Interesting Ryzen vs Intel benchmark results thanks to Linux

to a point I can agree with that.

`

Use the right tool for the right job. For example For a small to mid size company Windows Server is cheaper than Linux because of the total cost of ownership.

For a large corporation like google, Linux is a bargain because its free and they can deploy Millions of machines and can afford the technical expertise associated with the platform.

For example Visual Studio is Expensive but it works. compare it to some Linux editors, the amount of time wasted pays VS many times over (dev cost). For example have an inexperience developer setup a PHP editor with intelisence and see how many days it takes. days x hourly rate + extra lack of features > 1000.

However unless VS can work with all the open source frameworks its irrelevant in todays ecosystem (Except ASP.net core 1.0).

Everything is relative.

What is you are a lawyer, Windwows + Office would give you a piece of mind that at least you would be able to open other lawyers contracts, documents etc.. licenses are irrelevant.

And now we have windows 10 + cortana spying on you.... What is the price on that... for some high, for other irrelevant.

1 Like

Certainly no good answer to basic standard in the os world. The first one that does it, owns the world. I personally lean towards completely open source becoming mainstream.

gonna have to be permissive to persuade the business community. GPL is a cancer in that regard.

Hard thing to do in a world of pay walls. Needs to be a body that says this good for everyone. I have a problem with this is so special no one else can do it but for Christ sakes, all the basic crap. Needs to end. The bar has changed and windows :)

no problem....... since I can not edit :)

BSDs are slowly spreading everywhere that the GPL dissuades people. I don't see a world where the GPL ecosystem stays the currency of enterprise at this rate.

Throw around so ideals. It would spark true innovation verses the constant reiteration of the same old crap.

Not trying to be mean but is english your first language?

can not edit and I am doing about 15 things at the same time :) You are being a jerk on a forum that hosts people across the world.

That's why I asked. I figured I'd give you the benefit of the doubt.

Like I said, no judgement, I was just wondering given your sentence structure. I've had plenty of conversational to fluent ESL friends that write similarly.

I have actually been comparing the sub scores of the two tests.

Single core scores are just about identical. It all goes wrong with the multi core scores but it is not a wash.

Windows has better Crypto and memory scores but fails in the integer and floating point calculations.

That certainly could be due to compiler optimization, or at least lack of optimization

4 Likes

I have no creditability here. So chuck rocks all you want. :)~

this also operates on the assumption that the two operating systems handle all operations similarly assuming all other variables equal. Too much complexity to draw any conclusions for sure imo.

Geekbench is already known to be completely inconsistent across less compatible µarches, why assume it's consistent across these just because they share instruction sets?

@Freaksmacker Again, no malice here. Question was from a place of curiosity. You write very similarly to some of my naturalizing Korean friends.

I do some Korean mmo's :)

For finding why the scores are different.

1 Like

Called it
...kinda sorta

I guess all it takes for Linux to succeed is for Microsoft to do nothing :)

Again, yes and no... yes they are different OS.. but look at the Ryzen delta, vs the intel no delta.... you could say MSVC is not optimized for ryzen, well compile win geekbench with gcc...

a few of the arguments in that thread are weak..... This is not about specifics this is about having 2 machines one intel and one amd, and running the same tests on windows and linux.

The intel one shows consistent results across both systems.
The ryzen shows a big delta.

Compiles don't generate scheduling code. If there were difference on the compilers optimizations they would show on the intel platform as well.

1 Like

Assuming that the compiler doesn't do one thing when it detects Intel and something else when it detects not-intel...

...now that would be rude, but has happened before!

Used to be that the Linux compiler was trailing compilers like Intel C Compiler by a whole lot. Not so much anymore. GCC has gotten a lot better, and so has libc. Clang+LLVM is becoming a powerhouse, Clang is doing good progress even on Windows. I wouldn't surprise me one bit if GCC/Clang has more AMD-friendly optimizations down than any other compiler.

There are certainly differences due compiler optimizations. Until you analyze both compilers, their options, automatic optimizations and the resulting binaries, you simply don't know. A thing that usually differs a lot between Linux and Windows is that on the latter you can have very old binaries being distributed that were compiled ages ago (like the CPU-Z bench), while on Linux you usually have source and compilers at hand at can just compile your own binary just like that. It is not hard to see witch model is vastly superior. For example, The Stilt compiled the Euler3D benchmark with a modern compiler, got kinda different results compared to some others that used a kinda old, kinda not optimized binary.

1 Like