Intel Arc Graphics Enjoy Nice ~10% Speedup With Recent Open-Source Linux Driver

Merged to Mesa 23.2-devel recently was an Intel Arc Graphics driver change to improve performance. This ended up being a rather significant improvement to performance and in today's article is a look at the performance impact of the recent Mesa work by Intel engineers to better the Arc Graphics family.

The change enables L3 partial write merging for compressible surfaces and other suitable cases on DG2/Alchemist hardware. The patch explained:

"This enables L3 partial write merging for a number of cases that seem to be getting accidentally disabled by the kernel, which was causing a serious performance bottleneck on DG2 and MTL platforms. The "Compressible Partial Write Merge Enable", "Coherent Partial Write Merge Enable" and "Cross-Tile Partial Write Merge Enable" bits in L3SQCREG5 were expected to be enabled by default (and confusingly, they even read off as enabled if you ran 'intel_reg read 0xb158' on an idle system), but they are getting clobbered during 3D context initialization by an i915 workaround.

Enabling L3 partial write merging of compressible surfaces in particular seems to increase rendering fillrate by over 3x in some cases (e.g. the "VulkanFillRate/FillRateGPU/resolution:1[0-3]/format:*/blend:0" fillrate-bound microbenchmarks). Significant improvements can also be reproduced in most real-world workloads we've tested so far, e.g. Counter Strike GO improves by ~11%, Shadow Of the Tomb Raider improves by ~5.5%, and AztecRuins-VK improves by ~6.5% on DG2-512 -- Thanks a lot to Caleb Callaway for these figures. No regressions have been observed so far.

>Even though this patch might strike as surprisingly simple for such a large payoff, it's the result of @fjdegroo and I trying to root-cause the rendering performance gap of DG2 on Linux vs Windows on and off during the last year, and some of the OA statistics captured by Felix early this month were greatly helpful for me to connect the last few dots, so Felix deserves a big chunk of the credit for this work."

This "intel/gfx12.5: Enable L3 partial write merging for compressible surfaces among other cases" patch is also in the process of being back-ported to the Mesa 23.1 stable series and should appear with the next Mesa 23.1 point release. Additionally, there is a kernel patch for the i915 DRM kernel driver to fix context workarounds for non-masked registers.

I've now had the time to run through a variety of different Linux gaming and application OpenGL and Vulkan benchmarks for seeing the impact of this and other changes over the past month to Mesa. Linux 6.4 was running on the system throughout while comparing the Mesa Git state as of 12 June to 16 July with this round of testing. Unfortunately the Oibaf PPA builds were recently having issues that also delayed this testing in preferring to use PPA builds for the Mesa benchmarking to encourage and foster community reproducibility. An Intel Arc Graphics A770 graphics card was used for this round of benchmarking.

Let's move on to see how the Intel Arc Graphics A770 open-source Linux graphics performance has evolved simply in the past month alone.