NVMe Performance Compared: Windows Server 2025 vs. Ubuntu Server 24.04.4 LTS

[-]

Slasher1738@reddit

Interesting. I look forward to seeing a retest with Ubuntu 26.04

[-]

Maybe I'm jaded but the fact this article dropped 3 weeks before the next LTS version of Ubuntu drops makes me think there is a significant performance boost on these tests with a more recent Kernel. In which case my jaded self thinks we will be waiting 2 years for that retest.

[-]

Thotaz@reddit

Do you have any reason to suspect them having a bias against Ubuntu or for Windows?

[-]

RetdThx2AMD@reddit

Yes, the fact that they are doing this review out of the blue (none of their other reviews over the last few years are Windows vs Ubuntu) using an LTS version that gets replaced in 3 weeks.

[-]

Thotaz@reddit

But it's not out of the blue? This "Native NVMe" support is new for Windows. It was announced in December, and Storagereview made an article about it last month: https://www.storagereview.com/review/windows-server-native-nvme which btw, this article references in the very first line:

After publishing our article about Microsoft’s opt-in native NVMe feature on Windows Server 2025, we received multiple requests for a direct comparison of storage performance between Windows Server 2025 with native NVMe and a Linux-based server OS.

And they've used Ubuntu LTS in past reviews, so why wouldn't they also use Ubuntu LTS for this comparison? I think you should ask yourself if you would have been this skeptical over the review if Ubuntu had come out on top. To me this looks like a typical attempt at discrediting an article/test result because you didn't like the outcome.

[-]

RetdThx2AMD@reddit

I was responding to a guy who was expecting a retest with 26.04 LTS. You keep your eyes peeled and let me know when it drops. I won't hold my breath.

[-]

Thotaz@reddit

Again the timing was most likely intentional

That's what I was questioning in the original comment. You think they are maliciously trying to make Ubuntu look bad or make Windows look good. Your only argument for this is because the timing is bad for Ubuntu, but I've explained the likely reason behind this timing.
If you can't point to anything else then you don't really have a solid argument for thinking this way.

As for them waiting 3 weeks for the new release, why should they? The summary of new features makes no mention of improved disk performance: https://documentation.ubuntu.com/release-notes/26.04/summary-for-lts-users/
Maybe there will be some improvements anyway, but if they aren't worth highlighting there then there is no reason to think it would make a big difference in the results.

[-]

RetdThx2AMD@reddit

Why? Because their article becomes somewhat obsolete 3 weeks after publishing it.

[-]

Thotaz@reddit

Lol no it's not. If you are running LTS builds then you probably aren't going to be upgrading to the latest LTS version as soon as it's out. The current LTS stays relevant for a long time because that's kinda the whole point of them.
And like I said, there is no reason to suspect that the new version is way faster.

[-]

ViamoIam@reddit

Good point for existing users.

This article will be outdated in a month for 26.04LTS, which matters. If we are picking up a lot of storage, we may be upgrading the system. This could mean bringing in a new system online.

[-]

RetdThx2AMD@reddit

Maybe not, but if you were considering switching to Windows at all you wouldn't be that interested in comparing it to the old LTS, you would want to compare it to the new one. Which again leads me to think this article was pushed out before 26.04 LTS for a reason.

[-]

Slasher1738@reddit

It's possible. There are beta builds available.

[-]

DownloadTheInternet5@reddit

not suprised linux pulled ahead on the sequential stuff, the io scheduler and filesystem drivers have been getting crazy optimized for nvme over the last few kernel versions. the interesting part is the random 4k results tho, thats where you actually feel the difference in real workloads. would love to see them test with btrfs vs ext4 vs xfs too since filesystem choice matters almost as much as the OS at this point

[-]

Unusual_Pride_6480@reddit

Very surprised by the results to be honest