- every texture upload malloc'd a CPU copy of its pixels and never freed it (236 MB over the valley's boot, and again for every model loaded later): the pixels are converted straight into the mapped staging buffer - the per-level and per-layer views gvk_view_of made outlived their texture, and on MoltenVK a view keeps its Metal texture alive: a released texture takes its views with it, and the cache is keyed by numbers instead of a string built on every call - the Vulkan structs filled for a draw, pass, barrier, descriptor set, buffer, allocation or upload (about sixty call sites) come from a reused 1 MB scratch ring (gvk_tmp) - the descriptor-set cache and the retired buffers are emptied in place, not replaced; the grass cull's dispatch arguments are made once - macOS drains an autorelease pool each frame (Vk.frame_pool, lvk_frame_pool in vk_mac.ll) - R3D_VK_PROF reports Vulkan objects made and destroyed by kind, and every cache's length - examples/rendering/smooth presents through render3d, so it runs on Vulkan too The full valley on headless Vulkan loads to 2.18 GB and holds (it passed 8 GB while loading before). ludic-dev test 305/305, selfhost-test 33/33. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
1.2 KiB
bump: patch
type: fix
The Vulkan renderer no longer leaks memory as it loads and draws. Every texture upload
malloc'd a CPU copy of its pixels and never freed it (236 MB over the valley's boot, and again on
every model loaded later); the pixels are now converted straight into the mapped staging buffer.
The per-level and per-layer views made for mipmapped and layered targets outlived their texture,
and on MoltenVK a view keeps its Metal texture alive, so every rebuild of the screen-sized targets
left the last set in memory; a released texture now takes its views with it, and the cache is
keyed by numbers rather than a string built on every call. The Vulkan structs filled for a draw, a
pass, a barrier, a descriptor set or a buffer (about sixty call sites, several per draw) come from
a reused 1 MB scratch ring instead of a malloc each; the descriptor-set cache and the retired
buffer list are emptied and refilled rather than replaced; and on macOS the frame drains an
autorelease pool (Vk.frame_pool()), which MoltenVK and the layer rely on. R3D_VK_PROF now
also prints, every 120 frames, the Vulkan objects made and destroyed by kind and the length of
every cache, so a kind made each frame and never destroyed shows at once.