[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Qemu-discuss] Multi-node QEMU
From: |
Javier Picorel |
Subject: |
Re: [Qemu-discuss] Multi-node QEMU |
Date: |
Fri, 11 Apr 2014 20:36:20 +0200 |
Dear Wacha,
Thanks for your reply. The cycle-accurate models are built on top of QEMU. As
an example, MARSS86 in as cycle-accurate x86 simulator that
works on top of QEMU. Going back to my question, can we boot a multi-node setup
within one QEMU process? Thanks!
Regards,
Javier
On 11 Apr 2014, at 20:32, Wacha Gábor <address@hidden> wrote:
> Dear Javier,
>
> As far as I understand, QEMU itself is not usable for cycle accurate
> simulation even with only one node (because of the tcg).
>
> Regards,
> Gabor Wacha
>
> Javier Picorel <address@hidden> a következőt írta:
>
>> Dear all,
>>
>> I would like to know if there is the possibility in QEMU to boot several
>> machines that are synchronized in a single QEMU process. That means, that at
>> each cycle,
>> all the machines are emulated in a “round-robin fashion". Otherwise, the
>> results are not deterministic and the request rate is not controllable
>> because it will be defined
>> by the amount of “host cpu cycles” given to each of the different QEMU
>> processes (aka emulated machines). We are interested in a cycle-accurate
>> full-system
>> architectural simulator that works on top on QEMU. Is there any way to do
>> this?
>>
>> Regards,
>> Javier
>>
>>