AMD Radeon HD 8670D Preview On Linux

[phoronix]

Phoronix: AMD Radeon HD 8670D Preview On Linux

This past weekend I delivered benchmarks of the AMD A10-6800K Richland APU under Ubuntu Linux. This mild upgrade over AMD's Trinity APU ran faster on the CPU side and overclocked well, but how do the graphics performance under Linux? In this article are benchmarks of the Radeon HD 8670D running the Catalyst Linux driver on Ubuntu and compared to the previous-generation Radeon HD 7660D APU graphics.

http://www.phoronix.com/vr.php?view=18864

[tuke81]

Odd results. From anadtech test 5800k and 6800k are almost equal. Is richland some what less prone to memory bandwith or what is the reason such a high gap between the two. It's still the same gpu only slightly overclocked.

[remm]

So I'd say overall the Intel iris 5000 integrated GPUs look better than this, at least on Linux. With less hassle since you don't have to deal with the semi broken driver.

[lovenemesis]

Thought it's just 44MHz Difference

I thought the only difference between Radeon 7660D and Radeon 8670D is the latter one has 44MHz high frequency.

But this, wow!

[lovenemesis]

So I'd say overall the Intel iris 5000 integrated GPUs look better than this, at least on Linux. With less hassle since you don't have to deal with the semi broken driver.

Will Iris 5000 serious iGPU be available on DIY market? Last time I heard that is restricted to OEM supply.

Well Catalyst binary definitely has problems, but it's not that bad (as it's used to be?). I have been using it from rpmfusion repo for quite a while.
Nearly all Humble Indie titles work well, except Anomaly.

[curaga]

All under catalyst, yet such huge differences.

[bridgman]

If the Trinity and Richland parts are being swapped into the same mobo it's possible that the Trinity part doesn't play well with the BIOS upgrade that Richland requires.

[molecule-eye]

Something's definitely not right. The performance delta between the two is surprisingly higher than what most would think, given Anandtech's results, but the results here indicate something's broken on the Trinity platform. I would try to locate the problem and let us know what it is, something much more useful than the benchmark results shown in the article.

[spacetoilet]

Something's definitely not right. The performance delta between the two is surprisingly higher than what most would think, given Anandtech's results, but the results here indicate something's broken on the Trinity platform. I would try to locate the problem and let us know what it is, something much more useful than the benchmark results shown in the article.

you just missed what he said

If the Trinity and Richland parts are being swapped into the same mobo it's possible that the Trinity part doesn't play well with the BIOS upgrade that Richland requires.

[remm]

Will Iris 5000 serious iGPU be available on DIY market? Last time I heard that is restricted to OEM supply.

Well Catalyst binary definitely has problems, but it's not that bad (as it's used to be?). I have been using it from rpmfusion repo for quite a while.
Nearly all Humble Indie titles work well, except Anomaly.

They will be in the R series. But given the price of the i7 version, waiting for a real test is probably a good idea.