Deferred lighting default per GPU class (split vs uber) #824

New issue

Closed

opened 2026-10-17 18:10:02 +00:00 by sky_line · 1 comment

sky_line commented

2026-10-17 18:10:02 +00:00

Owner

Context: measured on a GTX 1080 (Pascal) at 4K with the test-control channel and the shader clock (RenderVKShaderClockLighting) / pipeline stats (RenderVKPipelineStats). The frame is GPU-bound but not bandwidth-bound (memory controller ~36% busy on the 1080, ~14% on a 5090): it is latency / occupancy bound, so trading a little memory traffic for fewer registers and fewer dependent loads should help every GPU generation.

On a GTX 1080 the uber kernel beats the split at both Firestorm-equivalent (-0.7 ms) and high settings (-2.0 ms): the split's combine runs at the same 126 registers as uber and the extra passes re-decode the G-buffer. The split was tuned on an RTX 5090. Revisit after the froxel grid and register work (both paths should get much lighter), then pick per GPU class; blocked on #809 for any live switch. Context: measured on a GTX 1080 (Pascal) at 4K with the test-control channel and the shader clock (RenderVKShaderClockLighting) / pipeline stats (RenderVKPipelineStats). The frame is GPU-bound but not bandwidth-bound (memory controller ~36% busy on the 1080, ~14% on a 5090): it is latency / occupancy bound, so trading a little memory traffic for fewer registers and fewer dependent loads should help every GPU generation.

sky_line added the

labels

2026-10-17 18:10:02 +00:00

sky_line added this to the Slipstream project

2026-10-17 18:10:02 +00:00

sky_line referenced this issue

2026-10-17 22:06:51 +00:00

AO: single-pass depth pyramid + cheaper GTAO tier for Pascal (AO is 4.9 ms at 4K on a GTX 1080) #828