Mesa 13.0 vs. 17.0 Performance For RADV/RadeonSI: Big Gains For Vulkan, OpenGL Boosts

[oooverclocker]

Well actually playing with 17.1 developer version these bars are not very surprising to me because that's the version that seems to wipe away most complaints magically. But it's still very impressive how it has improved when you see the actual numbers

[SaucyJack]

Xonotic is likely within error, but all three games easily hit >120 fps regardless.

Yeah, but it should really be 120fps minimum ideally rather than average.

[bridgman]

Aren't you just saying "in the late stages of a kernel release cycle only a few patches are allowed in so they all queue up for the next release" ?

[darkbasic]

Michael why you test Talos only at 4k? The FPS is so low, please consider including 1080p test next time/

Because Talos sucks and it always uses the highest resolution of the monitor.

[mibo]

Why was Deus Ex only benchmarked in low?

[Michael]

Why was Deus Ex only benchmarked in low?

Because it already struggles enough on low that pushing it higher won't add much value but just take longer to test.

[mibo]

I know that this shader compilation waiting is not nice, but the 64fps with the rx480 would be enough for me to switch to higher settings.
I you want to bench only one setting, please use a higher one next time. Thank you.

[mitch074]

I have DE :MD and a RX480; I hit 31fps at 1440p on Very high. Windows does 36 with the very same settings.

[mibo]

I have DE :MD and a RX480; I hit 31fps at 1440p on Very high. Windows does 36 with the very same settings.

Thanks, that sounds promising.

[Drago]

I have DE :MD and a RX480; I hit 31fps at 1440p on Very high. Windows does 36 with the very same settings.

Is that RadeonSI and not AMDGPU? If RadeonSI then that is fantastic!