Para-Bochs 0.2 is announced at:
http://grid.hust.edu.cn/cluster/VirtualMachine/index.html

Stanislav, I have tried to make the most current version of Bochs (2.3) to run in parallel, however, it just does not work well, and I am quite busy for these days, sorry!
I think you can read the Para-Bochs 0.2 source code, and write the related statements in your newest version bochs, anyway, my teammates and me will feel glad and proud for our contribution in Bochs community. If you have problems, pls send emails to me for discussion.

------------------    
Zhiyuan Shao
2007-06-13



2007/5/28, Stanislav <stl@fidonet.org.il>:
Could you send me please your final code ?
I tried to look on the code but I don't see anything which will avoid contention of memory writes to the same address by 2 different CPUs (memory coherency). In general any Bochs memory write should be locked also, as well as APIC and devices.

I already thought about making this para-Bochs a while ago and already build some todo-list so it will be interesting to see your final code.

Thanks,
Stanislav

-----Original Message-----
From: Zhiyuan Shao@HUST [mailto:zyshao.hust@gmail.com]
Sent: Sunday, May 27, 2007 5:04 PM
To: Stanislav; bochs-developers@lists.sourcefor
Subject: Re: RE: [Bochs-developers] Para-Bochs

Thanks for your reply and your attention.

We build Para-Bochs based on Bochs 2.2.6. As we have almost done few revisions of the source code except for the execution mode of multiple processors, we think the version number does not matter.

The published version Para-Bochs0.1 is NOT so stable, you may fail during booting the system even when two way SMP system is emulated. However, we have partly solved the problem now, and we can support 4-way SMP systems stably (only tested for RHEL guest systems). I will update our homepage soon, and publish Para-Bochs version 0.2.

Currently, we are working at:
Windows version Para-Bochs;
Documents that explains the internals of Para-Bochs.

Hope Para-Bochs will be part of the Bochs main source code in near future.


------------------
Zhiyuan Shao
2007-05-27

-------------------------------------------------------------
发件人:Stanislav
发送日期:2007-05-27 21:06:59
收件人:'Zhiyuan SHAO'; bochs-developers@lists.sourceforge.net
抄送:
主题:RE: [Bochs-developers] Para-Bochs

I am very interesting in your work and see it top priority in integration
into Bochs main source code.

I will explore your sources soon and definitely will come with questions.

Which Bochs version it based ?



If you have any questions to Bochs developers you are welcome to ask also.



Thanks for the job done,

Stanislav



_____

From: bochs-developers-bounces@lists.sourceforge.net
[mailto:bochs-developers-bounces@lists.sourceforge.net] On Behalf Of Zhiyuan
SHAO
Sent: Friday, May 25, 2007 12:58 PM
To: bochs-developers@lists.sourceforge.net
Subject: [Bochs-developers] Para-Bochs



We created Para-Bochs, which can run multiple CPU instruction streams in
parallel.

The acceleration (even above linear acceleration rate) can be achieved when
simulating SMP guest OS (we use Redhat 9.0 with SMP support) on real SMP (or
multi-core) host.

Here is it:
< http://grid.hust.edu.cn/cluster/VirtualMachine/index.html>
http://grid.hust.edu.cn/cluster/VirtualMachine/index.html

While, we are continuing working on it to make it more stable and
reliable...

Any comments are welcomed!