Fw: Aw: RE: bochs on MorphOS

Sorry sent it only to stanislav, here for others who might also be interessted in it:

Hello again

here the promised statistics result for booting WinXP on bochs on MorphOS (1GHz G4)

Boot time up to login screen: apprx 3 minutes

First time booting from login to desktop: 16-17 minutes

Every other boot time from login to desktop: 2-3 minutes

Games testet so far:

Minesweeper: very smooth

Solitair: very good playable

Hope to fix the crash bugs on ending the emulator soon. This is because of incorrect handling, this mistake is done by the first porter for the AmigaOS GUI. Also I had to make some changes to let it run on MorphOS correctly.

CPU latency is between 80% and 100% all the time, but MorphOS is designed, that it's mostly still able to react in the background. So this is not a real problem.

So I want to say thank you, to make the code much more portable, faster and better!

I recognized that most of the speed could be get lost by loading from disk image. I will try some stuff to speed it up later. If it's possible at all...

Regards

Thore

Gesendet: Mittwoch, 03. April 2013 um 17:04 Uhr

Von: Stanislav <[email protected]>

An: "'Thore Sittly'" <[email protected]>, [email protected]

Betreff: RE: [Bochs-developers] bochs on MorphOS

Hi,

I don’t care for installation times, if you really want you could install an OS at your fast Intel desktop PC and re-use the same disk image later.

How about boot time ?

Stanislav

From: Thore Sittly [mailto:[email protected]]

Sent: Wednesday, April 03, 2013 1:19 PM

To: [email protected]

Subject: [Bochs-developers] bochs on MorphOS

Hello

The MorphOS port is almost done, just need to check some minor issues with e.g. graphics.

But it's way too slow. On a 1GHz PowerPC host machine, WinXP needs more than 8 hours to install.

Do you have some ideas to speed it up? If it's the byte swap, what possibilities do we have?

Some months before we thought about pre-swapped memory or JIT. But this is really complicated.

I also thought about replacing instructions with fixed values by pre-swapped commands, ("inventing" new commands which behaves like big endian and replace the LE commands with this replacements at runtime, by using not used mnemonics?) but not sure if this could work.

Another idea could be improving the reading speed of the devices.

Any more idea or concerns? :)

Thank you

Kind regards

Thore Sittly

------------------------------------------------------------------------------
Minimize network downtime and maximize team effectiveness.
Reduce network management and security costs.Learn how to hire 
the most talented Cisco Certified professionals. Visit the 
Employer Resources Portal
http://www.cisco.com/web/learning/employer_resources/index.html

_______________________________________________
bochs-developers mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/bochs-developers