From: Pavel Begunkov <asml.silence@gmail.com>
To: "Li Wang" <liwang@kylinos.cn>,
"Christian König" <christian.koenig@amd.com>,
"Greg Kroah-Hartman" <gregkh@linuxfoundation.org>
Cc: Arnd Bergmann <arnd@arndb.de>,
Sumit Semwal <sumit.semwal@linaro.org>,
linux-media@vger.kernel.org, dri-devel@lists.freedesktop.org,
linaro-mm-sig@lists.linaro.org, io-uring@vger.kernel.org,
linux-kernel@vger.kernel.org,
Mengmeng Zhao <zhaomengmeng@kylinos.cn>
Subject: Re: [RFC PATCH] misc: fgds: enable GPU-NVMe direct I/O via POSIX and io_uring
Date: Sat, 12 Sep 2026 13:45:11 +0100 [thread overview]
Message-ID: <7b126162-f412-41d8-8002-6b31d2795f51@gmail.com> (raw)
In-Reply-To: <9447bb67-96e9-43b3-87c8-db231051fa2d@kylinos.cn>
On 9/10/26 04:57, Li Wang wrote:
> Hi Christian,
>
>>>>> Nor why "fgds" is the name, that's going to be hard to remember, does it
>>>>> stand for something?
>>>>>
>>>> "FGDS" stands for Fast GPUDirect Storage. GPUDirect Storage (GDS) is NVIDIA's
>>>> technology enabling direct I/O between GPU memory and files on NVMe,
>>>> widely used in LLM workloads to bypass CPU overhead.
>>>
>>> That's nvidia's specific solution, but this works on other devices,
>>> right? Or just for that one platform?
>>
>> That was nvidia's specific and very hacky out of tree solution
> Exactly, I completely agree with your viewpoint. In fact, we previously conducted
> a deep analysis of NVIDIA's GDS implementation code, which was also one of the
> motivations for us to develop fgds.
>
> Please allow me to introduce NVIDIA's GDS implementation briefly:
> NVIDIA introduced two kernel modules: one module called nvidia-fs, and the other module
> is a customized NVMe driver replacing the Linux kernel's default NVMe driver.
> nvidia-fs creates a device file for every NVIDIA GPU on the machine: /dev/nvidia-fs<gpu-id>.
I'm not an nvidia expert, but IIRC they switched to P2PDMA, which
is supported by the upstream nvme driver.
--
Pavel Begunkov
prev parent reply other threads:[~2026-09-12 12:45 UTC|newest]
Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-09-08 13:15 [RFC PATCH] misc: fgds: enable GPU-NVMe direct I/O via POSIX and io_uring Li Wang
2026-09-09 6:10 ` Greg Kroah-Hartman
2026-09-09 10:42 ` Li Wang
2026-09-09 13:35 ` Greg Kroah-Hartman
2026-09-09 13:45 ` Christian König
2026-09-10 3:57 ` Li Wang
2026-09-10 8:35 ` Christian König
2026-09-12 12:45 ` Pavel Begunkov [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=7b126162-f412-41d8-8002-6b31d2795f51@gmail.com \
--to=asml.silence@gmail.com \
--cc=arnd@arndb.de \
--cc=christian.koenig@amd.com \
--cc=dri-devel@lists.freedesktop.org \
--cc=gregkh@linuxfoundation.org \
--cc=io-uring@vger.kernel.org \
--cc=linaro-mm-sig@lists.linaro.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-media@vger.kernel.org \
--cc=liwang@kylinos.cn \
--cc=sumit.semwal@linaro.org \
--cc=zhaomengmeng@kylinos.cn \
/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