From: Franz Fehringer <fehrin2@gmail.com>
To: cygwin@cygwin.com
Subject: Re: Cygwin X11 on second (Nvidia) graphics card
Date: Mon, 30 Jan 2023 19:50:20 +0100 [thread overview]
Message-ID: <tr93hc$bn4$1@ciao.gmane.io> (raw)
In-Reply-To: <tr0q79$5tu$1@ciao.gmane.io>
Am 27.01.2023 um 16:22 schrieb Franz Fehringer via Cygwin:
> Hi,
>
> I have a question which seems quite natural but i cannot find anything
> useful using google & co.
> My (business) notebook has two graphics cards, one builtin identifying
> as Intel Iris Xe and an additional one, a Nvidia MX550.
> I start X11 (xwin) with defaults (-listen tcp is the only extra option).
> Using nvidia-smi and the Nvidia activity dialog i see that the Nvidia
> graphics card is not used.
> Using glxinfo i see that the builtin Intel Iris Xe is used instead.
> I already prefer Nvidia over Intel via the Nvidia control panel.
> How do i make X11 and OpenGL use the Nvidia graphics card instead of the
> Intel one?
>
> Thx in advance
>
> Franz
>
>
>
Investigating the issue further (incorporating the answers i got) the
situation looks as follows:
Both the builtin display and the external HDMI connected monitor are
connected to the internal Intel graphics card at first glance.
The idea in this constellation is that the user can decide both as a
general preference and application wise if the second graphics card
(better performance and higher power consumption) ot the first (builtin)
one is used.
These settings are duplicated (Nvidia control *and* M$ system control),
i use identical settings utilizing both approaches.
Now the XWin executable does not use the Nvidia graphics card.
But Cygwin executables like glxinfo or glxspheres do if i assign them
specifically to the Nvidia graphics card using the settings mentioned
(only setting the general preference does not suffice while it is
sufficient for several non Cygwin programs like teams or msedge).
In a sense the XWin server using the Nvidia graphics card would amount
to a (sort of) headless mode (graphics card but no monitor), at least
this is what i think: The XWin server does not use the Nvidia graphics
card because it is not connected to a display (monitor).
This is otherwise a valid scenario in the VirtualGL realm for example
and there is a short explanation how make this (i.e. X11 uses graphics
card with no monitor connected) work:
https://virtualgl.org/Documentation/HeadlessNV
In summary it seems to me that in this specific situation there is no
easy way to make XWin use the Nvidia graphics card, but i may well be
corrected about this.
It would be interesting what someone like Jon Turney has to say about this.
And i am still surprised that this seems to have been not discussed so far.
Thanks to all who answered (and those who will be answering).
prev parent reply other threads:[~2023-01-30 18:50 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-01-27 15:22 Franz Fehringer
2023-01-27 19:58 ` Brian Inglis
2023-01-29 17:18 ` Franz Fehringer
2023-01-29 22:05 ` Brian Inglis
2023-01-30 1:08 ` Duncan Roe
2023-01-30 18:50 ` Franz Fehringer [this message]
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='tr93hc$bn4$1@ciao.gmane.io' \
--to=fehrin2@gmail.com \
--cc=cygwin@cygwin.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;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).