A Look At The Windows vs. Linux Scaling Performance Up To 64 Threads With The AMD 2990WX

[DJ_Reflector]

Hi all,
Can I ask you (as specialists) ? Why processors Bus Speed 100MHz? 30 years and: 100MHz? What for all of this GHZ and Cores if bus speed so low?

[comp]

Now we see 2990WX benchmarks only for 32 GB single-rank RAM configuration.
But additional RAM ranks and banks and additional RAM volume can improve the performance.
I'd like to see benchmark results of 2990X for different RAM configurations:
- 32 GB single-rank
- 32 GB dual-rank
- 64 GB dual-rank

[boboviz]

The Windows Server licensing is kinda complex but as far as I understand it from just looking at the Windows Server 2016 Licensing sheet you need to purchase "Additional core license grants" for "Servers with > 8 cores per processor and > 16 cores per server". No mention of RAM so it looks to be core related only.

Windows server licensing is interested much more in socket on server than core per cpu.
Anyway this has nothing to do with performances

[boboviz]

I’d be interesting to see such a comparison, on the other hand if the balanced power plan really causes such performance issue on Windows I’d still say that this is a serious Windows defect.

Anyway, would anybody chip in for a benchmark of various CPU governors and its effect on the TR2? I definitely would.

I think thata "high performance" profile must be the default profile in a server....

[andrei_me]

Since when are premium-only articles a thing? I wasn't aware of that.

Since march? Don't remember exactly when but it is not a premium exclusive article, it is a premium first, so a couple days later the article is available to everyone

[andrei_me]

Was funny a few weeks ago someone on Twitter was suggesting to other review sites to checkout PTS... I think it was Gamer Nexus that responded that their tools were far superior to PTS but too good that they wouldn't want to release them publicly along with some other nonsense.

I guess AnAndTech also does some automation of their benchmark runs now, but that is all internal too and from the sounds of it just a collection of scripts and a lot of other comments on it of nonsense.

Yeah, PTS is the way to go, it is open-source, exists for a LONG time, it is test proven, big companies like Intel uses it, developers uses it, the guys improving POWER used it, and others sites claim that their in-house testing system is better!?

[F.Ultra]

Also, don't you also have to pay a lot extra when you have too many cores and too much RAM? Windows may not actually be using the hardware to its full potential if it's not licensed correctly. If I recall correctly, this was a potential issue on Windows Server 2012. I haven't touched Windows servers since that version though!

Just reminding you, Microsoft sucks.

The Windows Server licensing is kinda complex but as far as I understand it from just looking at the Windows Server 2016 Licensing sheet you need to purchase "Additional core license grants" for "Servers with > 8 cores per processor and > 16 cores per server". No mention of RAM so it looks to be core related only.

[nslay]

Was Windows Server's power management set to max performance before testing? By default it's set to run on quite conservative CPU clocks if at all possible.

If "yes" then Microsoft has quite a lot to think about.
If "No", then results would not reflect the performance differences correctly.

Also, don't you also have to pay a lot extra when you have too many cores and too much RAM? Windows may not actually be using the hardware to its full potential if it's not licensed correctly. If I recall correctly, this was a potential issue on Windows Server 2012. I haven't touched Windows servers since that version though!

Just reminding you, Microsoft sucks.

[F.Ultra]

For example, I work with an vendor application stack that is written and compiled for Windows first, then they go back and recompile it for Linux. For years we ran this app stack on Linux and took many months to tune it. When the Linux OS version became obsolete, we went back and worked with the vendor on a target OS plan. That is when they informed us that their stack was compiled Windows first, Linux second, and that the Windows version ran 10-15% faster because of that. The "why" was never revealed due to internal policies and we wanted that 15%, so we switched the stack to Windows.

This is why I am looking "under the covers" to see where people are getting their material. Because PTS is 100% open source, I can examine it all.

The "why" on your vendor stack is not hidden by internal politics. It's simply a result of your vendor writing and optimizing for Windows while their Linux version is a simply a compile target. The reverse happens with the software that my company produces, since we write, test and run everything on Linux so naturally it all is optimized for running on Linux while the Windows version is a compile target and thus performance is probably not optimal on Windows.

This is also why game ports from Windows will behave worse on Linux even when they use the same graphics stack (OpenGL/Vulkan) since everything else is optimized for Windows. Very few (I've heard of exactly no one) will invest the resources needed to treat every supported platform as first class citizens (and the #ifdef hell that would result in).

[F.Ultra]

They do. Windows doesn't scale. At least when defaults are benchmarked. And don't even think it has better response time than Linux. Stock exchange results speak for themselves. Only morons pay and use slower, less secure OS for serious tasks.

No they don't, while everything you write is true it has no correlation with the Netcraft numbers since none of the listed sites there use single servers. I.e it's not the OS that is making these sites stable, it's their High Availability setup with multiple load-balancers and so on that explains 100% of these numbers.