://blogs.sun.com/jag/entry/current_state_of_java_for
http://hal.inria.fr/inria-00312039/en
Mono has improved in performance quite a lot since 1.2.6 and is quickly gaining ground on Java6-server, especially with the new Linear IR branch which you can read about at http://www.mono-project.com/Linear_IL
The Linear IR branch is poised to make it into Mono 2.2 and is already as much as 30% faster than Mono 2.0 and there are lots more optimizations that can be done (and indeed, just last week SIMD optimizations were implemented which increase the performance even more).
Mono is also fully open source.
As far as waivers from Sun, I am unaware of these. Could you provide a link to these waivers?
I have never seen nor heard of any such waivers before.
Thanks for the compliment ;-)
This is all subjective, and let's not forget that Mono can continue even if Microsoft moves to something else next week, next month, next year, or next decade.
Mono is, after-all, a Free Software project that anyone can continue to maintain. It is not in any way dependent upon the life span of Microsoft's .NET just as C in Linux has been in use long after much of the Windows world switched to C++.