Interesting Ryzen vs Intel benchmark results thanks to Linux

Processors are not comparable, but it clearly shows there is something wrong with Windows regardless of AMD saying Windows is fine.

Intel CPU shows to be almost same score on both OS's while Ryzen gets 5000 extra points on Linux. So yes Microsoft needs to fix something in the Windows Kernel...

Ryzen 1800x
Geek Bench on linux score 27138
Geek Bench on Windows 10 score 22722

Intel Xeon E3-1241:
Geek Bench on linux score 14812
Geek Bench on Windows 10 score 14445

Original source:

I wish someone could do the same with the intel 7700K and 79XXK. And post the results.

What do you guys think?


I guess here they are:

7700K :

I wonder why we don't see big content creators doing this test?

Level1tech could be the first one :)

so linux is actually optimized to support a new product before windows is
Uhhh. Is this surprising or something? Kinda goes with the new product thing. Lol. And windows has such a great track record of supporting new products

Yes. Linux being ahead on hardware support to this degree is a pretty novel development.


Is it really that surprising though?
I mean it takes windows days if not weeks to fix security bugs; let alone support new Uarches.

The point being AMD said in their press release that Windows was fine.... WTF! They are hurting themselves or just don't want to throw their business partner under the bus. I wonder how the benchmarks would have looked with Windows patches, proper memory speed support and other bios updates.


Tbh, I think the whole sorta denial thing is just desperate hopes for 'recommended' deals and whatever the fuck helps in terms of business. Windows is CLEARLY fucked, but I think it's a blind eye situation.


It's 'ยต'

AMD is the first company since Sun Microsystems to lean in this hard on OSS, so yes, in the wider context of the industry, it is.

Not everyone has a fancy keyboard lol

And meh? This topic devolves real fast anyway. Oh linux. Ohhh windows.
Yeah? And?
It will be fixed; or won't.

What have I said that indicated a preference? I just think it's interesting that AMD is leaning in this hard on what is considered in business and enterprise to be a necessary 'evil.'

It has wider implications in industry. if your application performs 10% better on a linux stack, hiring that smug unix neckbeard may suddenly be a lot more attractive. It adds flavor and intrigue to their server launch offering.

Not you lol...

And yeah it may be an issue but we need all the testing with as many different people rigs, etc to see if this isn't just a freak score or is something bigger.

There is a Windows Creators release coming out in April, whatever that means, is it an official release? or some tech beta?

The way Windows Releases work is as follows:
Codebase freezes 3 months prior to release so that OEM's can optimize their hardware and drivers.
Only critical security patches are accepted, no more feature work as the ""BAR"" for code fixes is very high.
Code gets handed over to WinSE team, and windows devs/pm's start planning windows next.

WinSE wont fix something as big as Kernel Scheduling.... so unless Microsoft was aware of this issue 3 months ago, then folks will have to wait another year.... Unless AMD puts pressure on Microsoft and have the Windows dev team backport a fix of this cale to the stable version of Windows.

At least this is how it worked Windows XP, Windows Vista, 8, 8.1. I don't know if they are still driving development in this way.

were you part of the QA team that was fired when they decided to make users guinea pigs?

Yes and no, I could have moved to another team from the disappearing Windows 10 phone team we have just moved into like chess peaces. I had 2 months to look, 2 moths I spend at home watching TV :) I was leaving after a planned surgery and I got to leave with 6 months worth of paycheck :) were as medical leave plus leaving would have given me only 3 months at 70% salary. So I goy lucky... timing was perfect..

During my time there I worked as an SDET and also SDE, all the time on the Windows team.

1 Like

JayzTwoCents is doing a 30day challenge using 1800x as his main productivity driver.

From the same reddit post:

Geekbench results are not comparable between platforms unfortunately. The linux one is compiled using GCC the Windows one is compiled using MSVC. The MSVC compiler does not perform auto-vectorization for SIMD instructions, while GCC does and that's just the beginning of it. As far as compilers go MSVC is by far the slowest of them all, and why people who use blender on windows often compile it on their own using GCC and cygwin.


Well ? Innovation was never windows strong point :) Coming from the software side of things. The only focus windows has is stealing innovation not creating it..........................dig

well, that puts things into context.

What competent multiplat benchmark does shit like this?

Consider that not too smart. So early on.

Keep reading the thread and this claim is debunked by showing intel performing the same on both platforms if the statement was true, it would show similar results as with Ryzen.