Quite A Few Questions


lastaid

Still Fresh
Joined
Aug 14, 2009
Messages
21
Hello everyone.

1#:
I didn't read through the whole board, but i saw quite a reoccuring question:
Q:Can I execute Windows Applications with WINE?

We cannot execute Windows Binaries with WINE, because WINE IS NOT an EMULATOR and they use x86 Arch. We use an ARM RISC.
But Bochs is infact an emulator.
Windows would run slow as hell, especially with the whole linux kernel in background which brings me to my next Topic.

2#: Most of the applications depend on the whole kernel to run in background. I'm studying scientific programming, but the only language i can write in is Java, so I'm most likely to to become homeless and bipolar in the next 4-5 years. But is there a possibility to run bytecode directly like mainstream handhelds do? This would mean that we have infact the strongest handheld out there and without emulation or a linux kernel stealing res we could make the homebrew game experience go beyond anything seen before.

#3: LGPT :-D
Some of the GP32 people might know this nice piece of Software, its a Music Tracker similar to LSDj.
Everyone who wants to use the Pandora for performing Live music might be interested in this. The Interface only uses D-Pad and 2 Buttons + Shoulder Buttons ...
If Someone is out there with advanced knowledge of these Systems - using a Touchscreen could revolutionize Handheld music performance :-D


Ok, that's it for now.
Actually my only wish for the pandora is playing Tyrian, but the things above a quite interesting too.

Any Ideas?

Have a good day,

lastaid
 
I know the author of LGPT pretty well, and I'm sure there will be a port to Pandora in the future .. I offered to port it using the Beagleboard, but I guess he wanted to do it himself when he gets a Pandora ..
 
lastaid said:
Hello everyone.

1#:
I didn't read through the whole board, but i saw quite a reoccuring question:
Q:Can I execute Windows Applications with WINE?

We cannot execute Windows Binaries with WINE, because WINE IS NOT an EMULATOR and they use x86 Arch. We use an ARM RISC.
This has been discussed here: http://www.gp32x.de/board/index.php?/topic/47749-the-pandora-can-not-run-windows-or-wine/
Running Wine can be done, what is unknown is what the performance will be (extremely bad to very bad).
 
Last edited by a moderator:
lastaid said:
2#: Most of the applications depend on the whole kernel to run in background. I'm studying scientific programming, but the only language i can write in is Java, so I'm most likely to to become homeless and bipolar in the next 4-5 years. But is there a possibility to run bytecode directly like mainstream handhelds do? This would mean that we have infact the strongest handheld out there and without emulation or a linux kernel stealing res we could make the homebrew game experience go beyond anything seen before.

1. all linux applications depend on the linux kernel. software that does not depend on the linux kernel is not a linux application ; )
2. cortex-A8 has ARM's JVM-centric extension known as Jazelle. with the right software pandora should fare just as well at running java bytecode as any other jazelle-featuring gadget today (i.e. my cell phone, and likely yours too).

edit: ok, the above needs a correction. it appears cortex-A8 features 'Jazelle RCT', but not 'Jazelle DBX' - the latter being the JVM bytecode extension, and the former - a JIT-centric extension. so pandora may not be capable of running native java bytecode per se.

ps: i don't think java is going anywhere in the near future.
 
Last edited by a moderator:
darkblu said:
2. cortex-A8 has ARM's JVM-centric extension known as Jazelle. with the right software pandora should fare just as well at running java bytecode as any other jazelle-featuring gadget today (i.e. my cell phone, and likely yours too).

edit: ok, i think the above needs a correction. it appears cortex-A8 features 'Jazelle RCT', but not 'Jazelle DBX' - the latter being the JVM bytecode extension, and the former - a JIT-centric extension. so pandora may not be capable or running native java bytecode per se.

ps: i don't think java is going anywhere in the near future.
Seen the recent buzz about "Java" (→ actually JVM) 7? It basically enables you to run Python, Ruby, Scala, etc, on any machine that has a JVM, and usually at about 1.5x the speed compared to native (Python and Ruby) VMs or interpreters.

Imagine writing a game in your favourite language, and it'll run extremely quickly on all platforms known to man, and you can call Ruby code directly from Python code, and debug the whole thing with Scala code using a C# framework. Forget using Python "accelerated" C libraries; you won't have the need to. The Java platform is indeed going somehwere :p

However, Java is a dying platform (and I think that that's what you were talking about) because some wrong design choices were made when writing it, and we're now suffering from them (high memory usage, long loading times, bad interaction with native libs, etc). Heck, the fact that a technology like Jazelle exists is proof of that... If there was some alternative VM implementation, I would totally use it, but alas, there is none. (CLR on .NET? Are you kidding me? It's basically a Java copy with the same faults and even more bloat...)

Oh, and It's a pity that Jazelle is closed-source; we would really benefit from the extra performance.
 
Last edited by a moderator:
dflemstr said:
Seen the recent buzz about "Java" (→ actually JVM) 7? It basically enables you to run Python, Ruby, Scala, etc, on any machine that has a JVM, and usually at about 1.5x the speed compared to native (Python and Ruby) VMs or interpreters.

<snip>
sorry about the confusion caused. i meant that java is not leaving us in the foreseeable future ; )

while my view of the language is fairly remote, i'm well aware that it is a viable tool. and i know a few 'leet coders' who swear by it (as in: all conditions equal, they'd pick it over other, more 'established' languages).

as about a few hapless design decisions in the language, if that was a reason for programming languages to perish, C++ should've been dead, buried and anathema'd by the pope for a long time now ; )
 
Last edited by a moderator:
Eh have you forgotten LLVM it can be a jit compiler iirc... from what i remember it can interpret compile to bytecode or compile to native code
 
cb88 said:
Eh have you forgotten LLVM it can be a jit compiler iirc... from what i remember it can interpret compile to bytecode or compile to native code
Well, LLVM promises a lot but it doesn't seem to be popular enough to make me want to use it. Difficulties to find a *good* compiler for anything except C/C++ is an issue - not a big one but still an issue. Also, you don't really find precompiled libraries for LLVM and development is just generally difficult to get started with.

Something that would be great if you could compile JVM bytecode to LLVM byte code and convert libraries that way; would help LLVM gain popularity by a fair bit.

btw maybe we should get back on topic before it's too late ;)
 
Last edited by a moderator:
Thanks for the reply, that ended up in a java discussion :D [ somehow always happens when cpp and java people meet ]
I'm not sure if I got that part about using wine right, but now it seems like someone wants to use x86 emulation [ running on a linux kernel ] to run linux with wine to use windows apps. Now i have this image in my head of a scooter for fat people towing a firetruck uphill. Would be an entertaining prove of concept that you can physicaly connect a firetruck with a scooter but you canNOT expect performance.

About that LGPT stuff, I hope that the OP got a nice audio output without humming etc :D

Have a good night ;-)

lastaid
 
dflemstr said:
Something that would be great if you could compile JVM bytecode to LLVM byte code and convert libraries that way; would help LLVM gain popularity by a fair bit.

btw maybe we should get back on topic before it's too late ;)

VMKit for LLVM should do the job. (There, said my piece.)
 
Last edited by a moderator:
Samurai_Crow said:
dflemstr said:
Something that would be great if you could compile JVM bytecode to LLVM byte code and convert libraries that way; would help LLVM gain popularity by a fair bit.

btw maybe we should get back on topic before it's too late ;)

VMKit for LLVM should do the job. (There, said my piece.)


Thanks alot! I'm currently working on a big Project using, of course, Java :D, and until i find an alternative for all this polling, this should pump up the performance.
By the way: Anyone got an idea how to avoid these nasty while(true) loops for checking if a Queue or ioStream has new Items?

Have a good day

lastaid
 
Last edited by a moderator:
lastaid said:
Thanks alot! I'm currently working on a big Project using, of course, Java :D , and until i find an alternative for all this polling, this should pump up the performance.
By the way: Anyone got an idea how to avoid these nasty while(true) loops for checking if a Queue or ioStream has new Items?

Have a good day

lastaid
Some example code would be great.
 
Last edited by a moderator:
lastaid said:
By the way: Anyone got an idea how to avoid these nasty while(true) loops for checking if a Queue or ioStream has new Items?
Doesn't Java provide access to POSIX select?
 
Last edited by a moderator:
Laurent said:
lastaid said:
By the way: Anyone got an idea how to avoid these nasty while(true) loops for checking if a Queue or ioStream has new Items?
Doesn't Java provide access to POSIX select?
Why would it? Java needs to run on Windows too, and Windows doesn't implement POSIX at all as you know.
 
Last edited by a moderator:
dflemstr said:
Why would it? Java needs to run on Windows too, and Windows doesn't implement POSIX at all as you know.
Oh really?
http://msdn.microsoft.com/en-us/library/ms740141%28VS.85%29.aspx
http://en.wikipedia.org/wiki/Microsoft_Windows_Services_for_UNIX
http://java.sun.com/j2se/1.4.2/docs/api/java/nio/channels/Selector.html
 
Last edited by a moderator:
Laurent said:
dflemstr said:
Why would it? Java needs to run on Windows too, and Windows doesn't implement POSIX at all as you know.
Oh really?
http://msdn.microsoft.com/en-us/library/ms740141%28VS.85%29.aspx
http://en.wikipedia.org/wiki/Microsoft_Windows_Services_for_UNIX
http://java.sun.com/j2se/1.4.2/docs/api/java/nio/channels/Selector.html
Well, that's news to me then.
 
Last edited by a moderator:
dflemstr said:
Well, that's news to me then.
Honestly to me too, it's not like if I knew anything about Java or Windows :)
I was just suggesting select as it's the way I used to do things.
 
Last edited by a moderator:
I've heard of the POSIX layer in Windows for a long time and I don't think I've ever known anyone to actually use it for anything. I think something must not be all together about it or we wouldn't have things like Cygwin anymore.
 
Exophase said:
I've heard of the POSIX layer in Windows for a long time and I don't think I've ever known anyone to actually use it for anything. I think something must not be all together about it or we wouldn't have things like Cygwin anymore.
while posix has been present for ages under NT (and people do use it occasionally - one could hardly ignore it given that 99% of the planet's tcp software in written using it), select() in particular is somewhat deprecated in favor of completion ports, which is a native technique of a great deal higher efficiency, at least for SMP systems under windows.
 
Last edited by a moderator:
Back
Top