Slashdot Mirror

There was a JavaOne session on it last year: http://developers.sun.com/learning/javaoneonline/j1sessn.jsp?sessn=TS-13820&yr=2007&track=5

As far as I can see from the comparison of these FSes, BTRFS is a promising file system for Linux and is under development. Some say that it will be the ZFS of Linux or even better. I think time will say.
Some other say, now that Oracle owns Sun, Oracle can change the license of ZFS from CDDL to GPL2 and port to Linux. But porting ZFS to Linux it's another story...

As far as I can see from the comparison of these FSes, BTRFS is a promising file system for Linux and is under development. Some say that it will be the ZFS of Linux or even better. I think time will say.
Some other say, now that Oracle owns Sun, Oracle can change the license of ZFS from CDDL to GPL2 and port to Linux. But porting ZFS to Linux it's another story...

So when you say "sftp is an order of magnitude slower" you really mean ".... on openssh running on cygwin running on Windows". (I'm assuming that your workstations are not bottlenecks, since these systems typically have more CPU bandwidth than they know what to do with.)

I'm not an expert on this stuff, but I have to wonder if your sshd daemons are fully utilitizing your CPUs. Perhaps this is a stupid suggestion, but have you profiled core usage and verified that the cygwin and openssh in copssh are both compiled with multithreading support?

One obvious suggestion is that you try a native-to-Windows SSH server. But there don't seem to be any! Not open source, not commercial. Mind-boggling.

If there is a commercial Windows SSH server, you probably should give it a try. Your boss may hate spending money, but he'd hate being hacked even more.

And even if you guys are a Windows shop, you might consider serving your FTP off a system with a performance-oriented OS. Being a Sun employee, I'm required to tell you that Solaris is way cooler than Linux for high-performance apps. But even with Linux, the CPU bandwidth you need to do this properly is not that expensive.

</shameless plug>

So when you say "sftp is an order of magnitude slower" you really mean ".... on openssh running on cygwin running on Windows". (I'm assuming that your workstations are not bottlenecks, since these systems typically have more CPU bandwidth than they know what to do with.)

I'm not an expert on this stuff, but I have to wonder if your sshd daemons are fully utilitizing your CPUs. Perhaps this is a stupid suggestion, but have you profiled core usage and verified that the cygwin and openssh in copssh are both compiled with multithreading support?

One obvious suggestion is that you try a native-to-Windows SSH server. But there don't seem to be any! Not open source, not commercial. Mind-boggling.

If there is a commercial Windows SSH server, you probably should give it a try. Your boss may hate spending money, but he'd hate being hacked even more.

And even if you guys are a Windows shop, you might consider serving your FTP off a system with a performance-oriented OS. Being a Sun employee, I'm required to tell you that Solaris is way cooler than Linux for high-performance apps. But even with Linux, the CPU bandwidth you need to do this properly is not that expensive.

</shameless plug>

And in the far corner, weighing in at just over the weight of an IBM mainframe... SUN MODULAR DATACENTER!

He isn't out-of-date. I have to suffer Java 1.6 at my college. The VM is still slow to start, and the GUI libraries still suck.

Nah, not perfect though, but let's don't hyperbolize it either: it does not sucks really anymore. Personally I write a lot of Swing stuff and I would not say it that sucks as before Java 1.4 times. At least now 1.6 I'd say much better than even GTK or QT. Why?

Because:

• performance is same as QT or GTK. No, sorry. Performance sometimes better e.g. when you have a really huge amount of records and put it into table -- in that case I wish GTK performed better and GTK-based wxWindows did not crash.
• In Java Swing you can create any possible widget you want per a very reasonable time, having it relatively easy. Example: http://download.java.net/javadesktop/scenario/demos/repaint-NestedTest.jnlp
• While Swing looks like a cockpit of 747, yet it is over-flexible and thus very powerful. Yes, it is scary to new comers, but once you figured out, you have to say it is really cool, extendable and portable.
• To write these kind of GUIs, including Quake-2 port in QT or GTK would take much more time for you. Not saying it is impossible. I am saying about time/cost/quality/competence/final-result-asap things.
• I love JNLP concept and it works rock stable on controlled or semi-controlled networks, once done properly. JNLP (Java Web Start) is way better for internal software and enterprises than Ajax thing or equivalent GTK/QT stuff, since I've done all of this and can compare what was the best to achieve with average-skilled programmers, cheapest and fastest to develop-and-deploy. And you must know that either. :-)

Additionally, you can use your own look-and-feel (including GTK's) etc. Of course, it has its side effects: you have to emulate specific platform etc. SWT does it natively, but it suffers from portability a lot, plus some parts are written in plain Java. Since API of various platforms changes and sometimes very drastically (like Apple is gonna drop Carbon one day), SWT renders to be a fool trap that always trying to catch up current status, but actually performs great only on Windows.

Add to the story really great try of NetBeans team to bring you GUI visual editor and you have to agree that at least you can create GUI stuff literally in front of your boss's eyes. I agree, Swing not the best yet, but also not that horrible as before anymore either. That's why we can commit our patches to Java 7! :-)

I know you were joking, but have you used Java's decimal library before?

BigDecimal opA = new BigDecimal(a);
BigDecimal opB = new BigDecimal(b);
BigDecimal result = opA.add(opB);
System.out.println(result);

Not quite as bad as your fake code, but the difference is... BigDecimal is actually a real class that exists in the Java standard library.

The same code in C#:

decimal result = a + b;

(provided you don't overflow the decimal value)

Funny how all MY links were on the first page, and were higher.

Also, your Sun link isn't on the first page.
Also, your Sun link doesn't define JBOD.
Also, your Sun link is arguing against you.
The guy says he got 144 drives in 6 JBODs.

http://blogs.sun.com/brendan/resource/analytics-1/analytics-1-hardware-crop.png

Yeah, that's 24 "1 TB" drives in 6 JBODs.
Nothing in that blog post says what the Sun box is presenting the volumes (externally) as. All this is about is using their fun little GUI to tell that he cabled it wrong (on purpose, to make the blog post).

Your dciginc link also doesn't define JBOD, but it proves itself to be even more wrong than you:
"JBODs use aging RAID5/6 protection schemas to maintain data integrity. Raid 5/6 is becoming more impractical as disk drive sizes grow and rebuild times to recover failed disk drives increase. Further, JBODs provide no or limited self-healing capabilities to detect errors that can occur in disk drives over time and automatically correct them."

JBOD using RAID 5/6? What?

Your aberdeeninc link is for a 4-channel SAS box.
It uses the term JBOD a lot, but it never defines it. It's marketing.

Looks like you couldn't find anything at all to support your claim, yet you posted some arbitrary links to try to save face.

When I google JBOD, from the *first page* of results:

http://www.aberdeeninc.com/abcatg/kitjbod-1003.htm

http://opensolaris.org/jive/thread.jspa?threadID=83108&tstart=0

http://blogs.sun.com/brendan/entry/jbod_analytics_example

What if he was a Mac OS X pundit, and asked this question of Mac OS X? "Does Apple innovate too much to be competitive in the desktop market?" What people really want is System 7, all this innovation and UNIX underpinning are just developers writing for developers. Why do we need an object oriented runtime library, and a constantly changing AP that supports concurrent processingI? Most Mac users were happy with one mouse button, why do they need their track pads to sense multiple points? Changes like this just confuse the user and make them learn new ways to do old tasks.

Clearly he has a point. It just isn't a very good one. The real problem with the linux desktop has been INSUFFICIENT innovation. And I don't mean replacing X. I mean designing software that makes computing ubiquitous, transparent, and accessible. Why I as a user should ever be concerned about files, drives, network connections, applications, processes, etc. you know all those metaphors programmers have invented for themselves, is beyond me.

What the linux desktop sucks at (and this is true for all WIMPy interfaces) is cross-task operations. When I'm working on a project that involves, text, drawings, tables, and some computations, using a system like Linux, Windows, or Mac OS X is an exercise in frustration. Each "task" as defined by the system's designers requires a different set of tools. By generating a report is only one task from my point of view as a user. As a result, I will end up using 4 or more tools, because no one tool has my work flow in mind.

There are interfaces that solve this problem, however, and they've been around since the late 1970s and early 1980s. Some of us still use them today. Now there is an experimental implementation of one for linux here you can run them in your web browser here and you can package up you can roll your own tools with this here. But all of these are still from the user's point of view in their infancy.

... it depends on what you're trying to do. Just today, Sun published a Sun BluePrint Solid State Drives in HPC: Reducing the I/O Bottleneck showing that by using SSD as part of an HPC configuration, you can improve performance pretty dramatically. If buying 64GB of SSD can double performance on an I/O bound HPC program (these were finite element analysis programs) it might be a pretty good investment.

You can have your cake and eat it in 7-STABLE with ZFS v13; it has functional L2ARC support.

Or go for both and use the SSD as cache.

Well said, random reads are an order of magnitude faster. What's interesting is to see the article neglect this mark, when the desktop/laptop computing experience relies most on this exact variable.

I don't expect enterprise data centers to be using SSD to host my flikr photos any time soon (outside of a few specialized workloads such as database write cache), but the laptop and the solid state disk are a match made in heaven.

http://mediacast.sun.com/share/ChrisM/BeAGameDesigner.mpg

So now they have an unprofitable hardware division that also happens to be in a pond that is getting smaller every day due to Intel.

Sun makes commodity computers too. They haven't marketed them very effectively because the sales organization is still in a Sparc-uber-alles mindset. Despite this, some of Sun's biggest profit centers are in the x64 arena.

Some of these products are pretty cool. This beast squeezes 8 Intel-compatible CPUs into a 4 rack-unit space, something nobody else can do. This one does the same with 48 terabytes of disk. This blade module squeezes four Nehalem processors into an absurdly small space and and been generating tons of orders from people looking to build huge HPC clusters. And of course there are the usual 1U and 2U systems.

Oracle is saying they can squeeze $1.5 billion in profit out of Sun in the first year. Assuming that this isn't BS, the only way they can do this is by drastically boosting hardware sales. Which is actually doable, with better marketing (including more emphasis on systems people actually want to buy) and sales, and access to Oracle's huge sales channels. (Oracle has more sales people than Sun has employees.) In any case, that scenario makes a lot more sense than the common assumption that they're dropping $5 billion just to acquire some not very profitable software assets.

So now they have an unprofitable hardware division that also happens to be in a pond that is getting smaller every day due to Intel.

Sun makes commodity computers too. They haven't marketed them very effectively because the sales organization is still in a Sparc-uber-alles mindset. Despite this, some of Sun's biggest profit centers are in the x64 arena.

Some of these products are pretty cool. This beast squeezes 8 Intel-compatible CPUs into a 4 rack-unit space, something nobody else can do. This one does the same with 48 terabytes of disk. This blade module squeezes four Nehalem processors into an absurdly small space and and been generating tons of orders from people looking to build huge HPC clusters. And of course there are the usual 1U and 2U systems.

Oracle is saying they can squeeze $1.5 billion in profit out of Sun in the first year. Assuming that this isn't BS, the only way they can do this is by drastically boosting hardware sales. Which is actually doable, with better marketing (including more emphasis on systems people actually want to buy) and sales, and access to Oracle's huge sales channels. (Oracle has more sales people than Sun has employees.) In any case, that scenario makes a lot more sense than the common assumption that they're dropping $5 billion just to acquire some not very profitable software assets.

So now they have an unprofitable hardware division that also happens to be in a pond that is getting smaller every day due to Intel.

Sun makes commodity computers too. They haven't marketed them very effectively because the sales organization is still in a Sparc-uber-alles mindset. Despite this, some of Sun's biggest profit centers are in the x64 arena.

Some of these products are pretty cool. This beast squeezes 8 Intel-compatible CPUs into a 4 rack-unit space, something nobody else can do. This one does the same with 48 terabytes of disk. This blade module squeezes four Nehalem processors into an absurdly small space and and been generating tons of orders from people looking to build huge HPC clusters. And of course there are the usual 1U and 2U systems.

Oracle is saying they can squeeze $1.5 billion in profit out of Sun in the first year. Assuming that this isn't BS, the only way they can do this is by drastically boosting hardware sales. Which is actually doable, with better marketing (including more emphasis on systems people actually want to buy) and sales, and access to Oracle's huge sales channels. (Oracle has more sales people than Sun has employees.) In any case, that scenario makes a lot more sense than the common assumption that they're dropping $5 billion just to acquire some not very profitable software assets.

So now they have an unprofitable hardware division that also happens to be in a pond that is getting smaller every day due to Intel.

Sun makes commodity computers too. They haven't marketed them very effectively because the sales organization is still in a Sparc-uber-alles mindset. Despite this, some of Sun's biggest profit centers are in the x64 arena.

Some of these products are pretty cool. This beast squeezes 8 Intel-compatible CPUs into a 4 rack-unit space, something nobody else can do. This one does the same with 48 terabytes of disk. This blade module squeezes four Nehalem processors into an absurdly small space and and been generating tons of orders from people looking to build huge HPC clusters. And of course there are the usual 1U and 2U systems.

Oracle is saying they can squeeze $1.5 billion in profit out of Sun in the first year. Assuming that this isn't BS, the only way they can do this is by drastically boosting hardware sales. Which is actually doable, with better marketing (including more emphasis on systems people actually want to buy) and sales, and access to Oracle's huge sales channels. (Oracle has more sales people than Sun has employees.) In any case, that scenario makes a lot more sense than the common assumption that they're dropping $5 billion just to acquire some not very profitable software assets.

So now they have an unprofitable hardware division that also happens to be in a pond that is getting smaller every day due to Intel.

Sun makes commodity computers too. They haven't marketed them very effectively because the sales organization is still in a Sparc-uber-alles mindset. Despite this, some of Sun's biggest profit centers are in the x64 arena.

Some of these products are pretty cool. This beast squeezes 8 Intel-compatible CPUs into a 4 rack-unit space, something nobody else can do. This one does the same with 48 terabytes of disk. This blade module squeezes four Nehalem processors into an absurdly small space and and been generating tons of orders from people looking to build huge HPC clusters. And of course there are the usual 1U and 2U systems.

Oracle is saying they can squeeze $1.5 billion in profit out of Sun in the first year. Assuming that this isn't BS, the only way they can do this is by drastically boosting hardware sales. Which is actually doable, with better marketing (including more emphasis on systems people actually want to buy) and sales, and access to Oracle's huge sales channels. (Oracle has more sales people than Sun has employees.) In any case, that scenario makes a lot more sense than the common assumption that they're dropping $5 billion just to acquire some not very profitable software assets.

So now they have an unprofitable hardware division that also happens to be in a pond that is getting smaller every day due to Intel.

Sun makes commodity computers too. They haven't marketed them very effectively because the sales organization is still in a Sparc-uber-alles mindset. Despite this, some of Sun's biggest profit centers are in the x64 arena.

Some of these products are pretty cool. This beast squeezes 8 Intel-compatible CPUs into a 4 rack-unit space, something nobody else can do. This one does the same with 48 terabytes of disk. This blade module squeezes four Nehalem processors into an absurdly small space and and been generating tons of orders from people looking to build huge HPC clusters. And of course there are the usual 1U and 2U systems.

Oracle is saying they can squeeze $1.5 billion in profit out of Sun in the first year. Assuming that this isn't BS, the only way they can do this is by drastically boosting hardware sales. Which is actually doable, with better marketing (including more emphasis on systems people actually want to buy) and sales, and access to Oracle's huge sales channels. (Oracle has more sales people than Sun has employees.) In any case, that scenario makes a lot more sense than the common assumption that they're dropping $5 billion just to acquire some not very profitable software assets.

Gosling says::

Lots of folks ask "why doesn't sun just do the JDK for Mac?". The real answer is "because Apple wanted to do it". They've wanted to do all sorts of customization and integration that only they could do - because they own the OS.

In fiscal year 2008, Sun sold 4,532 $ millions in SPARC servers, and only 707 millions in x64 servers (source).
I don’t think it would have been wise for them to kill their biggest-selling product.

Since .NET is an entire platform "just like Java", let's compare their bloat.
The latest version of the Java runtime is 15.50 MB big. And it will run apps written for *any* version of Java.
The latest version of the .NET framework is 231 MB big. And it will only run apps written for that version, requiring the "side by side" installation of other runtimes for other versions.

In order to really get an idea of how fast (or slow) Java is, I tried the following on my outdated machine, a 1600 MHz AMD Turion. First, I tried to measure the dreaded JVM cold startup time by running Apache Rhino:

peppe@tikal:~$ time java -jar /usr/share/java/js-1.7R2.jar -help
[...]
real 0m1.444s
user 0m0.232s
sys 0m0.084s

Then I’ve done it a second time to see what the delay becomes on a hot start:

peppe@tikal:~$ time java -jar /usr/share/java/js-1.7R2.jar -help
[...]
real 0m0.358s
user 0m0.252s
sys 0m0.036s

...that’s a little more than a third of a second. So there is a one-time delay to pay if you write your app in java, but it’s not so user-noticeable as many believe.

So far for the startup time. Then I tested the run time.
Probably the JVM cannot compete with native code as a byte-pusher. But how well does it fare when it comes to support high-level languages? I wrote two small nonsense programs, that aim to exercise some random high-level functions I expect to be common in today’s software. The two programs try to do exactly the same thing, the first in C, and the latter Java.
Link to the C source
Link to the Java source
(sorry, slashdot didn’t let me put the snippets inline without making them unreadable)

I compiled and timed the C implementation:

peppe@tikal:~$ gcc -v
Using built-in specs.
Target: x86_64-linux-gnu
[...]
gcc version 4.4.0 (Ubuntu 4.4.0-6ubuntu2)
peppe@tikal:~$ gcc -O3 -march=athlon64 cperf.c -Wall -o cperf.exe
cperf.c: In function `main':
cperf.c:18: warning: ignoring return value of `asprintf', declared with attribute warn_unused_result
peppe@tikal:~$ time ./cperf.exe

real 0m43.429s
user 0m43.351s
sys 0m0.028s

...then I did the same with the Java version:

peppe@tikal:~$ javac -version
javac 1.6.0_14
peppe@tikal:~$ javac JavaPerf.java
peppe@tikal:~$ time java JavaPerf

real 0m28.300s
user 0m27.770s
sys 0m0.464s

Not only Java performance was comparable to native, but Java was even faster in this case. And that included the JVM startup penalty.

Finally, about the JVM size versus the .NET framework size, the latest win32 version of the JRE weighs 15.50 MB (link), while the latest win32 version of the .NET framework weighs 231 MB (link).

And Java has done that for years, FileOpenService, FileSaveService and PersistenceService, JNLP (Web Start) one of the most useful API on Java that nearly everyones ignores

# # An unexpected error has been detected by HotSpot Virtual Machine: # # SIGSEGV (0xb) at pc=0xb77acafa, pid=21855, tid=1833073568 # # Java VM: Java HotSpot(TM) Server VM (1.5.0_14-b03 mixed mode) # Problematic frame: # V [libjvm.so+0x23dafa] # # An error report file with more information is saved as hs_err_pid21855.log # # If you would like to submit a bug report, please visit: # http://java.sun.com/webapps/bugreport/crash.jsp # Abort

Citation needed.

Even if not phython, what does Fortran have over modern compiled languages, for example?

The main thing that Fortran (and Ada too) has is a standard floating point model. There are some minor differences between compilers. This is unlike C, Java, Python which rely on the underlying FP hardware. Then to look at the loopholes in the IEEE FP standard for what NaN means. For citations, look at the Numerical Recipes series of books for a discussion of the differences. Also check out "What every computer scientist should know about floating point"

http://docs.sun.com/source/806-3568/ncg_goldberg.html

IBM has released a set of libraries for C and Java that provide the Fortran semantics of floating point, and there is a wrapper for them that uses operator overloading in C++ to make them easier to use. But they are still a pain to use compared to the

The one thing ZFS won't have is hard linked directories, which is needed for time machine.

ZFS out-time-machines Time Machine. Very good snapshots are available by default. OpenSolaris implements something known as Time Slider which provides similar features. Really, Time Machine would have been easier to implement in the first place if they switched to ZFS with Leopard.

I wouldn't use MS' ODF, last time I wanted to export ODF from MS Office, I used the plug in provided by Sun microsystems. I haven't used it lately, but it's up to version 3.1. Last version I used was 1.1.

Sun ODF Plugin

http://www.sun.com/storage/disk_systems/unified_storage/7110/specs.xml

You can even download a simulator.

Anybody that has been out of touch for one year or more with a technology should check his assumptions first before talking....

If Apple can't use the technology I think it is their loss frankly.

it has been vigorously discussed in project coin list. and it most likely to be included in the Java 7. Along with collection literals. there is a javaone presentation about language changes. http://blogs.sun.com/darcy/resource/JavaOne/J1_2009-TS-4060.pdf

Where does it say on the page that ARM is accepted (I've no doubt that it was submitted)? I don't see it on the tentative list of changes here, and it's not in top 3 in the poll.

Last thing I heard on the subject was Mark Reinhold's update on the status of Java 7 dating back to December last year - and it doesn't list ARM. But I must admit that it may well be outdated (since the project you've linked to came later), so what's the current definite status, and where to keep track of it?

a better list: http://blogs.sun.com/darcy/entry/project_coin_consideration_round_2

Java 7 will have something very similar, called the fork/join framework: http://developers.sun.com/learning/javaoneonline/j1sessn.jsp?sessn=TS-5515&yr=2008&track=javase Doug Lea also offers a preliminary package for Java 6 that can be downloaded from here: http://gee.cs.oswego.edu/dl/concurrency-interest/

Well, the keys are right next to each other...

By the way, the research paper describing G1 is here.

Sun didn't "quietly edit" the release notes; they announced it publicly and appologized for having been unclear (which seems like a bit dishonest, but not quiet).

Hehe. We have a large Sun/Storage Tek tape library at my work. The SL300000 http://www.sun.com/storagetek/tape_storage/tape_libraries/sl3000/ . It runs Win2k. The question is what is a new $120k device (~70k but then that is before you get the drives for the library :-)) from an old school UNIX vendor doing running an out of support version of Windows :-) . We also have microscopes that are controlled by windows but the GUI is in Linux (they come with both computers in one case). It all comes down to what the developers were comfortable with at the time, and whether device drivers are available I guess.

The biggest challenge facing Windows is its size and hardware requirements

Well, most people would say its reliability and security issues. But I guess those are also effects of feature bloat, driven by a desire to please lots of diverse customers.

Nobody seems immune from this. Apple people I know blame it for the meltdown in OS 8 development. I work at Sun, and I noticed that our latest crop of Sun Ray thin clients come with RS-232 ports — this at a time when such ports are disappearing from most products, including previous Sun Rays. When I asked about it, I was told that it was to accommodate a big customer deploying point-of-sale systems. (Bar code scanners, card swipers, etc., still haven't migrated to USB.)

http://www.sun.com/servers/coolthreads/t5440/specs.xml

4 UltraSparc T2 processors with 4 processors x 8 cores per processor x 8 threads per core = 256 threars

Press "Get it". Prices start from $91,995.00 with 256 threads and $51,795.00 with 128 threads

From http://www.sun.com/processors/UltraSPARC-T2/features.xml

"Features and Benefits
With eight cores and 64 threads on one chip, integrated 10 GbE networking, crypto, and PCI-Express expansion, you have the jump on anything else on the market. The opportunities for system consolidation and virtualization are here like never before. Consumes less power per core and thread than any processor in its class - without compromising on performance. The UltraSPARC T2 processor gives OEMs a massively threaded, multi-core alternative to more power-hungry, less threaded processors from competing vendors."

Sun SPARC anyone? http://www.sun.com/processors/opensparc/

Static imports?
http://java.sun.com/j2se/1.5.0/docs/guide/language/static-import.html

You're a funny guy but obviously have zero idea about what you are talking about. It's a bit aggravating that ignorant folk still come out with the same old 'slow' mantra and unfortunately even more ignorant folk buy into it. That keeps people writing crappy software on C++ or C when quite often Java would be a good fit for them and the performance is mostly a non-issue these days (unless you write very inefficient programs).

I used to eschew Java for the speed of C++ but now I've completely switched to Java for most development tasks (apart from some C glue-code [JNI] when I need to integrate some scientific instrument or another). I'm doing instrument control, image processing and analysis and I need both reasonable latency, multithreading support, and every performance cycle I can get, and yet Java is plenty fast enough for me (and even embedded devices these days have relatively large amounts of RAM this is far less of an issue than it used to be).

Seriously Java is very fast these days. Graphics2D is all done in Direct3D or OpenGL shaders, the VM is very optimised and quite often approaches FORTRAN (which is faster than C). Don't believe me? check out the links below.

So please, next time stop with the FUD (that used to be marginally true 5 years ago) and start with an open mind.

• From http://en.wikipedia.org/wiki/Java_performance#Use_for_High_Performance_Computing
  "However, High Performance Computing applications written in Java have recently won benchmark competitions. In 2008, Apache Hadoop, an open source High Performance Computing project written in Java was able to sort a terabyte of integers the fastest.[46]"
• INRIA (French scientifc Institution) report on High Performance Computing with Java [Summary: Java approaching FORTRAN for speed, although some network intensive speed bumps still remaining]
  http://hal.inria.fr/inria-00312039/en
• From http://blogs.sun.com/jag/entry/current_state_of_java_for
  Current State of Java for HPC
  
  At the last JavaOne I did a walk-on talk during the AMD keynote where I talked about how incredible HotSpot's performance had become - beating the best C compilers. I ended my talk with a joking comment that "the next target is Fortran". Afterwards, Denis Caromel of Inria came up to me and said "you're already there". He and some colleges had been working on some comparisons between Java and Fortran for HPC. Their final report Current State of Java for HPC has been made available as a Tech Report and makes pretty interesting reading. There are a lot of HPC micro benchmarks in it which look great. Thanks! Permalink Comments [3]

See the Object.finalize() javadoc.

Java has a concept of "new," but it doesn't have a "delete." (Well, the concept exists behind the cutain, but programmers never delete things themselves when using Java.)

finalize() is what the GC called to destroy an object. Every object inherits it from the Object class (duh), so if you want to do some memory management yourself, go ahead and give finalize() a call.

http://www.sun.com/software/javaseforbusiness/faq.jsp#c2q1
Internal/In-house use: The Java SE platform binaries (JDK and JRE) are licensed under Sun's Binary Code License (BCL) with supplemental terms. For most developers and end-users, the binary JDK and binary JRE are all that's needed to experience the world of Java technology. USE: The binary JDK and JRE are available at no fee from Sun (per terms of the BCL) for use with desktop personal computers. JDK or JRE use for embedded devices and other computing environments may require a license fee from Sun.

Anyone with Solaris 10 can run dtrace on any adobe programs and see hundreds of thousands (100.000+) of wasted system calls constantly being made.

I did this with Acrobat Reader a few years ago and saw hundreds of get_current_time() calls every second. This is a read-only application, so why does it care what time it is?

Want to know more about DTrace? http://www.sun.com/bigadmin/content/dtrace/

Seriously, your headline is exactly backwards. Read what the pay for support is for. The G1 is beta so you pay for support... the older versions of Java that are out of date are also pay for support.