Intel Iris Pro Linux Graphics Yield Some Wins Against Windows
Phoronix: Intel Iris Pro Linux Graphics Yield Some Wins Against Windows
For the past few days at Phoronix we have begun looking extensively at the Intel Iris Pro 5200 graphics under Linux, since receiving the System76 Galago UltraPro. The Iris Pro 5200 are the new high-end Intel Haswell graphics that have 128MB of embedded video RAM on the die, which should yield a nice performance boost when properly implemented within the Intel Linux driver. Already our testing has found the Iris Pro performance on Linux has doubled with open-source driver improvements since Haswell's launch. Now we're out today with our first Intel Iris Pro OpenGL gaming benchmarks between Ubuntu Linux and Microsoft Windows 8 for this Intel Core i7 Ultrabook.
Even when it was behind, Linux was pretty close in most of these with the new drivers. And considering where we were just recently, this is a dramatic improvement! I only hope it keeps approaching a graceful and powerful level of performance as development continues.
Is there another article already that compares Iris Pro to Ivy Bridge?
I just come here to put out one suggestion. Given that Steam is on Linux for a while, can we start using some games on Steam as benchmarks? Just as many Windows reviews do? That could both provide more useful information for the Linux gaming crowd, and foster grounds of improvement for real games that many people play. (I am not saying current benchmarks are useless, but there is definitely a disconnection between Windows reviews and Linux ones. Using common and popular games could significantly reduce this gap.)
Then whenever someone doubts the usability of Linux as a gaming platform, we can simply point it to these benchmark: this is the performance you know in Windows, and Linux is better/same/slightly slower, etc.
Last edited by wujj123456; 09-19-2013 at 02:48 AM.
+1 I was going to ask the same. Michael has said in the past that he has had some issues benchmarking steam games, but we really need it, at least as a baseline. Does the source engine (half-life2 / portal / dota / cs / left4dead) compare to openarena/xonotic and has better performance, or is it like unigine and has half the performance?
Originally Posted by wujj123456
How about Serious Sam 3? Natural Selection 2? Brutal Legend? Killing Floor? Dungeon Defenders?
Especially since sometimes developers don't test them and Michael ends up uncovering issues in his testing and raising some eyebrows. We really need testing on those titles.
Not all of these propose reproducable benchmarking.
Originally Posted by [Knuckles]
That's incorrect. Every game with Unreal engine (Killing Floor uses a 2.5 variant) has a benchmark mode. Try:
Or with your own demos:
I dont have got DD installed and it crashed very often when i tried it, but if you could record demos it should work similar. Btw. it would work with an offline KF install as well (if you used it as trial before it still can be used for solo mode/offline).
That still relies upon Steam where the game could be updated without knowing and not for sure at the same version as others.
Originally Posted by Kano
So I see the issues you are putting forward. That's doesn't relieve the curiosity of your readers (including myself) to know how these games perform! I mean they don't have to be part of your regular regimine of game benchmarks, but I would like to see these games (including some other games that have been recently released on linux) be included in the more extensive reviews. You could include a disclaimer with benchmarks for those games.
Originally Posted by Michael
k thnx bai.
Show us in benchmark min/max FPS because avg. nothing say about preformance. If Ubuntu wins in OpenArena with Windows, and you say it is very good than I have no more comments about you test procedure ...