Zink OpenGL-On-Vulkan Seeing Some 50~100% FPS Gains

[phoronix]

Phoronix: Zink OpenGL-On-Vulkan Seeing Some 50~100% FPS Gains

After working on getting the Zink OpenGL-over-Vulkan driver up to OpenGL 4.6 with still pending patches, former Samsung OSG engineer Mike Blumenkrantz has been making remarkable progress on the performance aspect as well...

http://www.phoronix.com/scan.php?pag...00p-Perf-Boost

[EmbraceUnity]

What were the previous benchmarks? If Zink used to be three times slower, then now it is only 50% slower. If it used to be 50% slower, now it is up to 25% faster.

[R41N3R]

What were the previous benchmarks? If Zink used to be three times slower, then now it is only 50% slower. If it used to be 50% slower, now it is up to 25% faster.

This number was compared only to Zink before and after the optimizations. We will have to wait for some fresh benchmarks from Phoronix before you can compare this to radeonsi.
But I would be more interested in seeing Zink evolve and in getting it merged into Mesa than comparing this in it's current stage directly to some drivers like radeonsi.

[rmfx]

Vulkan is really the best thing that happened in graphics over the past 10 years...

Now we can have dxvk and this happening too

[Snaipersky]

Actually, since zink is a gallium driver, doesn't that mean we now have a cross vendor, and nearly cross OS state tracker? Gallium nine and 10/11 would see more value as they don't have the dxgi resource sharing issues that dxvk suffers from.

[You-]

I am looking forward to this being available on distributions. It will open up new options for software stacks.

However there is still a lot of work to get it merged upstream - a greater dificulty as there will need to be review capacity to consider too.

[Prescience500]

If Zink can be made performance competitive, it would make bringing up new open drivers really easy where OpenGL support doesn't already exist. It's exciting.

[marek]

If the performance is 0, then 50-100% faster is still 0.

[GreenByte]

fiy on the previous Zink post it (perf) was roughly the same as native opengl driver (idk was it intel or radeon), sometimes faster, so if it's compared to that, then these news are insane.

[smitty3268]

fiy on the previous Zink post it (perf) was roughly the same as native opengl driver (idk was it intel or radeon), sometimes faster, so if it's compared to that, then these news are insane.

The previous post was not actual performance, so this is not in comparison to that. That was mostly concerned with shader compile speed, not runtime.

I do think it would be interesting to get some actual benchmark numbers from Michael. A bunch of people seem to believe Zink is fast for some reason, so it'd be good to get actual confirmation on that - or that, as expected, it's still really slow.