Windows on ARM, how fast it can be implemented?

classic Classic list List threaded Threaded
21 messages Options
12
Reply | Threaded
Open this post in threaded view
|

Windows on ARM, how fast it can be implemented?

Max Vlasov
If this news is right
  http://www.bloomberg.com/news/2010-12-21/microsoft-is-said-to-announce-version-of-windows-for-arm-chips-at-ces-show.html
There will be ARM Win32 in January. Looking at how many platforms fpc supports currently, there should be no big deal to implement this, right? Or do I miss something and there are hidden challenges?

Max Vlasov


_______________________________________________
fpc-pascal maillist  -  [hidden email]
http://lists.freepascal.org/mailman/listinfo/fpc-pascal
Reply | Threaded
Open this post in threaded view
|

Re: Windows on ARM, how fast it can be implemented?

Sven Barth-2
On 22.12.2010 20:05, Max Vlasov wrote:
> If this news is right
> http://www.bloomberg.com/news/2010-12-21/microsoft-is-said-to-announce-version-of-windows-for-arm-chips-at-ces-show.html
> There will be ARM Win32 in January. Looking at how many platforms fpc
> supports currently, there should be no big deal to implement this,
> right? Or do I miss something and there are hidden challenges?

What the?! O.o

Ok... even if they would indeed be doing that, FPC has all needed things
in place basically. With WinCE we have the necessary ARM backend support
and the normal Win32 API might work without much changes if Microsoft
keeps the APIs compatible (which they should do if they want ANY popular
software running at least by recompiling).

We will see how that evolves.

Regards,
Sven

_______________________________________________
fpc-pascal maillist  -  [hidden email]
http://lists.freepascal.org/mailman/listinfo/fpc-pascal
Reply | Threaded
Open this post in threaded view
|

Re: Windows on ARM, how fast it can be implemented?

Max Vlasov


On Wed, Dec 22, 2010 at 10:13 PM, Sven Barth <[hidden email]> wrote:
On 22.12.2010 20:05, Max Vlasov wrote:
If this news is right...
There will be ARM Win32 in January.

What the?! O.o

Ok... even if they would indeed be doing that, FPC has all needed things in place basically. With WinCE we have the necessary ARM backend support and the normal Win32 API might work without much changes if Microsoft keeps the APIs compatible ...


I was  surprised also. If it's Windows CE with more marketing effort, I think it will be the end of surprise. If it's desktop Win32 with slogan "you only need to recompile", it's an interesting story. I doubt Embarcadero will be ready fast enough unless they're already working under nda. But the latter is still possible since for MS encouraging everybody and not having skype ready to run on the new platform will look a little bit strange. But it's just my speculation

Max

_______________________________________________
fpc-pascal maillist  -  [hidden email]
http://lists.freepascal.org/mailman/listinfo/fpc-pascal
Reply | Threaded
Open this post in threaded view
|

Re: Windows on ARM, how fast it can be implemented?

Florian Klämpfl
In reply to this post by Max Vlasov
Am 22.12.2010 20:05, schrieb Max Vlasov:
> If this news is right
>  
> http://www.bloomberg.com/news/2010-12-21/microsoft-is-said-to-announce-version-of-windows-for-arm-chips-at-ces-show.html
> There will be ARM Win32 in January. Looking at how many platforms fpc
> supports currently, there should be no big deal to implement this,
> right?

No. Maybe MS uses different calling conventions or a different FPU
implementation but even those are no big show stoppers.
_______________________________________________
fpc-pascal maillist  -  [hidden email]
http://lists.freepascal.org/mailman/listinfo/fpc-pascal
Reply | Threaded
Open this post in threaded view
|

Re: Windows on ARM, how fast it can be implemented?

Max Vlasov
Hi,

It was long time ago this post was sent.
A platform appeared closed by design
But since RT platform was jailbroken (http://forum.xda-developers.com/showthread.php?t=2092158), I'm curious whether compiling into ARM Exe would work out of the box or will require some change on the compiler side?

Thanks

Max

On Thu, Dec 23, 2010 at 11:57 AM, Florian Klämpfl <[hidden email]> wrote:
Am 22.12.2010 20:05, schrieb Max Vlasov:
> If this news is right
>
> http://www.bloomberg.com/news/2010-12-21/microsoft-is-said-to-announce-version-of-windows-for-arm-chips-at-ces-show.html
> There will be ARM Win32 in January. Looking at how many platforms fpc
> supports currently, there should be no big deal to implement this,
> right?

No. Maybe MS uses different calling conventions or a different FPU
implementation but even those are no big show stoppers.
_______________________________________________
fpc-pascal maillist  -  [hidden email]
http://lists.freepascal.org/mailman/listinfo/fpc-pascal


_______________________________________________
fpc-pascal maillist  -  [hidden email]
http://lists.freepascal.org/mailman/listinfo/fpc-pascal
Reply | Threaded
Open this post in threaded view
|

Re: Windows on ARM, how fast it can be implemented?

Sven Barth-2

Am 26.03.2013 15:01 schrieb "Max Vlasov" <[hidden email]>:
>
> Hi,
>
> It was long time ago this post was sent.
> A platform appeared closed by design
> But since RT platform was jailbroken (http://forum.xda-developers.com/showthread.php?t=2092158), I'm curious whether compiling into ARM Exe would work out of the box or will require some change on the compiler side?

Disclaimer: Microsoft had responded that they don't guarantee that the jailbreak will continue to work in future builds (and they congratulated the developer that discovered the hack ^^)

That said it mostly depends on how much Microsoft changed their existing ARM compiler. If the changes are minimal for the purpose of user space applications then it should be rather easy to adjust the Win32 target for this. The biggest problem might be unaligned accesses in the Win32 RTL code, but AFAIK newer ARMs can handle these...

If you own a Windows RT device and are willing to jailbreak it, I could try to hack arm-win32 support together for you to test, but finegrained bug fixing (if the port should indeed work) needs to be done by someone owning the device.

Regards,
Sven


_______________________________________________
fpc-pascal maillist  -  [hidden email]
http://lists.freepascal.org/mailman/listinfo/fpc-pascal
Reply | Threaded
Open this post in threaded view
|

Re: Windows on ARM, how fast it can be implemented?

Max Vlasov
Thanks for the info, Sven. I will probably have a chance to get a device and jailbreak it, although not right now. If someone on the list does this before me, I'll be glad. I did not find any prior discussion about this topic, seems like it's not that hot to be discussed and everything.

Max

On Tue, Mar 26, 2013 at 8:05 PM, Sven Barth <[hidden email]> wrote:

Am 26.03.2013 15:01 schrieb "Max Vlasov" <[hidden email]>:


That said it mostly depends on how much Microsoft changed their existing ARM compiler. If the changes are minimal for the purpose of user space applications then it should be rather easy to adjust the Win32 target for this. The biggest problem might be unaligned accesses in the Win32 RTL code, but AFAIK newer ARMs can handle these...

If you own a Windows RT device and are willing to jailbreak it, I could try to hack arm-win32 support together for you to test, but finegrained bug fixing (if the port should indeed work) needs to be done by someone owning the device.







_______________________________________________
fpc-pascal maillist  -  [hidden email]
http://lists.freepascal.org/mailman/listinfo/fpc-pascal
Reply | Threaded
Open this post in threaded view
|

Re: Windows on ARM, how fast it can be implemented?

shiruba2012

Hi,

Well windows has only about 1.5% of the tablet market, and its doubtful if that will grow - plus, as mentioned, it's closed by design.  Then again, Apple's app store is restricted, and fpc supports that.

Thank you,
    Noah Silva

2013/03/27 4:15 "Max Vlasov" <[hidden email]>:
Thanks for the info, Sven. I will probably have a chance to get a device and jailbreak it, although not right now. If someone on the list does this before me, I'll be glad. I did not find any prior discussion about this topic, seems like it's not that hot to be discussed and everything.

Max

On Tue, Mar 26, 2013 at 8:05 PM, Sven Barth <[hidden email]> wrote:

Am <a href="tel:26.03.2013%2015" value="+12603201315" target="_blank">26.03.2013 15:01 schrieb "Max Vlasov" <[hidden email]>:


That said it mostly depends on how much Microsoft changed their existing ARM compiler. If the changes are minimal for the purpose of user space applications then it should be rather easy to adjust the Win32 target for this. The biggest problem might be unaligned accesses in the Win32 RTL code, but AFAIK newer ARMs can handle these...

If you own a Windows RT device and are willing to jailbreak it, I could try to hack arm-win32 support together for you to test, but finegrained bug fixing (if the port should indeed work) needs to be done by someone owning the device.







_______________________________________________
fpc-pascal maillist  -  [hidden email]
http://lists.freepascal.org/mailman/listinfo/fpc-pascal

_______________________________________________
fpc-pascal maillist  -  [hidden email]
http://lists.freepascal.org/mailman/listinfo/fpc-pascal
Reply | Threaded
Open this post in threaded view
|

Re: Windows on ARM, how fast it can be implemented?

Sven Barth-2
Am 27.03.2013 01:22, schrieb Noah Silva:
>
> Well windows has only about 1.5% of the tablet market, and its
> doubtful if that will grow - plus, as mentioned, it's closed by
> design.  Then again, Apple's app store is restricted, and fpc supports
> that.
>
This is not about Windows apps. Support for WinRT (aka Metro) would be a
completely different topic. This is about running normal desktop
applications compied for arm-win32 on a Windows RT tablet and those
won't be found in the Windows app store, because you need to jailbreak
the Windows RT device to be able to run unsigned desktop applications.

Regards,
Sven
_______________________________________________
fpc-pascal maillist  -  [hidden email]
http://lists.freepascal.org/mailman/listinfo/fpc-pascal
Reply | Threaded
Open this post in threaded view
|

Re: Windows on ARM, how fast it can be implemented?

Michael Schnell
On 03/27/2013 09:20 AM, Sven Barth wrote:
>  you need to jailbreak the Windows RT device to be able to run
> unsigned desktop applications.
Never buy such a system !

I am not even sure if this is legal in Europe.... I.e. you are the owner
of the hardware.

-Michael
_______________________________________________
fpc-pascal maillist  -  [hidden email]
http://lists.freepascal.org/mailman/listinfo/fpc-pascal
Reply | Threaded
Open this post in threaded view
|

Re: Windows on ARM, how fast it can be implemented?

Marco van de Voort
In reply to this post by Sven Barth-2
In our previous episode, Sven Barth said:

> > Well windows has only about 1.5% of the tablet market, and its
> > doubtful if that will grow - plus, as mentioned, it's closed by
> > design.  Then again, Apple's app store is restricted, and fpc supports
> > that.
> >
> This is not about Windows apps. Support for WinRT (aka Metro) would be a
> completely different topic. This is about running normal desktop
> applications compied for arm-win32 on a Windows RT tablet and those
> won't be found in the Windows app store, because you need to jailbreak
> the Windows RT device to be able to run unsigned desktop applications.

Does Metro btw make a difference between the visual, language and system
libraries?   (like GDI/MSVCRT/win32 in Win32/64)
_______________________________________________
fpc-pascal maillist  -  [hidden email]
http://lists.freepascal.org/mailman/listinfo/fpc-pascal
Reply | Threaded
Open this post in threaded view
|

Re: Windows on ARM, how fast it can be implemented?

Sven Barth-2
Am 27.03.2013 09:41, schrieb Marco van de Voort:

> In our previous episode, Sven Barth said:
>>> Well windows has only about 1.5% of the tablet market, and its
>>> doubtful if that will grow - plus, as mentioned, it's closed by
>>> design.  Then again, Apple's app store is restricted, and fpc supports
>>> that.
>>>
>> This is not about Windows apps. Support for WinRT (aka Metro) would be a
>> completely different topic. This is about running normal desktop
>> applications compied for arm-win32 on a Windows RT tablet and those
>> won't be found in the Windows app store, because you need to jailbreak
>> the Windows RT device to be able to run unsigned desktop applications.
> Does Metro btw make a difference between the visual, language and system
> libraries?   (like GDI/MSVCRT/win32 in Win32/64)
I've not looked at this in detail, but AFAIK you only have certain
libraries and functions available and it will be checked by the Windows
store and the OS that you don't load anything else that you shouldn't
load... (you don't even have full access to the filesystem :( ). That's
why a complete new OS port would be necessary to support WinRT...

Regards,
Sven
_______________________________________________
fpc-pascal maillist  -  [hidden email]
http://lists.freepascal.org/mailman/listinfo/fpc-pascal
Reply | Threaded
Open this post in threaded view
|

Re: Windows on ARM, how fast it can be implemented?

Michael Schnell
On 03/27/2013 10:00 AM, Sven Barth wrote:
> That's why a complete new OS port would be necessary to support WinRT...
>
A new OS would be more appropriate :-P .

-Michael
_______________________________________________
fpc-pascal maillist  -  [hidden email]
http://lists.freepascal.org/mailman/listinfo/fpc-pascal
Reply | Threaded
Open this post in threaded view
|

Re: Windows on ARM, how fast it can be implemented?

Marco van de Voort
In reply to this post by Sven Barth-2
In our previous episode, Sven Barth said:
> > Does Metro btw make a difference between the visual, language and system
> > libraries?   (like GDI/MSVCRT/win32 in Win32/64)
> I've not looked at this in detail, but AFAIK you only have certain
> libraries and functions available and it will be checked by the Windows
> store and the OS that you don't load anything else that you shouldn't
> load... (you don't even have full access to the filesystem :( ). That's
> why a complete new OS port would be necessary to support WinRT...

That's what I heard too, and in addition that most functionality is exposed
as (COM) interfaces. But I haven't seen more detail. I hoped for an article
in the C'T about it, but either I missed it, or there hasn't been.
_______________________________________________
fpc-pascal maillist  -  [hidden email]
http://lists.freepascal.org/mailman/listinfo/fpc-pascal
Reply | Threaded
Open this post in threaded view
|

Re: Windows on ARM, how fast it can be implemented?

Sven Barth-2
In reply to this post by Michael Schnell
Am 27.03.2013 10:06, schrieb Michael Schnell:
> On 03/27/2013 10:00 AM, Sven Barth wrote:
>> That's why a complete new OS port would be necessary to support WinRT...
>>
> A new OS would be more appropriate :-P .
I'm not a fan of Windows 8 either, but if I see the possiblity to run
FPC apps there, I'll take it. (Seeing that this is also something that
Delphi can't do :P )

Regards,
Sven
_______________________________________________
fpc-pascal maillist  -  [hidden email]
http://lists.freepascal.org/mailman/listinfo/fpc-pascal
Reply | Threaded
Open this post in threaded view
|

Re: Windows on ARM, how fast it can be implemented?

Sven Barth-2
In reply to this post by Marco van de Voort
Am 27.03.2013 10:07, schrieb Marco van de Voort:

> In our previous episode, Sven Barth said:
>>> Does Metro btw make a difference between the visual, language and system
>>> libraries?   (like GDI/MSVCRT/win32 in Win32/64)
>> I've not looked at this in detail, but AFAIK you only have certain
>> libraries and functions available and it will be checked by the Windows
>> store and the OS that you don't load anything else that you shouldn't
>> load... (you don't even have full access to the filesystem :( ). That's
>> why a complete new OS port would be necessary to support WinRT...
> That's what I heard too, and in addition that most functionality is exposed
> as (COM) interfaces. But I haven't seen more detail. I hoped for an article
> in the C'T about it, but either I missed it, or there hasn't been.
Some of the low level API functions are available (but only selected
ones) and then there are the COM APIs. Seeing that our support for COM
has matured in 2.6 a WinRT port should be doable, but one would need (a)
time and (b) a Windows 8 system. I don't have either (the latter mostly,
because my university has not yet managed to provide free versions... -.-).

Regards,
Sven
_______________________________________________
fpc-pascal maillist  -  [hidden email]
http://lists.freepascal.org/mailman/listinfo/fpc-pascal
Reply | Threaded
Open this post in threaded view
|

Re: Windows on ARM, how fast it can be implemented?

Michael Schnell
In reply to this post by Sven Barth-2
On 03/27/2013 10:09 AM, Sven Barth wrote:
> I'm not a fan of Windows 8 either, ...

Is Win 8 on normal PCs crippled in a similar manner or is only the ARM
version affected ?

-Michael

_______________________________________________
fpc-pascal maillist  -  [hidden email]
http://lists.freepascal.org/mailman/listinfo/fpc-pascal
Reply | Threaded
Open this post in threaded view
|

Re: Windows on ARM, how fast it can be implemented?

Sven Barth-2
Am 27.03.2013 10:15, schrieb Michael Schnell:
> On 03/27/2013 10:09 AM, Sven Barth wrote:
>> I'm not a fan of Windows 8 either, ...
>
> Is Win 8 on normal PCs crippled in a similar manner or is only the ARM
> version affected ?
Only the ARM version. Though I'd not want to use the PC version either :P

Regards,
Sven
_______________________________________________
fpc-pascal maillist  -  [hidden email]
http://lists.freepascal.org/mailman/listinfo/fpc-pascal
Reply | Threaded
Open this post in threaded view
|

Re: Windows on ARM, how fast it can be implemented?

Michael Schnell
On 03/27/2013 10:16 AM, Sven Barth wrote:
> Only the ARM version. Though I'd not want to use the PC version either :P
>
I'm planning a (kind of) embedded device that need high performance
(quad core, 64 Bit, 4 Gig RAM, SSD).

I need to use Windows, because of the 3rd party software I need to install.

I was told for this Win 8 would offer a better infrastructure than Win 7.

In fact  I was told that the abysmal Desktop can be deactivated in Win 8.

What do you think ?

Thanks.
-Michael
_______________________________________________
fpc-pascal maillist  -  [hidden email]
http://lists.freepascal.org/mailman/listinfo/fpc-pascal
Reply | Threaded
Open this post in threaded view
|

Re: Windows on ARM, how fast it can be implemented?

Marco van de Voort

(reply-to set to fpc-other since this is getting off-topic)

In our previous episode, Michael Schnell said:
> > Only the ARM version. Though I'd not want to use the PC version either :P
> >
> I'm planning a (kind of) embedded device that need high performance
> (quad core, 64 Bit, 4 Gig RAM, SSD).
>
> I need to use Windows, because of the 3rd party software I need to install.
>
> I was told for this Win 8 would offer a better infrastructure than Win 7.

As windows 6.2, Win8 is a win6 (vista) point release. The main differences
are just like Vista->Win7, some minor performance (mainly startup) and
startup updates (USB3 in preboot env), some new SMB/NTFS options etc.  And performance
mainly in the startup.
  As with any new release, the drivers.cab is updated, so most current
common HW is detected out of the box.

There are some downsides too, though they mostly are related to the defaults
than real issues. (the quick startup is default, but not compatible with many
industrial drivers, many shutdown options are hidden by default etc etc )

And of course, the whole startmenu episode is a bit sad, but IMHO a minor
footnote just like Vista's gadgets (that have mysteriously disappeared in
the next version too).  And I hope that electronic delivery will cut out the
middle man and thus some of the price in the long run, so the whole shop
idea is not that bad, if it isn't forced on you too much.  (it is only for
MS services like messenger, MS based mail and Skype (I only use skype), and
you have to dig out the desktop games from out of the store). Only time will
tell.
 
But I never used the startmenu much in the first place, and I consider it an
anti RSI measure (if you miss the startmenu in the default location, you
probably use the mouse too much :-)

Yes, it is still not an improvement overall, but it shouldn't exaggerated.
IMHO critics are still a bit power-high over the Vista revolt, and gunning
for an encore. The whole discussion only detracts from the real issues and
differences.

> In fact  I was told that the abysmal Desktop can be deactivated in Win 8.

Not by standard means. There is a lot of 3rd party cruft out there of
course, and even OEMs deliver some, but I wouldn't recommend it.

> What do you think ?

Win8 vs Win7 is the least important bit of your future device choice.

I mildly biassed towards win8, but I mostly migrated because of attractive
offers, not features.  I would get a 64-bit edition though.
_______________________________________________
fpc-pascal maillist  -  [hidden email]
http://lists.freepascal.org/mailman/listinfo/fpc-pascal
12