Latest AMDGPU-PRO Ubuntu Linux Performance vs. NVIDIA, Including The GTX 1080

[bridgman]

That's a logical assumption if you feel a bit lazy. Please excuse me being so inconvenient - I just mean well for you - but I don't think that you can afford to just make assumptions. The Fury isn't represented as well - where should it be? It would really make no sense to put it under R9 200 series while even the HD 8800 series is counted separately - being in use by 1.5 percent??? So more than the HD 7900 and nearly the same like the R9 200?

No worries, I like inconvenient... I spend most of my life being inconvenient to other people at work. More or less my job description...

Good question about the Fury... there is some potential fuzziness about 3xx but not Fury. Will see what I can find out...

[oooverclocker]

Edit2: I deleted what I previously wrote here because it was a bit confusing I guess. I finally noticed that the GPU data is presented differently when you look at the survey for Windows only vs. Win/Mac/Linux combined. This caused me to think that Valve would have changed it. But as the Fury is visible in the Data for Windows only with low percentage(less than I would have believed so it doesn't show up in the combined results) it is pretty sure that the survey is just fine when the 300 series is grouped with the 200 - sadly.. Sorry for the false alarm. It looks like you need to sell many 4XX

[Marc Driftmeyer]

They're not though... look at Luxmark which is more of a real world app, where a non-X Fury is competitive with a 1080. The amdgpu stack is pretty new, don't think anyone has had time to do much performance tuning for synthetic benchmarks yet. I'll ask though...

I expect, as the stack matures that AMD overtakes Nvidia in all OpenCL operations.

[pq1930562]

Michael :

Thanks for doing Linux benchmarks. But why don't your Linux benchmarks include comparisons with Windows performance?

What we really want to see is Linux vs. Windows performance to see if Linux can hold up to Windows.

That would be much more interesting than just seeing the numbers for Linux.

Please also show us the numbers for Windows in direct comparison. In all benchmarks please.

[Michael]

Michael :

Thanks for doing Linux benchmarks. But why don't your Linux benchmarks include comparisons with Windows performance?

What we really want to see is Linux vs. Windows performance to see if Linux can hold up to Windows.

That would be much more interesting than just seeing the numbers for Linux.

Please also show us the numbers for Windows in direct comparison. In all benchmarks please.

Because I don't have enough resources to include the Windows numbers all the time but only once in a great while... You can thank ad-blockers and the 99% of people who opt not to subscribe to Phoronix Premium as among the limitations for not being able to do Windows benchmarks so often.

[pq1930562]

Because I don't have enough resources to include the Windows numbers all the time but only once in a great while... You can thank ad-blockers and the 99% of people who opt not to subscribe to Phoronix Premium as among the limitations for not being able to do Windows benchmarks so often.

Sorry, I do not understand. In http://www.phoronix.com/scan.php?pag...est-Automation you said that test automation would be much easier on Windows. So why would it eat a considerable amount of resources to include the Windows numbers all the time?

Would be nice if you could explain.

[Michael]

Sorry, I do not understand. In http://www.phoronix.com/scan.php?pag...est-Automation you said that test automation would be much easier on Windows. So why would it eat a considerable amount of resources to include the Windows numbers all the time?

Would be nice if you could explain.

Because it's still more time needed before I can publish an article.... It may be not time as much I need to be interacting, but still more time needed to complete an article before I can post it online. Additionally, there's time involved in setting up Windows and maintaining it with updates, drivers, etc. Additionally, many of my SSDs are only 120GB drives. Already when installing a couple Steam games on a clean Ubuntu install, they get mostly filled up and wouldn't even have the room for a Windows install. Not all of my Linux benchmarking test profiles for the different games have also been ported to Windows yet.

[pq1930562]

Because it's still more time needed before I can publish an article.... It may be not time as much I need to be interacting, but still more time needed to complete an article before I can post it online. Additionally, there's time involved in setting up Windows and maintaining it with updates, drivers, etc. Additionally, many of my SSDs are only 120GB drives. Already when installing a couple Steam games on a clean Ubuntu install, they get mostly filled up and wouldn't even have the room for a Windows install. Not all of my Linux benchmarking test profiles for the different games have also been ported to Windows yet.

I see...

Previously you mentioned:

You can thank ad-blockers and the 99% of people who opt not to subscribe to Phoronix Premium as among the limitations for not being able to do Windows benchmarks so often.

But, when including Windows numbers, the articles surely would become more interesting for a broader audience, wouldn't they?

For me it is like this:

I appreciate the benchmarks on Phoronix, but to me they mean nothing more than "Driver 0.2 is faster than driver 0.1".

It's nice to see that drivers on Linux are evolving. But what I really want to know is if the performance on Linux can hold up to the performance of Windows, which the Phoronix benchmarks are not telling me.

Furthermore, I think the Phoronix benchmarks should include more triple A games.

For example: Why do you benchmark Xonotic but not F1 2015 or Tomb Raider for example?

I absolutely do not care about Xonotic, and I would assume I am not alone with this attitude.

So, including Xonotic (or Nexuiz) in the benchmark is an absolute waste of time IMHO.

Your benchmarks would be way more interesting if they would consist of most triple A games and would include Windows comparison benchmarks.

That's what gamers really care about.

Please consider it.

Regards

[Michael]

I see...

Previously you mentioned:



But, when including Windows numbers, the articles surely would become more interesting for a broader audience, wouldn't they?

For me it is like this:

I appreciate the benchmarks on Phoronix, but to me they mean nothing more than "Driver 0.2 is faster than driver 0.1".

It's nice to see that drivers on Linux are evolving. But what I really want to know is if the performance on Linux can hold up to the performance of Windows, which the Phoronix benchmarks are not telling me.

Furthermore, I think the Phoronix benchmarks should include more triple A games.

For example: Why do you benchmark Xonotic but not F1 2015 or Tomb Raider for example?

I absolutely do not care about Xonotic, and I would assume I am not alone with this attitude.

So, including Xonotic (or Nexuiz) in the benchmark is an absolute waste of time IMHO.

Your benchmarks would be way more interesting if they would consist of most triple A games and would include Windows comparison benchmarks.

That's what gamers really care about.

Please consider it.

Regards

The lack of AAA games has been explained many times before including the link you referred to that they don't expose the necessary support for test automation and reproducibility.