Intel Core i9 11900K: Five Linux Distros Show Sizable Lead Over Windows 11 (phoronix.com) 82
Phoronix: Now that Windows 11 has been out as stable and the initial round of updates coming out, I've been running fresh Windows 11 vs. Linux benchmarks for seeing how Microsoft's latest operating system release compares to the fresh batch of Linux distributions. First up is the fresh look at the Windows 11 vs. Linux performance on an Intel Core i9 11900K Rocket Lake system. Microsoft Windows 11 Pro with all stable updates as of 18 October was used for this round of benchmarking on Intel Rocket Lake. The Windows 11 performance was being compared to all of the latest prominent Linux distributions, including: Ubuntu 20.04.3 LTS, Ubuntu 21.10, Arch Linux (latest rolling), Fedora Workstation 35, Clear Linux 35150. All the testing was done on the same Intel Core i9 11900K test system at stock speeds (any frequency differences reported in the system table come down to how the information is exposed by the OS, i.e. base or turbo reporting) with 2 x 16GB DDR4-3200 memory, 2TB Corsair Force MP600 NVMe solid-state drive, and an AMD Radeon VII graphics card.
Each operating system was cleanly installed and then run at its OS default settings for seeing how the out-of-the-box OS performance compares for these five Linux distributions to Microsoft Windows 11 Pro. But for the TLDR version... Out of 44 tests run across all six operating systems, Windows 11 had just three wins on this Core i9 11900K system. Meanwhile Intel's own Clear Linux platform easily dominated with coming in first place 75% of the time followed by Fedora Workstation 35 in second place with first place finishes 9% of the time. The geometric mean for all 44 tests showed Linux clearly in front of Windows 11 for this current-generation Intel platform. Ubuntu / Arch / Fedora were about 11% faster overall than Windows 11 Pro on this system. Meanwhile, Clear Linux was about 18% faster than Windows 11 and enjoyed about 5% better performance overall than the other Linux distributions.
Each operating system was cleanly installed and then run at its OS default settings for seeing how the out-of-the-box OS performance compares for these five Linux distributions to Microsoft Windows 11 Pro. But for the TLDR version... Out of 44 tests run across all six operating systems, Windows 11 had just three wins on this Core i9 11900K system. Meanwhile Intel's own Clear Linux platform easily dominated with coming in first place 75% of the time followed by Fedora Workstation 35 in second place with first place finishes 9% of the time. The geometric mean for all 44 tests showed Linux clearly in front of Windows 11 for this current-generation Intel platform. Ubuntu / Arch / Fedora were about 11% faster overall than Windows 11 Pro on this system. Meanwhile, Clear Linux was about 18% faster than Windows 11 and enjoyed about 5% better performance overall than the other Linux distributions.
No end user cares if Linux is 18% faster (Score:5, Insightful)
They care if Linux supports the applications and hardware they want to use.
Too often, this is lacking.
Re: No end user cares if Linux is 18% faster (Score:5, Insightful)
Re: No end user cares if Linux is 18% faster (Score:3)
Re: (Score:3)
That's the key point. But games are designed around Windows, it's APIs, the DirectX version that's in vogue, etc. Most games developers won't even bother with OSX and it's a much larger share of the market than desktop Linux. Video card support is often proprietary. Yes, a few decent games will run on Linux, but it's still just a tiny handful.
Re: (Score:2)
Well it's because Microsoft works with hardware vendors to produce an API for their newest technologies and roll it out to Windows users. I'm a big advocate for Vulkan and OpenGL but there's no getting around the fact that if you want to be on the cutting edge then DirectX is going to get the new hardware features first. By the time vendor extensions are ratified and implemented by all IHVs developers have already had the opportunity to make extensive use of them via Microsoft's APIs.
For example by the time
Re: (Score:2)
If games run better on Linux, developers will develop for it and users will buy hardware that works on it and so OEMs will make drivers
That's a big "if", what's that got to do with these benchmarks? Many (but not all) of these are CPU rendering benchmarks, things that are generally GPU-accelerated or in the case of games always GPU-accelerated, nobody is doing CPU rendering of games. If you can consistently get say 15-20% higher performance in games then I would agree with you, but that's not what these benchmarks are showing.
Even if you're doing CPU-only rendering (why aren't you using LuxCore on the GPU?) if you look at the benchmarks, t
Re: (Score:2)
Re: (Score:2)
It is all a push towards cloud computing. Microsoft is slowly deprecating desktop and as network becomes more accessible and ubiquitous we are looking at the advent of thin clients running "cloud native" applications like office 365.
The only real challenge remains gaming. Somehow it is still remaining to be cracked. I suppose it may take another 10 years or so for ray tracing to become availalbe over internet, and then it will be game over for PC.
Re: No end user cares if Linux is 18% faster (Score:1)
download caps will make cloud only hard and pricei (Score:2)
download caps will make cloud only hard and pricing.
let's see games at full cost and then service fees.
game saver account suspension fee $5/mo (you don't do want to lose the $1000's in games that you paided for right?)
legacy mode (for games older then XX years) $5-$10/mo
music rights fees (yes that music licensed in the games does not cover steaming)
$10/mo 4K upgrade fee
$10-$15/mo unlimited usage (game in prime time with out getting kicked out)
$5-$10/mo for XX more storage (for saves, mods and more)
$10/mo mu
Re: (Score:3)
Which Windows is fast forgetting. With mandatory arbitrary CPU requirments, TPM, etc
It's doing nothing of the sort. People upgrade OSes with their PCs. That's how it always has been except for a brief quirk of history where people were getting Windows 10 updates every 6 months which did nothing for them. No one cares.
Your new computer will run Windows 11 *guaranteed*. Not so guaranteed is how well it will run Linux because it's 2021 and we're still in a world where vendors ship broken binary blobs, missing drivers (thank god my printer is okay thinking it's a different model), and non ACPI
Re: (Score:2)
It does. Tom's Hardware did a big writeup on how to install Windows 11 to a raspberry pi 4
Re: (Score:2)
Hmm, that's funny cuz the last computer I bought was a Raspberry Pi 4 and I'm pretty sure it won't run Windows 11.
Why are you pretty sure? Is it ignorance? https://www.notebookcheck.net/... [notebookcheck.net]
Re: (Score:2)
Of course, in defense of the 'ACPI sleep' situation, ACPI S3 is pretty superfluous now. The OS has much more fine grained access to all the resources and can get to S3 level function/power consumption without actually having the firmware do S3. Better, the OS can elect to have connected standby, to have almost as low as S3, but the occasional window for things like networking to process messages.
I haven't had bad luck with hardware support in Linux in quite a long time, but maybe I just am lucky in the mo
Re: (Score:2)
I haven't had bad luck with hardware support in Linux in quite a long time, but maybe I just am lucky in the models I pick.
The hardware is important but it also comes down to what you actually mean by Linux. One thing these benchmarks do demonstrate pretty clearly is that Linux isn't consistent. In some benchmarks Linux beats Windows, in some benchmarks Windows beats Linux, in some benchmarks Windows both beats Linux and loses to Linux, and in some benchmarks Linux beats Windows by the same amount that Linux loses to Linux.
It's not simply a matter of 'switching to Linux' because whether it will be better or not - assuming you c
Re:No end user cares if Linux is 18% faster (Score:4, Informative)
Linux supports all the Windows stuff it can.
The issue isn't what Linux supports. The issue is that developers support Windows, and in some cases (games and CAD stuff mostly) jump through some hoops to disable Linux support for their products.
But again: It's not "Linux supports a program". Windows doesn't "support" shit. What happens is, developers support Windows.
And of course this is just an issue with some sections of desktop software and occasionally some important niche thing compiled for Windows XP and a special serial device or whatever.
Re: (Score:2)
But again: It's not "Linux supports a program". Windows doesn't "support" shit. What happens is, developers support Windows.
Well yes it's not that "Windows" supports something or "Linux" supports something but it's rather that Microsoft produces APIs, tools and documentation for developers and, if you're a AAA studio, you'll find it's pretty easy to get Microsoft and hardware vendors together to help optimize and troubleshoot things. Microsoft does indeed support, in particular, AAA game studios very well just as Sony does for game studios on their Playstation platform.
That kind of scenario doesn't really exist in the Linux envi
Re:No end user cares if Linux is 18% faster (Score:5, Interesting)
Re: (Score:2)
Re: No end user cares if Linux is 18% faster (Score:1)
Re: (Score:3)
>"A cron job for a weekly reboot"
Why? I mean, I can see that if you were doing that for some kernel update, but no Linux machine I have ever used required regular rebooting. And I have run many hundreds, of many distros, and many types of machines, across many tasks and types of users. Generally they NEVER need rebooting unless I want to apply a kernel-related update.
In any case, I set up a Linux machine for my Mom many years ago and she has no problems with it at all (other than the problems she woul
Re: (Score:2)
>"A cron job for a weekly reboot"
Why?
It didn't "need" it. My mother was one of those people who would leave every tab open in the browser eating RAM. It was the easiest way to just give it a weekly fresh start. I would often walk up to her computer and find a browser with 60 or more tabs open, on a mid end laptop from 7ish years ago. I suppose I could have croned a browser restart, but it was just easier to tell her it was part of the update process. She was extremely non technical. So much so, that she kept a book with notes on the exact pro
Re: (Score:2)
>"It didn't "need" it. My mother was one of those people who would leave every tab open in the browser eating RAM. "
OK, that makes sense :)
I have the opposite problem with my mom- she can't seem to handle the concept of tabs and keeps closing Firefox over and over again, instead of switching to a new tab. But, whatever... works for her!
Re: (Score:3, Insightful)
No end user cares if Linux is 18% faster
Gamers and power-users sure as fuck do; an 18% difference in single-threaded performance is fucking astronomical.
They care if Linux supports the applications and hardware they want to use.
Too often, this is lacking.
How disingenuous and/or utterly ignorant of you to lump app and hardware support together - on average, Linux has better hardware support than the last twenty years of Windows releases put together.
Any other stupidity needing correcting??
Re: No end user cares if Linux is 18% faster (Score:2)
Re: (Score:2)
Re:No end user cares if Linux is 18% faster (Score:4, Interesting)
But the slowness of Windows that drives me nuts isn't in its max performance, IOPS and so on... it's the cruft. Windows always has 10 things running that I didn't ask for, mostly running under some multi-purpose OS service so I can't even tell who it is or what it is. And then disk space, Windows just gradually expands to take over however much there is.
Re: (Score:3)
Hardware? It's been a very, VERY long time since I ran across hardware that's not supported in Linux. If it's server hardware, it is almost always supported since server hardware folks know the breadth of Linux usage in server world. For desktops, it's either supported out of the box, by the vendor, or someone wrote something that supports it.
P.S. This is the part where you google "hardware that's not supported in Linux" just to prove me wrong.
Re: (Score:3)
I care. I run Windows 10 in a VM and it sucks royally there. That environment proves all the crap that Windows does behind the scenes, because even if I am not moving the mouse or typing anything, that OS is busy busy busy doing crap. Indexing your files, compiling .net bytecode for "improved user experience", preloading applications, sending data back to the mothership, etc. My mouse will stutter as I move it across the screen and I've already given it 8GB. Linux in the same VM is amazingly fast and res
Re: (Score:2)
Re: (Score:2)
o Give Windows VM a host-only adapter
o Use ssh and port-forwarding to connect it to a raspberry pi (or other VM) running squid and pi-hole for Internet connectivity
o Log everything it's connecting to
o Block as appropriate using the pihole's web console
o Profit(?)
o Also make sure you have whatever guest additions installed in the VM, should fix the mouse - and ensure any hardware-specific mouse drivers or apps are uninstalled or disabled
Re: (Score:2)
How kind of you to speak for all users.
Re: (Score:1)
Re: (Score:2)
Intel's Clear Linux (Score:3)
Re:Intel's Clear Linux (Score:5, Interesting)
Clear Linux OS [wikipedia.org] is highly optimized for Intel (and, due to near-identical architecture, AMD) CPUs. I assume (but can't find definitive proof that) Clear Linux OS uses the Intel C++ Compiler [wikipedia.org] rather than GCC. I haven't paid attention to this in many years, but Intel's proprietary compiler has at least historically produced radically more performant outputs than GNU's free/open-source GCC because Intel's primary design focus is performance on Intel chips while GCC is chiefly concerned with portability across ~all chips).
Re: (Score:2)
I should really try Clear Linux. I've been trying to get my DAW performance under Linux to be consistent and keep getting momentary millisecond pauses in recording mode. Which is pretty shit when you're a minute and some change into the perfect take after biffing the last five tries.
Re: (Score:2)
This sounds like a scheduler issue? Have you looked into the real-time support stuff?
Re: (Score:2)
Yup. Audio processing has full access in the scheduler and everything, literally EVERYTHING else is deprioritized. And this is across something like four different distributions using two different real time kernel options and one of the low-latency kernels. I almost wonder if it's not something with my interface, but that same interface works fine under Mac OS or Windows, Windows on the exact same hardware, so it's got me pretty lost.
Re: (Score:2)
I'd like to try it, but it doesn't include Mate desktop which I prefer. However, it supports the new Xfce4 DE, which is almost as good, so I might give it a spin. Gnome desktop is bordering ridiculous. I don't know why they chose it as the default DE. I'm using Debian because I like to have a nice set of choices for DEs, and I always felt like different DEs are Linux' strength, not a weakness like some people claim.
Re: (Score:2)
Re: (Score:2)
Thank you so much, mate. Appreciated! d= :)
Re: (Score:3)
Intel's compilers produce excellent code, but it's code that deliberately sabotages all non-Intel processors.
Intel compilers generate code that checks the CPU ID value for the user's processor, with high-performance code under the Intel side of the check and correct but horribly slow code under the other side.
It is because of this that I try very hard to not buy anything from Intel. I've been buying AMD processors for years... even during the time where Intel chips were genuinely better than AMD chips. No
Re: (Score:2)
You must have little experience with Linux to even be acting like a silly troll as you are trying to do.
No need (Score:2)
And for anyone else, that doesn't need a power-hungry workstation, power consumption, longevity, cost etc is way more important. The other important power-user market - the gamer - is typically more worried about its GPU and OS compatibility for favorite games.
We are long past the days that 'power tasks' are a an issue. Snappiness of system is one, but has very little relation with any benchmark and is only dependent on bloatware and script-loaded websites aka surfing without ad blocker. Everyone enjoys a g
Re: (Score:2)
If we are talking compute power, remember that mobile phones the last couple of years are more powerful than most laptops/desktops folks were using 10 years ago, so yeah ... they are small but you can do a lot more than Tiktoking and whatever it is these zoomers do. ;)
Re: (Score:2)
Well, for my games, I care if they run. Most won't run on Linux without effort at getting WINE tweaked just right. Quite a large number just won't run even with WINE. And those that do run don't take advantage of my video card (such as it runs fine but with zero anti-aliasing). I don't care about new games, I have 20+ years worth of games I still play now and then. I'm happy with lower game performance, I think 30fps is great, but on Linux if it's 0fps it won't cut it :-)
Re: No need (Score:1)
Re: (Score:2)
No-one, except a few professionals, need a workstation. To prove that point: most people are happy to use a smart phone as primary computing device.
Back when I first went onto Slashdot, the talk then was "desktop PCs are ok, but real users need workstations". In the intervening years, it is likely that the workstation market is still the same size, but the desktop PC market will have grown by maybe 3 orders of magnitude.
In 2020, there were around 275 million PCs (of all kinds - Windows, MacOS and a bit of Linux) sold around the world. By comparison, there were around 1,375 million mobile phones sold (iOS and Android) and 160 million tablet devices (iP
Re: (Score:2)
From the Summary (and maybe the article as well) (Score:3)
>>Now that Windows 11 has been out as stable...
What?
Re: (Score:2)
One of the problems of all the pre-release, leak, and absurd requirements talks has been that the official "stable" release of win11 went by without much notice. It hasn't been automatically pushed to many win10 users yet as microsoft is currently doing the responsible thing and deploying in small batches and only to hardware that's expected to be fully compatible.
Of course, if you don't want to wait, or don't want a TPM based backdoor on your system you can still get a script to download a patched ISO wit
Information abuse. (Score:1)
Linux is beating Windows for a particular processor ( which is a rather high end processor ) is really misleading as it is exaggerating the success of Linux and making Windows 11 seem like a failure.
On a construction site, the number of bulldozers outnumber the number of Toyota Camrys, So Toyota must be failing at making cars people want, and everyone wants a Caterpillar.
The Intel i7 is overboard for most consumers, the i9 even more so. So people will often have specialized tasks that they may want to us
Re: (Score:3)
is really misleading as it is exaggerating the success of Linux and making Windows 11 seem like a failure.
I can guarantee you that the number of people whom were affected by this news and opted for Linux instead of Windows because of it is precisely zero.
Re: (Score:3)
I'm building a machine at home where I want to run Linux, FreeBSD, and Windows (for various bits of software each supports well). One of them needs to run on the bare metal and run VMs for the others. This is a good argument not to use Windows as the bare metal OS.
Re: (Score:2)
Re: (Score:2)
> If it's performance you're looking for
Or just not needlessly wasting it in a case where I now know better
Re: (Score:3)
Agreed. It's interesting information, but not really all that relevant for day to day computing needs of most people. It's kind of the same deal when talking about how Python now "beats" all other languages in terms of popularity of web searches. There are a lot of different metrics which may make a particular language (or desktop OS) the best choice.
One of Linux's strengths is its open-source nature, meaning you can customize it and tune if for specific characteristics. I think this just highlights one
I Knew It! (Score:4, Funny)
Re: (Score:3)
No no no, you have it backwards. You want Linux at the lowest level, so boot Linux but run Windows in a VM so you get all the performance of Linux and all the software of Windows. When I need to bang out a movie script quickly, I run DOSBox + Wordstar on top of that.
What's the LOAD? (Score:2)
Or does this measure the respective operating systems sitting around on the desktop doing nothing? And if so, what value is there in benchgmarking an operating system with no load?
Re: (Score:2)
You could go to the link in above and actually look. Whodathunkit.
https://www.phoronix.com/scan.... [phoronix.com]
Re: (Score:2)
I didn't look far enough... Sorry.
Not a fair comparison! (Score:3)
This isn't a fair comparison at all! Linux developers actually care about making Linux better! ;)
Containerization of Apps? (Score:2)
Fishy results (Score:1)
So why would you use the geometric mean to compare clock speeds when all the results should be on the same scale? You should use the arithmetic mean.
Also, the results imply some variance (Windows won certain results, and not others). If that's true, and the results came from an actual experiment with multiple repeats, why not report the statistical significance?
It seems like the "researchers" went out of their way to make Windows look bad, and don't even report the full results. What's the point?
For some, it doesn't matter (Score:2)
For people who only need stuff available on linux, it's a great alternative
For others who need stuff that only runs on windows, it doesn't matter
Other than email and web browsing, everything I do requires stuff that only runs on windows
And please, don't tell me that there are linux alternatives. When I need Solidworks, I need Solidworks, not a less capable alternative
I was impressed with antiX Linux (Score:2)
If your looking for performance (Score:1)
If your looking for performance, why aren't you running AMD? clearly a paid promotion by Intel.
Certified fresh. (Score:2)
It's apples to oranges (Score:2)
Actually, Apple products are not involved, so it's more like penguins to ... lemons?
Two "don't cares" for the price of one (Score:2)
Windows and Intel? They've both had more strikes than I'm going to give them. Neither will ever sit on hardware I chose again. Both are insecure, unreliable piles of shit.