Login
Username:

Password:

Remember me



Lost Password?

Register now!
Main Menu
Who's Online
1 user(s) are online (1 user(s) are browsing Forum)

Members: 0
Guests: 1

more...
   All Posts (michalsc)


« 1 (2) 3 4 5 ... 39 »


Re: iAROS just for Core i-CPUs

Joined:
2004/3/30 23:19
From Clausthal-Zellerfeld, Germany
Group:
Member
Posts: 391
Offline
Quote:

RC_tech wrote:
Sorry, I'm strictly against paying. If the programmers of every driver for Linux,


most of them either do that for
a) the hardware they own and want it to be supported, or
b) they are working and payed by the hardware vendor

The case (a) is exactly the same as in AROS. Developer has given HW and want it supported, then developers writes a driver, if he can of course. Regarding (b) - guess why so many drivers in linux tree contain not only the GPL notice, but also copyright note from Intel, NVidia, Freescale and others?

Quote:

OSx86,


Sure, there are drivers written by third party developers (and many thanks there!), but the majority of work done is fiddling with .plist files and trying to add the correct ProductID/VendorID and checking whether the driver provided by Apple is working or not. And guess what, the people who write the real drivers are doing that for the hardware they own, same as in case of AROS developers.

Quote:

the x86-portation of Android


which is relative simple, since linux kernel is full of already working hardware drivers

Quote:

or whatever had only started after they got enough money, we wouldn't have a lot less alternative operationg systems.


Development of every alternative operating system focuses on the hardware owned by the developers. If you know any alternative OS, where developers buy and write drivers for hardware requested by the community, for free of course, name it.

Quote:

I'm comparing again with the OSx86 scene.


You should not. They were re-using 99.99% of the drivers written and provided by Apple already. Do you think we have tons of drivers here, on AROS, which we need only to fix or adapt to make them working?

Quote:

Do you really think AROS is getting further that way?


A lot of bugs were found and fixed only because AROS is working in so many variants and on so many platforms. Some bugs were visible only on certain HW native, or hosted. Also performance of AROS internals was drastically improved in many places only because at some point we started to support some slower hardware. Thanks to the way AROS is written, porting it to completely new architecture, like 64-bit x86 or ARM many years ago was made by one person within few months. Is it really so bad?

Posted on: 2013/7/24 11:59
_________________
Click to see original Image in a new window
Transfer the post to other applications Transfer


Re: iAROS just for Core i-CPUs

Joined:
2004/3/30 23:19
From Clausthal-Zellerfeld, Germany
Group:
Member
Posts: 391
Offline
Quote:

what do you thing to reduce a new generation of AROS only to Intel Core i3, i5 and i7 CPUs of the last generations, Sandy Bridge, Ivy Bridge and Haswell? Reducing the programming of AROS to a few chipsets would have lot of adavatages:


I guess the low-level developers would need to buy such hardware then?

Quote:

- Core iX-systems would allow a 64bit-version as default with 32bit-support. Multicore support also could be possible later.


64bit version of AROS is available for years. The 32-bit support is not available in 64 bit version of AROS and it's not a matter of CPU used. Multicore will be worked on some day. As for now, all CPU cores are woken up from the initial suspend state and enter endless sleep loop again. Can be woken up anytime by sending an IPI interrupt though.

Quote:

- For programming, special command sets like SSSE3 could be used for more efficient code.


You can use SSE3 in AROS already. If you want to write a software which works without SSE3 too, provide either two binaries, or, even better, provide your software with two sets of functions (using SSE3 and using something else instead) and decide which to use in running software already.

Quote:

- there are a lot of Core iX Mini-ITX-Boards and laptops, so you have small or portable iAROS systems with a lot of power.


A lot of Mini-ITX boards and laptops means, there is a lot of different hardware configurations. And I just thought you were talking about reducing the hardware base?

I read you are using OSx86. So you're surely familiar with things like buyer's guides containing hints and advices which HW to use and how well is given hardware supported by OSX. Guess what, we have the same. Look at this page: http://en.wikibooks.org/wiki/Aros/Platforms/x86_support

Why do you accept the fact, that you are forced to buy new hardware to be able to install OSX there, and at the same time you refuse AROS to offer you the same?

Yes, AROS is open source project and all developers are working on it in their free time, just for fun. Therefore, it's rather unpleasant to hear, what shall we do with our free time. You don't like the situation, then try to influence on AROS development in some other way. Telling us what we should and what we should not is surely not the right way, though.

Posted on: 2013/7/24 8:19
_________________
Click to see original Image in a new window
Transfer the post to other applications Transfer


Re: iAROS just for Core i-CPUs

Joined:
2004/3/30 23:19
From Clausthal-Zellerfeld, Germany
Group:
Member
Posts: 391
Offline
Quote:

RC_tech wrote:
I heard sometimes that's a problem that AROS is open-source, but that's not the problem. The problem is, that there's no directed development, like erbsenbirne said. Actually we also have three distributions, all Intel 32-bit. Linux has more than 100 distros, but they lowered down, a lot are using Ubuntu today.


So you say if AROS had only one distro, we would get better hardware support? Do you think that people who put AROS distros together, are good low-level developers capable of writing hardware drivers?

Or do I miss something?


Posted on: 2013/7/24 8:04
_________________
Click to see original Image in a new window
Transfer the post to other applications Transfer


Re: Thread.library gets stuck

Joined:
2004/3/30 23:19
From Clausthal-Zellerfeld, Germany
Group:
Member
Posts: 391
Offline
Quote:

serk118uk wrote:

Hello ppl

I am currently Testing (Thread.Library) with small example trying to run aVorbisPlayer as thread using thread library but plays the 1st filled buffer than wont receive any more and gets stuck at above code but if i run the code without thread than all fine and plays entire ogg..

Please download and take a look and see if i am doing wrong or aros related problem.


Quick reply without compiling the code. You've made one huge mistake. You call ahi_allocate_resources from the main thread. Hence, the MsgPort which you create there is assigned to the main thread too (mp_SigTask). Even worse, the CreateMsgPort function allocates a signal (mp_SigBit) for the main thread too. Therefore, if you call Wait(SIGBREAKF_CTRL_C | (1L << ahi_msg_port->mp_SigBit)) in your player thread, it will wait for SIGBREAKF_CTRL_C only.

Solution: create MsgPort and IORequests in the player thread. Actually you could call the whole ahi_allocate_resources from the player thread :)


PS. It is a very VERY VERY bad habit to put the code into .h file (output_ahi.h). Fix it please!

Posted on: 2013/7/9 2:44
_________________
Click to see original Image in a new window
Transfer the post to other applications Transfer


Re: How useable is AROS on a slow system like 68K, PPC, ARM or old x86?

Joined:
2004/3/30 23:19
From Clausthal-Zellerfeld, Germany
Group:
Member
Posts: 391
Offline
Quote:

biggun wrote:

I think a developer or designer can easily or accidentally create a theme or some code which runs very swift on the high end system but totally overloads a low end system - without noticing this.


I have had that already. When I was porting AROS to EfikaPPC I have found few places which slowed the whole system significantly. But apart from that EfikaPPC with AROS and Radeon driver (not VESA) was quite snappy and fast.

Soon I should be able to tell something about the performance of EfikaMX running native AROS.

Quote:

Personally I'm very curious to know if AROS could be the ideal Operating system for a "semi legacy system" that has 1 GB of memory and a performance comparable to the EFIKA PPC.

This is of course a slow system compared to todays x86 - but out of nostalgic reasons I would like to use the system very much.


EfikaPPC was fast with AROS, as already said. One could have notice the slowdown though when raw cpu power or fast HD IO performance were needed. The hello world in C++ took nearly a minute to compile, mostly due to a huge IO load. You know, c++ and std namespace...

PS. With 1.6GHz Atom system and my video driver for Intel GMA I was able to watch 720p movies. However, it was a bit too slow for 1080p (most likely due to lack of overlay)

Posted on: 2013/6/4 0:38
_________________
Click to see original Image in a new window
Transfer the post to other applications Transfer


Re: ppc efika

Joined:
2004/3/30 23:19
From Clausthal-Zellerfeld, Germany
Group:
Member
Posts: 391
Offline
Quote:

mt12345 wrote:
so I copied the missing handlers from nightly x86.
no errors this time but still hangs at


Erm... Do I understand right, you have copied x86 binaries to ppc AROS and expected it to work?

PS. The nightly build for EfikaPPC you guys are testing was an automated build on Markus Weiss machine. I'm not sure he was even able to test it...

Posted on: 2013/5/29 5:38
_________________
Click to see original Image in a new window
Transfer the post to other applications Transfer


Re: ppc OS X hosted

Joined:
2004/3/30 23:19
From Clausthal-Zellerfeld, Germany
Group:
Member
Posts: 391
Offline
Quote:

ausPPC wrote:
The one I've got is only about a week old but I take your point.

Who put the OS X release out?


These are not releases, but rather nightly builds -- automated builds done from the current revision of source tree. They are not checked by any humans after compilation is done, but rather uploaded automatically to the sourceforge.

Quote:

Or are these things just the product of a sophisticated build environment that isn't supposed to need individual support for specific platforms?


As I wrote the nightly builds are not controlled by anyone. Even worse - supporting e.g. OSX PPC hosted is relatively hard, since we do not have the Rosetta PPC emulator since OSX 10.7... I personally do not own any PPC mac anymore and cannot test the builds.

Posted on: 2013/5/29 0:29
_________________
Click to see original Image in a new window
Transfer the post to other applications Transfer


Re: Amiga 1200 + ACA 1232 aros m68k

Joined:
2004/3/30 23:19
From Clausthal-Zellerfeld, Germany
Group:
Member
Posts: 391
Offline
Quote:

twilen wrote:
Arosbootstrap probably got broken due to recent memory subsystem changes.



Can't be, unless you give MEMF_MANAGED flag to the krnCreateMemHeader function... But I'm sure you don't.

Posted on: 2013/5/22 10:00
_________________
Click to see original Image in a new window
Transfer the post to other applications Transfer


Re: Performance improvements

Joined:
2004/3/30 23:19
From Clausthal-Zellerfeld, Germany
Group:
Member
Posts: 391
Offline
Let me tease you a little. Test machine - virtualbox on MacBook Air running AROS hosted. I have executed each test several times. The time mentioned here is a mean value.

Default AROS hosted kernel:
test0 = 8.200 (clib's malloc/free)
test1 = 8.925 (exec's allocvec/freevec)

My tweaked AROS hosted kernel (with MEMF_MANAGED memory and completely different memory allocator):
test0 = 7.352 (clib's malloc/free)
test1 = 0.127 (exec's allocvec/freevec)

nice, isn't it? ;)

Posted on: 2013/5/9 14:42
_________________
Click to see original Image in a new window
Transfer the post to other applications Transfer


Re: Genesi: Added an AROS link

Joined:
2004/3/30 23:19
From Clausthal-Zellerfeld, Germany
Group:
Member
Posts: 391
Offline
Updated build with scout and few other things. Still, some software will not run or even will crash AROS (gcc is an example), so please test and report :)

Link: https://docs.google.com/file/d/0B4cwc1 ... 2RXFvTU0/edit?usp=sharing

PS. In case you have issues with keyboard, go to SYS:Devs/Keymaps/X11 and delete the keycode2rawkey.table file.

Posted on: 2013/3/27 4:10
_________________
Click to see original Image in a new window
Transfer the post to other applications Transfer



 Top
« 1 (2) 3 4 5 ... 39 »




Search
Top Posters
1 paolone
paolone
3838
2 nikolaos
nikolaos
3474
3 magorium
magorium
3357
4 phoenixkonsole
phoenixkonsole
3275
5 deadwood
deadwood
2419
6 ncafferkey
ncafferkey
2215
7 mazze
mazze
2097
8 clusteruk
clusteruk
2051
9 Kalamatee
Kalamatee
1808
10 damocles
damocles
1775
© 2004-2014 AROS Exec