On Wed, Sep 15, 2010 at 7:38 AM, Nick Couchman <Nick.Couchman@xxxxxxxxx> wrote:
>>>> On 2010/09/15 at 03:20, Rudi Ahlers <Rudi@xxxxxxxxxxx> wrote:
>> Hi all,
>>
>> I'm just curios and would like some input from the community on this
>> one. We're busy budgeting for a couple of new servers and I thought it
>> would be good to try out the Core i7 CPU's, but see the majority of
>> them don't offer VT-d, but just VT-x. Looking at the LGA1366 range,
>> only the "Intel lga1366 i7 980XE" (from the list of what our suppliers
>> stock) have VT-d, and it costs 4x more than "Intel lga1366 i7 930" or
>> 2x more than "Intel lga1366 i7 960". From a budget perspecitve I could
>> purchase 4 more CPU's, which could translate to 40x - 80x more VM's
>> being hosted for the same capital outlay. Experience has shown that we
>> under-utilize CPU's by a great margin and memory / HDD IO is our
>> biggest bottleneck on any server.
>
> That's interesting...is it the motherboards you're seeing this lack of
> support on, or the chips? I'm pretty sure the i7 processors themselves
> support VTd, you just have to have the BIOS and MB support for it.
>
>>
>> So, if VT-d really necessary?
>> We mainly host XEN virtual machine for the hosting industry, i.e. we
>> don't need / use graphics rendering inside VM's, or need DAS on the
>> VM's, etc.
>
> In an environment where you're planning on running many VMs on a single host,
> I don't imagine VTd is going to be very helpful to you. I use it here where
> I work, but I use it on desktop systems where folks need access to graphics
> cards, serial cards, etc., from Windows, and not on my server systems. It
> really depends on what you're doing as to whether you think you'll need it or
> not - if you cannot think of a situation where *any* type of PCI card needs
> to be forwarded through to an HVM domU, then you're probably okay without it.
> The few situations I can think of are:
> - Serial port cards, but this is really more common on desktops
> - Phone/Voicemail systems - If you're using any sort of telephony card with
> an HVM domU you'll need direct PCI access, which requires VTd.
> - Other, vendor-specific add-in PCI cards, for applications like industrial
> automation, etc. But you won't see much of this in a data center.
>
> Also, keep in mind that, IIRC, PV domUs can access PCI devices without VTd,
> so if you're going to be running PV kernels on these systems, or the PV
> kernels you're running are the ones that need access to the PCI devices, then
> you're probably fine without it. I'm sure it offers some performance
> enhancements over software-based IOMMU, but I don't know what those are.
>
> -Nick
>
Apologies in advance for hyjacking your thread, but this is an
interesting question.
For cases where PCI passthougth is the heart of the application, is it
better to use hvm/VTd or pv-ops/iommu=soft? Or is there some other
combination where you can still leverage VTd from pv-ops?
-Bruce
>
>
> --------
> This e-mail may contain confidential and privileged material for the sole use
> of the intended recipient. If this email is not intended for you, or you are
> not responsible for the delivery of this message to the intended recipient,
> please note that this message may contain SEAKR Engineering (SEAKR)
> Privileged/Proprietary Information. In such a case, you are strictly
> prohibited from downloading, photocopying, distributing or otherwise using
> this message, its contents or attachments in any way. If you have received
> this message in error, please notify us immediately by replying to this
> e-mail and delete the message from your mailbox. Information contained in
> this message that does not relate to the business of SEAKR is neither
> endorsed by nor attributable to SEAKR.
>
> _______________________________________________
> Xen-users mailing list
> Xen-users@xxxxxxxxxxxxxxxxxxx
> http://lists.xensource.com/xen-users
>
_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-users
|
Home