public inbox for pve-devel@lists.proxmox.com
 help / color / mirror / Atom feed
From: Maximiliano Sandoval <m.sandoval@proxmox.com>
To: "DERUMIER, Alexandre" <alexandre.derumier@groupe-cyllene.com>
Cc: "pve-devel@lists.proxmox.com" <pve-devel@lists.proxmox.com>
Subject: Re: [PATCH manager] ui: processor edit: allow setting more than 4 sockets
Date: Thu, 26 Mar 2026 14:35:39 +0100	[thread overview]
Message-ID: <s8oy0jeogr8.fsf@toolbox> (raw)
In-Reply-To: <85b678425cfe0ac844e64a9c2e5f20c67203005e.camel@groupe-cyllene.com> (Alexandre DERUMIER's message of "Thu, 26 Mar 2026 13:04:35 +0000")

"DERUMIER, Alexandre" <alexandre.derumier@groupe-cyllene.com> writes:

> Hi,
> isn't it better to extend our current numa option (which create 1 node
> by socket), to something like "numa_nodes_by_socket" ?

As of now, fine-tuning the NUMA configuration is achieved via the
`numa[n]` parameters.

> Also, I think it could be great to enable numa by default for the vms
> too.

I have not been able to detect any improvement when running benchmarks
on my guest with NUMA enabled with or without defining a topology. The
content of this patch is to simply reconcile the documentation and API
with the UI.

I would be interested however in setups+benchmarks that show any
improvement.

> Alexandre
>
> -------- Message initial --------
> De: Maximiliano Sandoval <m.sandoval@proxmox.com>
> À: pve-devel@lists.proxmox.com
> Objet: [PATCH manager] ui: processor edit: allow setting more than 4
> sockets
> Date: 26/03/2026 11:50:29
>
> As per our NUMA documentation [1]:
>
>> If the NUMA option is used, it is recommended to set the number of
>> sockets to the number of nodes of the host system.
>
> Nowadays, it is common to find CPU models with 8 NUMA nodes. Two
> examples are Xeon Platinum 8160 and EPYC 73F3.
>
> The number 16 was picked to allow for some future proofing while
> remaining conservative.
>
> The schema at qemu-server does not limit the socket amount.
>
> [1] https://pve.proxmox.com/pve-docs/pve-admin-guide.html#_numa
>
> Signed-off-by: Maximiliano Sandoval <m.sandoval@proxmox.com>
> ---
>  www/manager6/qemu/ProcessorEdit.js | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/www/manager6/qemu/ProcessorEdit.js
> b/www/manager6/qemu/ProcessorEdit.js
> index ffaf014e..11912e3c 100644
> --- a/www/manager6/qemu/ProcessorEdit.js
> +++ b/www/manager6/qemu/ProcessorEdit.js
> @@ -113,7 +113,7 @@ Ext.define('PVE.qemu.ProcessorInputPanel', {
>              xtype: 'proxmoxintegerfield',
>              name: 'sockets',
>              minValue: 1,
> -            maxValue: 4,
> +            maxValue: 16,
>              value: '1',
>              fieldLabel: gettext('Sockets'),
>              allowBlank: false,

-- 
Maximiliano




  reply	other threads:[~2026-03-26 13:35 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-03-26 10:50 Maximiliano Sandoval
2026-03-26 13:04 ` DERUMIER, Alexandre
2026-03-26 13:35   ` Maximiliano Sandoval [this message]
2026-03-26 14:27   ` DERUMIER, Alexandre

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=s8oy0jeogr8.fsf@toolbox \
    --to=m.sandoval@proxmox.com \
    --cc=alexandre.derumier@groupe-cyllene.com \
    --cc=pve-devel@lists.proxmox.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
Service provided by Proxmox Server Solutions GmbH | Privacy | Legal