Intel DRM-Next Performance Tests With Skylake (Linux 4.8)

[Kano]

@debianxfce

Is there a reason why you can not use PTS directly and show the resulting links? The deb should install fine with Debian testing.

[darkbasic]

Because he knows nothing about benchmarking and he thinks to be way smarter than he is.

[Passso]

Nope, you didn't count the debian testing + xfce magic boost.

If ( /dev/cpuinfo | grep Intel == true ) { SetCpuMhz = 800; } // 0.8Ghz

[Passso]

More academic way is to publish important parameters in English than pointing readers to the source code.

debianxfce goes wild!!!

[dungeon]

Because he knows nothing about benchmarking and he thinks to be way smarter than he is.

He will once figure out that this openarena with custom settings is actually good bandwidth cap benchmark

This benchmark bans all low bandwidth devices like iGPUs (yes if you have iris pro with edram expect more in this bench) and is good to test cards on real world scenario bandwidth... but also drivers on more or less CPU cap too

[franglais125]

Using pts for one game benchmark causes extra work and I am against php because the php code is not human readable and tend to be difficult to maintain in many web projects. Also pts uses non repository (pts openarena version) software that is a security risk.

I would actually be interested in knowing what result you would get. Michael doesn't have many APUs to compare against (and a lot on his plate anyhow).
Would you mind running the test again, using PTS?

Stating the obvious: the reason it's important to use PTS is so that the comparison is apples to apples. If a different version of OpenArena (or engine, etc) is used, then you are not comparing hardware alone, but also different software.

I expect your APU to be faster than whatever the intel iGPU (since no eRAM) can deliver. But the question is: by how much? And this is not a question one can answer if not using the same software. Granted, you are not using exactly the same computer, but at least one big variable (game version) is removed.

Politely, and cheers,
Fran

[Kano]

debianxfce

Best shut up if you don't want to provide a comparable result. Your manually tested ones interest nobody.

[Passso]

I am against php because the php code is not human readable.

if you cannot read php: learn the logo language and move the turtle like I did it when I was 8.

Everyone can progress, hang on! The world is yours!

[fakenmc]

the php code is not human readable

Yet in the past you claimed to look at C/C++ source code and even finding bugs. You know that PHP uses a similar syntax to C/C++, while being much simpler than either, right?

php code (...) tend to be difficult to maintain in many web projects

So, Wikipedia, Facebook or any of the other 250+ million sites using it? What are they? Figments of our imagination?

Also pts uses non repository (pts openarena version) software that is a security risk.

As opposed to using a non-repository custom kernel?

[fakenmc]

It is different thing, device driver source code and phoronix magazine.

Of course it is. Driver source code is much more complex than the PHP code in PTS (or "magazine", or whatever you seem to think it is). Since you can't even read the latter, what can we conclude about your knowledge of the former?

PHP does have cryptic operators

Not even comparable with C/C++. PHP is a high-level language, and is very simple to understand for C/C++ programmers since it shares a similar syntax and similar operators. If anything that seems cryptic in PHP just go to the docs see what a certain function or method does. On the other hand, no docs will help you with cryptic stuff in C/C++, you really have to know what you're doing, e.g. tell me what this code does:

void**(*x) (int &, char**(*)(char*, char**));

Now that is cryptic, and easily seen in non-obfuscated C code. No way you have similar non-obfuscated stuff in PHP.

and usually the code is not object oriented.

It wasn't in 2001. But we're in 2016. Nowadays, most of the PHP API and most of the PHP frameworks are OO and make heavy use of OO design patterns.

I know that php is widely used, but so are many other shit things too.

You've said PHP "tend to be difficult to maintain in many web projects". That is true of any programming language when you don't know what you're doing. The fact that Facebook and Wikipedia exist, and the fact that most websites are developed in PHP, are proof that its not that hard to maintain when you know what you're doing.

Python code is way more readable

Yes it is. Especially when showing simple examples for people who don't know how to program. Since its so simple, what do you make of this "readable" Python code:

lambda itsfun: (lambda q: q(q))(lambda u: itsfun(lambda *args: u(u)(*args)))

python does have better error reporting.

It has nice error reporting indeed. PHP also has nice error reporting, very flexible, maybe more complex to initially setup than Python's approach. However, saying that Python is "better" in this regard is more a matter of opinion than fact. Could you explain in your own words why Python has better error reporting than PHP please? No copy-paste from the web or links to random sites? Thank you.

I did guess that this was coming. Don't you think this is safe: https://www.kernel.org/

You really missed the point didn't you? Let me put it this way, may you'll understand:

1. What has the potential to be more unsafe, an OS kernel or a sandboxed game? What controls every aspect of your PC, including networking and security, the OS kernel or the game?
2. You claim the game used in PTS is unsafe because it is an upstream version (not from distro repo). However, you use an upstream kernel (not from distro repo). Thus you are contradicting yourself, because a kernel is by definition much more unsafe than a sandboxed game.