AMD Perf 1080p
Above are the upper-low-end to low-end video cards compared at 1080p with the “Ultimate” preset graphics settings. We can see clear improvements with Vulkan over DX11 at the highest settings, though not as high as the previous cards. Radeon RX 5600 XT shows Vulkan to be 5% faster than DX11, Radeon RX 5500 XT 8GB shows Vulkan to be 10% faster than DX11 and Radeon RX 5500 XT 4GB shows Vulkan to be 6% faster than DX11.
This graph is interesting because it shows us several things. Firstly, it makes sense that these more GPU dependent video cards would show less of a performance increase, as it’s not the CPU bottlenecking them but the GPUs themselves. Also, at “Ultimate” graphics settings we are VRAM bottlenecked on all the video cards except the 8GB RX 5500 XT. And lo and behold, it resulted in the largest performance difference because it isn’t VRAM capacity limited at “Ultimate” settings. The very bottlenecked 4GB Radeon RX 5500 XT at “Ultimate” is just flat out stalled because of a lack of VRAM capacity in this setting.
To prove the above, we ran the game at “Medium” preset in this graph at 1080p. At “Medium” the VRAM counter in the game was not at red in any of the video cards and was below 4GB of usage. This meant the GPUs could flex their full muscle. When this happened, now we see huge Vulkan improvements over DX11.
Radeon RX 5600 XT Vulkan is 30% faster than DX11, Radeon RX 5500 XT 8GB Vulkan is 20% faster than DX11 and even the 4GB Radeon RX 5500 XT Vulkan is now 16% faster than DX11. When VRAM isn’t bottlenecked, Vulkan performance really takes off, even on these low-end video cards.
The 4GB Radeon RX 5500 XT really shows a huge improvement at “Medium” settings versus “Ultimate” where it is bottlenecked. What’s more, the 8GB Radeon RX 5500 XT is still faster than the 4GB Radeon RX 5500 XT, despite the video cards having the same GPU clock speed. It shows that even at “Medium” having 8GB onboard versus 4GB is advantageous to performance.