Intel Allows Buffer Compression For Higher Gaming Efficiency With DXVK

Merged at this time to Mesa 26.3-devel is a pleasant efficiency enhance for Intel graphics with Valve’s Steam Play when utilizing DXVK for Direct3D 8 / 9 / 10 / 11 over the Vulkan API. The open-source Intel ANV driver is seeing as much as 5% higher efficiency with DXVK.

The enchancment now merged to Mesa is enabling buffer compression for the Intel ANV driver with DXVK. Previously the buffer compression choice was enabled for VKD3D with Direct3D 12 on Vulkan as many DX12 video games had been seeing higher efficiency whereas now it is opened as much as DXVK too. On Intel Arc Battlemage there was 5.26% higher efficiency with Shadow of Tomb Raider, 4.6% with Total War Warhammer 3, and a pair of% with Total War Pharaoh. On Intel Xe3 Panther Lake was 4.8% larger efficiency with Shadow of Tomb Raider, and Total War video games within the 2% vary, among the many video games examined.

The numbers will be present in this commit setting anv_enable_buffer_comp for DXVK.

As a part of the merge was then determined to simply drop the anv_enable_buffer_comp choice that had been gated behind VKD3D, Source2 video games, and now DXVK to now enable it everywhere for attainable efficiency advantages in different video games/layers out of this shade management floor (CCS) buffer compression.

This change is merged for Mesa 26.3 that can debut as secure in November whereas for now’s out there from Mesa Git.