Here's Why Radeon Graphics Are Faster On Linux 3.12

[chrisb]

You can use that, but don't bother me with the benchmark results, because as a driver developer I am not interested. It's not only the graphics driver being benchmarked here, it's also cpufreq. And who knows what cpufreq will do next time.

You, as a GPU driver developer, may not personally care about low GPU performance results due to the CPU, but AMD as a corporation should care, because it negatively impacts end-user experience of AMD products.

[log0]

The testing is about the default for the distribution that most people utilize.

Even if the default is completely bananas? I mean 90% performance difference due to default governor setting is crazy.

[log0]

You, as a GPU driver developer, may not personally care about low GPU performance results due to the CPU, but AMD as a corporation should care, because it negatively impacts end-user experience of AMD products.

So what do you expect AMD to do? Provide custom governors for all the cpus out there, that might be used with AMD graphics cards?

Imho this is a linux kernel issue and should be fixed there, which is what happening already I assume.

[Michael]

Even if the default is completely bananas? I mean 90% performance difference due to default governor setting is crazy.

Then it should be fixed or the default setting changed within the kernel.

[jonnor]

Even if the default is completely bananas? I mean 90% performance difference due to default governor setting is crazy.

Especially when the default is completely bananas. Because the default then needs to be changed, or the default behavior simply fixed.

[Figueiredo]

You, as a GPU driver developer, may not personally care about low GPU performance results due to the CPU, but AMD as a corporation should care, because it negatively impacts end-user experience of AMD products.

If I'm not mistaken, Brigdman once said that he or AMD sees the open drivers as the more adequate, long term, to the consumers. Thus, I assume that, long term, the plan would be to pretty much dump fglrx for radeon cards and apus on linux. If this is right, I would expect AMD to work closely with target distro in order to iron out these naty little bugs. If not for Ubuntu, than at least for SteamOS... Nvidia certainly is doing something similar for their blob.

[GreatEmerald]

Especially when the default is completely bananas. Because the default then needs to be changed, or the default behavior simply fixed.

At the same time, you'll never know whether it's the scaling governor or the graphics drivers that are bananas. You'll just know that something in there is.

[Gps4l]

As I remember developers from Croteam discover that default ondemand governor slow down their engine in Novermber of 2012. They recommend everyone switch to performance governor before starting Serious Sam 3. Hopefully, we probably doesn't need switch to performance governor any more.

Yes they did, they told us to switch off things like cool and quiet in the bios, as I did.

Setting the governor to performance also improved things allot.

(The biggest performance boost came with one of the latest catalyst drivers though.)

AMD phenom II X4 at 3.2 ghz, 4 gig ram, HD5750 1 gig ram. openSUSE 12.3 64bit

[s_j_newbury]

I'm not sure I really buy the default argument. The default it to have FPS rate limited to the display refresh rate, and rightly so. If there's a call to measure performance with games far below the capabilities of the gfx hw such that they are cpu bound, wouldn't a more meaningful benchmark be overall power utilisation? What use case is the in running at multiple 100s fps more than you can see?

[menfin87]

You can use that, but don't bother me with the benchmark results, because as a driver developer I am not interested. It's not only the graphics driver being benchmarked here, it's also cpufreq. And who knows what cpufreq will do next time.

You are not interested as a driver developer ? So if a user complains about performance problems, do you just tell him "I don't see it so there is no problem" or do you maybe try to understand what is going on ? You should be glad that someone is testing an average user configuration to spot problems that would be blamed on you.