Fw: about vNUMA for HVM

Fw: about vNUMA for HVM

[butine]

[-- Attachment #1.1: Type: text/plain, Size: 409 bytes --]




-----原始邮件-----
发件人: butine@zju.edu.cn
发送时间: 2013年10月24日 星期四
收件人:抄送:主题: about vNUMA for HVM



hello,Elena,

I have implemented the vNUMA for HVM guest,I use the NPB(OMP and MPI) benchmark to test the performance,but no improvement.Howerver,VMWare declares that the vNUMA Can greatly enhance the performance of 30%,why?

And about your vNUMA for guest?

 

Butine Huang

2013-10-24  

[-- Attachment #1.2: Type: text/html, Size: 521 bytes --]

[-- Attachment #2: Type: text/plain, Size: 126 bytes --]

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel

Re: Fw: about vNUMA for HVM

[Elena Ufimtseva]

[-- Attachment #1.1: Type: text/plain, Size: 971 bytes --]

Hi
On Oct 24, 2013 12:16 PM, <butine@zju.edu.cn> wrote:
>
>
>
>
> -----原始邮件-----
> 发件人: butine@zju.edu.cn
> 发送时间: 2013年10月24日 星期四
> 收件人: 抄送: 主题: about vNUMA for HVM
>
> hello,Elena,

Hi Butine.

At the moment I did not perform that kind of evaluation and as soon the
major issues will be resolved that will be the next step to evaluate
performance gain.

Though in general I would like to see how VMware evaluated the performance
gain in 30% and what applications they were using for such tests. I think
the results may vary pretty dramatically in regards to workload type.
Do you have any more information in this?

Elena
>
> I have implemented the vNUMA for HVM guest,I use the NPB(OMP and MPI)
benchmark to test the performance,but no improvement.Howerver,VMWare
declares that the vNUMA Can greatly enhance the performance of 30%,why?
>
> And about your vNUMA for guest?
>
>
>
> Butine Huang
>
> 2013-10-24

[-- Attachment #1.2: Type: text/html, Size: 1295 bytes --]

[-- Attachment #2: Type: text/plain, Size: 126 bytes --]

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel