# Vulkan **URL:** https://forums.developer.nvidia.com/c/gaming-and-visualization-technologies/apis/vulkan/205.md [Latest](https://forums.developer.nvidia.com/latest.md) · [Categories](https://forums.developer.nvidia.com/categories.md) · [Tags](https://forums.developer.nvidia.com/tags.md) --- ## [Vulkan SC Overview and Resources](https://forums.developer.nvidia.com/t/vulkan-sc-overview-and-resources/315754)
**Author:** [@andyk2](https://forums.developer.nvidia.com/u/andyk2)\ **Replies:** 0\ **Last updated:** [December 4, 2024, 7:14pm UTC](https://forums.developer.nvidia.com/t/vulkan-sc-overview-and-resources/315754 "2024-12-04T19:14:55Z")
What is Vulkan Safety Critical? The design of Vulkan SC gives safety-critical application developers detailed control of GPU acceleration in a way that can be rigorously specified and tested to meet safety certification… --- ## [Welcome to Vulkan SC forums!](https://forums.developer.nvidia.com/t/welcome-to-vulkan-sc-forums/315603)
**Author:** [@andyk2](https://forums.developer.nvidia.com/u/andyk2)\ **Replies:** 0\ **Last updated:** [December 3, 2024, 9:46pm UTC](https://forums.developer.nvidia.com/t/welcome-to-vulkan-sc-forums/315603 "2024-12-03T21:46:29Z")
Hello visitors! Please feel free to post any questions regarding the Vulkan SC SDK or any questions related to platform offerings here. To find out more about Vulkan SC, please visit the NV website: Vulkan SC | NVIDIA D… --- ## [Task shader reads the mesh descriptor set when layouts are stage-exclusive and UPDATE\_AFTER\_BIND is enabled](https://forums.developer.nvidia.com/t/task-shader-reads-the-mesh-descriptor-set-when-layouts-are-stage-exclusive-and-update-after-bind-is-enabled/385129)
**Author:** [@felix.richter](https://forums.developer.nvidia.com/u/felix.richter)\ **Replies:** 0\ **Last updated:** [October 5, 2026, 2:27pm UTC](https://forums.developer.nvidia.com/t/task-shader-reads-the-mesh-descriptor-set-when-layouts-are-stage-exclusive-and-update-after-bind-is-enabled/385129 "2026-10-05T14:27:06Z")
Summary In a task + mesh + fragment graphics pipeline, the task shader reads the descriptors of the mesh shader’s descriptor set instead of its own set when all of the following are true: The task set layout uses stage… --- ## [Incoherent offsets in my vertex shader for some meshes](https://forums.developer.nvidia.com/t/incoherent-offsets-in-my-vertex-shader-for-some-meshes/385041)
**Author:** [@laurentduroisin](https://forums.developer.nvidia.com/u/laurentduroisin)\ **Replies:** 2\ **Last updated:** [October 4, 2026, 3:00pm UTC](https://forums.developer.nvidia.com/t/incoherent-offsets-in-my-vertex-shader-for-some-meshes/385041 "2026-10-04T15:00:10Z")
Hi! I’m stuck on a display bug since several weeks now. I defines offsets with reference vertices and indices positions of first vertices/indices mesh on the buffers, depending on the visible mesh and the lod level, and … --- ## [Vulkan RayQuery + descriptor heap: stable on 610.88, GPU page faults on 616.x / 617.x; Nsight Graphics dies at vkCreateDevice](https://forums.developer.nvidia.com/t/vulkan-rayquery-descriptor-heap-stable-on-610-88-gpu-page-faults-on-616-x-617-x-nsight-graphics-dies-at-vkcreatedevice/384792)
**Author:** [@Dioxyd](https://forums.developer.nvidia.com/u/Dioxyd)\ **Replies:** 0\ **Last updated:** [September 30, 2026, 10:43pm UTC](https://forums.developer.nvidia.com/t/vulkan-rayquery-descriptor-heap-stable-on-610-88-gpu-page-faults-on-616-x-617-x-nsight-graphics-dies-at-vkcreatedevice/384792 "2026-09-30T22:43:03Z")
Hi, I’m seeing what looks like a driver regression with Vulkan ray query on recent Game Ready drivers, and I wanted to check whether others hit the same thing before I invest in a minimal standalone repro. Setup Custo… --- ## [Windows Vulkan: unused compute pipeline causes device loss on RTX 4090 / 4070 SUPER](https://forums.developer.nvidia.com/t/windows-vulkan-unused-compute-pipeline-causes-device-loss-on-rtx-4090-4070-super/383848)
**Author:** [@shogan](https://forums.developer.nvidia.com/u/shogan)\ **Replies:** 3\ **Last updated:** [September 30, 2026, 5:41am UTC](https://forums.developer.nvidia.com/t/windows-vulkan-unused-compute-pipeline-causes-device-loss-on-rtx-4090-4070-super/383848 "2026-09-30T05:41:17Z")
We are reporting a reproducible failure that strongly points to an NVIDIA Vulkan driver defect. Creating a second compute pipeline causes an otherwise correct shader dispatch to fail with VK\_ERROR\_DEVICE\_LOST (-4). The s… --- ## [RTX 4060: Fragment shader reads the wrong descriptor for any heap index other than 0](https://forums.developer.nvidia.com/t/rtx-4060-fragment-shader-reads-the-wrong-descriptor-for-any-heap-index-other-than-0/383842)
**Author:** [@jiachen.wang](https://forums.developer.nvidia.com/u/jiachen.wang)\ **Replies:** 4\ **Last updated:** [September 28, 2026, 8:01pm UTC](https://forums.developer.nvidia.com/t/rtx-4060-fragment-shader-reads-the-wrong-descriptor-for-any-heap-index-other-than-0/383842 "2026-09-28T20:01:25Z")
VK\_EXT\_descriptor\_heap: fragment shader reads the wrong descriptor for any heap index other than 0 Reading a layout(descriptor\_heap) array from a fragment shader at any index other than 0 returns the wrong descriptor on … --- ## [RTX 4070: Linux Vulkan atomic add throughput (u32, random access) degrades above 1MiB buffer size](https://forums.developer.nvidia.com/t/rtx-4070-linux-vulkan-atomic-add-throughput-u32-random-access-degrades-above-1mib-buffer-size/383912)
**Author:** [@kasten.m](https://forums.developer.nvidia.com/u/kasten.m)\ **Replies:** 2\ **Last updated:** [September 27, 2026, 6:24pm UTC](https://forums.developer.nvidia.com/t/rtx-4070-linux-vulkan-atomic-add-throughput-u32-random-access-degrades-above-1mib-buffer-size/383912 "2026-09-27T18:24:00Z")
Atomic add throughput (u32, random access) on RTX 4070 is ~32G adds / s up to 16MiB buffer size except on Linux Vulkan, which degrades above 1MiB buffer size. The graph is from a microbenchmark ( GitHub - Draradech/a… --- ## [\[Vulkan\]\[615.71.09\]\[RTX 4050\]\[Descriptor Heap\] Incorrect RowMajor MatrixStride handling for whole matrix OpLoad through an untyped Uniform pointer](https://forums.developer.nvidia.com/t/vulkan-615-71-09-rtx-4050-descriptor-heap-incorrect-rowmajor-matrixstride-handling-for-whole-matrix-opload-through-an-untyped-uniform-pointer/383932)
**Author:** [@Opposite](https://forums.developer.nvidia.com/u/Opposite)\ **Replies:** 0\ **Last updated:** [September 21, 2026, 8:06pm UTC](https://forums.developer.nvidia.com/t/vulkan-615-71-09-rtx-4050-descriptor-heap-incorrect-rowmajor-matrixstride-handling-for-whole-matrix-opload-through-an-untyped-uniform-pointer/383932 "2026-09-21T20:06:55Z")
Environment: OS/distribution: Arch Linux Kernel: 7.2.6-arch2-1 GPU: NVIDIA GeForce RTX 4050 Laptop GPU Driver: 615.71.09 Vulkan API: 1.4.351 Vulkan SDK: 1.4.357 SPIRV-Tools: 2026.3 Description: When having descr… --- ## [Sparse texture binding is painfully slow](https://forums.developer.nvidia.com/t/sparse-texture-binding-is-painfully-slow/259105)
**Author:** [@antoinerichermoz](https://forums.developer.nvidia.com/u/antoinerichermoz)\ **Replies:** 30\ **Last updated:** [September 19, 2026, 6:32pm UTC](https://forums.developer.nvidia.com/t/sparse-texture-binding-is-painfully-slow/259105 "2026-09-19T18:32:08Z")
Sparse texture binding / unbinding can be painfully slow: it can take multiple seconds (yes, not milliseconds) to bind 1000 pages in a 1024^3 texure. (My setup is an A4000 with a dual Xeon 4210R running on Ubuntu 22.04 w… --- ## [Ray-Box Intersection in Vulkan Ray Tracing Pipeline](https://forums.developer.nvidia.com/t/ray-box-intersection-in-vulkan-ray-tracing-pipeline/383795)
**Author:** [@meetbanthia](https://forums.developer.nvidia.com/u/meetbanthia)\ **Replies:** 0\ **Last updated:** [September 19, 2026, 6:19pm UTC](https://forums.developer.nvidia.com/t/ray-box-intersection-in-vulkan-ray-tracing-pipeline/383795 "2026-09-19T18:19:36Z")
Hi all, I have been working on one problem which require use of ray - AABB intersection, but i have noticed that in Vulkan, ray box intersection can happen in RT core itself, but it does not expose distance of where thi… --- ## [RTX 5070 / Driver 616.92.0: VK\_ERROR\_DEVICE\_LOST](https://forums.developer.nvidia.com/t/rtx-5070-driver-616-92-0-vk-error-device-lost/383602)
**Author:** [@tima.tcvetkov2](https://forums.developer.nvidia.com/u/tima.tcvetkov2)\ **Replies:** 0\ **Last updated:** [September 18, 2026, 10:21am UTC](https://forums.developer.nvidia.com/t/rtx-5070-driver-616-92-0-vk-error-device-lost/383602 "2026-09-18T10:21:27Z")
Item Failing Working 1 Working 2 GPU RTX 5070 RTX 2070 SUPER Quadro T2000 Max-Q (4 GB) Architecture Blackwell Turing Turing Driver 616.92.0 610.88.0 Vulkan 1.4.351 1.4.341 1.4.341 OS Windows 8 (6.2) Wind… --- ## [Pipeline creation is 6.5x slower with OpFunction None than DontInline](https://forums.developer.nvidia.com/t/pipeline-creation-is-6-5x-slower-with-opfunction-none-than-dontinline/381276)
**Author:** [@asfix1](https://forums.developer.nvidia.com/u/asfix1)\ **Replies:** 2\ **Last updated:** [September 17, 2026, 8:17pm UTC](https://forums.developer.nvidia.com/t/pipeline-creation-is-6-5x-slower-with-opfunction-none-than-dontinline/381276 "2026-09-17T20:17:34Z")
I’m seeing a pretty large shader compiler cliff on an RTX 5090 with the 610.57.04 Linux driver. The attached SPIR-V pair is the same size and differs at exactly ten bytes. The only assembly changes are ten OpFunction … --- ## [Analyse cache and register usage in my vulkan shaders](https://forums.developer.nvidia.com/t/analyse-cache-and-register-usage-in-my-vulkan-shaders/383144)
**Author:** [@meetbanthia](https://forums.developer.nvidia.com/u/meetbanthia)\ **Replies:** 0\ **Last updated:** [September 12, 2026, 7:33pm UTC](https://forums.developer.nvidia.com/t/analyse-cache-and-register-usage-in-my-vulkan-shaders/383144 "2026-09-12T19:33:32Z")
Hi all!! I have been working on a Vulkan ray tracing project from some time. I don’t have much experience in Vulkan Programming. Issue i have been facing is that whenever i try to render a full resolution image(3118 \* … --- ## [RTX 5070 Ti nvgpucomp64.dll exception thrown when shader uses GL\_NV\_shader\_sm\_builtins](https://forums.developer.nvidia.com/t/rtx-5070-ti-nvgpucomp64-dll-exception-thrown-when-shader-uses-gl-nv-shader-sm-builtins/343335)
**Author:** [@michal.mwa87](https://forums.developer.nvidia.com/u/michal.mwa87)\ **Replies:** 1\ **Last updated:** [September 12, 2026, 3:19pm UTC](https://forums.developer.nvidia.com/t/rtx-5070-ti-nvgpucomp64-dll-exception-thrown-when-shader-uses-gl-nv-shader-sm-builtins/343335 "2026-09-12T15:19:33Z")
I have following compute shader: #version 460 core #extension GL\_NV\_shader\_sm\_builtins : require layout(set = 0, binding = 0) uniform writeonly image2D image; layout(push\_constant) uniform Settings { layout(offse… --- ## [Add support for VK\_AMD\_shader\_image\_load\_store\_lod](https://forums.developer.nvidia.com/t/add-support-for-vk-amd-shader-image-load-store-lod/361061)
**Author:** [@antoinerichermoz](https://forums.developer.nvidia.com/u/antoinerichermoz)\ **Replies:** 1\ **Last updated:** [September 10, 2026, 7:10pm UTC](https://forums.developer.nvidia.com/t/add-support-for-vk-amd-shader-image-load-store-lod/361061 "2026-09-10T19:10:16Z")
When an image is binded to a shader as a sampled image, you can access the mip levels of the image by specifying the desired mip level in the texture access call: uniform sampler2D sampledImage; void main() { ivec2 … --- ## [(Potential) VK\_EXT\_descriptor\_buffer synchronization/corruption bug](https://forums.developer.nvidia.com/t/potential-vk-ext-descriptor-buffer-synchronization-corruption-bug/382912)
**Author:** [@weteringt](https://forums.developer.nvidia.com/u/weteringt)\ **Replies:** 0\ **Last updated:** [September 10, 2026, 5:54pm UTC](https://forums.developer.nvidia.com/t/potential-vk-ext-descriptor-buffer-synchronization-corruption-bug/382912 "2026-09-10T17:54:13Z")
I have encountered a synchronization or corruption bug which seems to be a driver issue (on 3080 12GB, Windows 11). This only happens when descriptor buffers are used. Some resources written to at the start of frame fla… --- ## [Random fragment's rasterization with mesh shaders](https://forums.developer.nvidia.com/t/random-fragments-rasterization-with-mesh-shaders/382899)
**Author:** [@laurentduroisin](https://forums.developer.nvidia.com/u/laurentduroisin)\ **Replies:** 1\ **Last updated:** [September 10, 2026, 3:55pm UTC](https://forums.developer.nvidia.com/t/random-fragments-rasterization-with-mesh-shaders/382899 "2026-09-10T15:55:15Z")
Hi! It’s been a few days I’m trying to figure out why after my mesh shader invocation I’ve random pixel colors on the screen. But I can’t find anything wrong in my code : -My descriptors are corrects. Everytihign works … --- ## [VK\_KHR\_swapchain\_maintainance1 present fences don't get signaled after multiple OUT\_OF\_DATE](https://forums.developer.nvidia.com/t/vk-khr-swapchain-maintainance1-present-fences-dont-get-signaled-after-multiple-out-of-date/382630)
**Author:** [@karlsasssie](https://forums.developer.nvidia.com/u/karlsasssie)\ **Replies:** 0\ **Last updated:** [September 7, 2026, 9:13pm UTC](https://forums.developer.nvidia.com/t/vk-khr-swapchain-maintainance1-present-fences-dont-get-signaled-after-multiple-out-of-date/382630 "2026-09-07T21:13:47Z")
Acquiring multiple frames from a swapchain, then presenting all of them with presentation fences can lead to some of them not becoming signaled even tho OUT\_OF\_DATE was returned. I’m not 100% if i read the vulkan specs … --- ## [Windowed Vulkan HDR is presented through the SDR path when VkApplicationInfo::pEngineName is exactly "DXVK"](https://forums.developer.nvidia.com/t/windowed-vulkan-hdr-is-presented-through-the-sdr-path-when-vkapplicationinfo-penginename-is-exactly-dxvk/382517)
**Author:** [@krisztailor](https://forums.developer.nvidia.com/u/krisztailor)\ **Replies:** 0\ **Last updated:** [September 6, 2026, 5:22pm UTC](https://forums.developer.nvidia.com/t/windowed-vulkan-hdr-is-presented-through-the-sdr-path-when-vkapplicationinfo-penginename-is-exactly-dxvk/382517 "2026-09-06T17:22:06Z")
Summary. On this driver, a windowed Vulkan application that presents an R16G16B16A16\_SFLOAT swapchain in VK\_COLOR\_SPACE\_EXTENDED\_SRGB\_LINEAR\_EXT receives correct HDR output, unless it identifies itself with VkApplica… --- ## [vkFreeMemory leaks for D3D11 imported images](https://forums.developer.nvidia.com/t/vkfreememory-leaks-for-d3d11-imported-images/382249)
**Author:** [@nateglasser](https://forums.developer.nvidia.com/u/nateglasser)\ **Replies:** 0\ **Last updated:** [September 3, 2026, 6:36pm UTC](https://forums.developer.nvidia.com/t/vkfreememory-leaks-for-d3d11-imported-images/382249 "2026-09-03T18:36:16Z")
Driver: 581.57.0.0 Device: GTX 1080 OS: Windows 10.0.19045 (I realize this is somewhat outdated, but it’s what I have!) The driver does not free the underlying handle in vkFreeMemory for allocations imported from D3D… --- ## [VK\_EXT\_descriptor\_heap ray query causes VK\_ERROR\_DEVICE\_LOST](https://forums.developer.nvidia.com/t/vk-ext-descriptor-heap-ray-query-causes-vk-error-device-lost/376527)
**Author:** [@qian-o](https://forums.developer.nvidia.com/u/qian-o)\ **Replies:** 3\ **Last updated:** [August 29, 2026, 8:40am UTC](https://forums.developer.nvidia.com/t/vk-ext-descriptor-heap-ray-query-causes-vk-error-device-lost/376527 "2026-08-29T08:40:50Z")
Summary A minimal native Vulkan 1.4 compute application consistently returns VK\_ERROR\_DEVICE\_LOST when an inline ray query uses an acceleration structure obtained from ResourceHeapEXT data. The shader loads a raw 64-bit… --- ## [Vulkan multiview: primitive fully clipped in view 0 is incorrectly discarded in view 1](https://forums.developer.nvidia.com/t/vulkan-multiview-primitive-fully-clipped-in-view-0-is-incorrectly-discarded-in-view-1/379857)
**Author:** [@nvidia2757](https://forums.developer.nvidia.com/u/nvidia2757)\ **Replies:** 0\ **Last updated:** [August 11, 2026, 7:24pm UTC](https://forums.developer.nvidia.com/t/vulkan-multiview-primitive-fully-clipped-in-view-0-is-incorrectly-discarded-in-view-1/379857 "2026-08-11T19:24:13Z")
I am observing what appears to be incorrect per-view primitive clipping behavior in Vulkan multiview on NVIDIA. Hardware GPU NVIDIA GeForce RTX 5070 Laptop GPU OS / Driver Model Windows, WDDM NVIDIA Driver … --- ## [Possible stack-buffer-overflow-by-1 caused overwrite callee-saved register in vkGetPipelineKeyKHR when pPipelineCreateInfo == nullptr (get global key)](https://forums.developer.nvidia.com/t/possible-stack-buffer-overflow-by-1-caused-overwrite-callee-saved-register-in-vkgetpipelinekeykhr-when-ppipelinecreateinfo-nullptr-get-global-key/373306)
**Author:** [@alexseykutasov](https://forums.developer.nvidia.com/u/alexseykutasov)\ **Replies:** 3\ **Last updated:** [August 7, 2026, 4:30pm UTC](https://forums.developer.nvidia.com/t/possible-stack-buffer-overflow-by-1-caused-overwrite-callee-saved-register-in-vkgetpipelinekeykhr-when-ppipelinecreateinfo-nullptr-get-global-key/373306 "2026-08-07T16:30:28Z")
\##Details: os: Windows 11 IoT Enterprise 22H2 driver version: Game Ready Driver 610.47 hardware: GeForce RTX 3060 12 GB ##How to reproduce: create VkDevice with {extensionNames = \[VK\_KHR\_PIPELINE\_BINARY\_EXTENSIO… --- ## [Descriptor heap access to nested structs ignores offsets](https://forums.developer.nvidia.com/t/descriptor-heap-access-to-nested-structs-ignores-offsets/379370)
**Author:** [@alexseykutasov](https://forums.developer.nvidia.com/u/alexseykutasov)\ **Replies:** 0\ **Last updated:** [August 6, 2026, 5:50pm UTC](https://forums.developer.nvidia.com/t/descriptor-heap-access-to-nested-structs-ignores-offsets/379370 "2026-08-06T17:50:44Z")
I face a really strange behaviour when use VK\_EXT\_descriptor\_heap extension with Storage Buffer contains a nested structs. I’m not sure - it might be the bug of glslc or incorrect GLSL shader. Validation layers don’t rep… --- ## [\`createComputePipeline\` crash when SPIR-V module does not bind to first descriptor](https://forums.developer.nvidia.com/t/createcomputepipeline-crash-when-spir-v-module-does-not-bind-to-first-descriptor/379117)
**Author:** [@aidas.tekorius](https://forums.developer.nvidia.com/u/aidas.tekorius)\ **Replies:** 0\ **Last updated:** [August 4, 2026, 10:28am UTC](https://forums.developer.nvidia.com/t/createcomputepipeline-crash-when-spir-v-module-does-not-bind-to-first-descriptor/379117 "2026-08-04T10:28:19Z")
Hello, I am working on a research project that uses Vulkan for GPU accelerated computation. I encountered a very weird issue where no validation layer output was being printed, but trying to use \`vk::raii::Device::creat… --- ## [Decoding H265 with Vulkan Video results in Device Removal](https://forums.developer.nvidia.com/t/decoding-h265-with-vulkan-video-results-in-device-removal/378933)
**Author:** [@agrael](https://forums.developer.nvidia.com/u/agrael)\ **Replies:** 0\ **Last updated:** [August 2, 2026, 4:34pm UTC](https://forums.developer.nvidia.com/t/decoding-h265-with-vulkan-video-results-in-device-removal/378933 "2026-08-02T16:34:36Z")
Hi, I am trying to implement HEVC decoding and for starters I have decided to decode a simple HEIF. Everything seems normal: Aligned buffer for input (256) No DPB, everywhere there are nullptrs dimensions for the outpu… --- ## [Massive performance drop with VK\_QUERY\_PIPELINE\_STATISTIC\_TASK/MESH\_SHADER\_INVOCATIONS\_BIT\_EXT on RTX 4060](https://forums.developer.nvidia.com/t/massive-performance-drop-with-vk-query-pipeline-statistic-task-mesh-shader-invocations-bit-ext-on-rtx-4060/376792)
**Author:** [@tamaspeti3451](https://forums.developer.nvidia.com/u/tamaspeti3451)\ **Replies:** 0\ **Last updated:** [July 14, 2026, 8:11am UTC](https://forums.developer.nvidia.com/t/massive-performance-drop-with-vk-query-pipeline-statistic-task-mesh-shader-invocations-bit-ext-on-rtx-4060/376792 "2026-07-14T08:11:09Z")
Hi, I’ve run into what seems like a very unusual performance issue when enabling the Task/Mesh shader pipeline statistics counters. I’m working on a fully GPU-driven Vulkan renderer based on mesh shaders. Visibility is… --- ## [Vulkan regression report: transform-feedback application breaks between driver 596.49 and 610.x (RTX 3050 Laptop)](https://forums.developer.nvidia.com/t/vulkan-regression-report-transform-feedback-application-breaks-between-driver-596-49-and-610-x-rtx-3050-laptop/376564)
**Author:** [@faisal.sumak](https://forums.developer.nvidia.com/u/faisal.sumak)\ **Replies:** 0\ **Last updated:** [July 11, 2026, 12:57pm UTC](https://forums.developer.nvidia.com/t/vulkan-regression-report-transform-feedback-application-breaks-between-driver-596-49-and-610-x-rtx-3050-laptop/376564 "2026-07-11T12:57:05Z")
A Vulkan application that relies on VK\_EXT\_transform\_feedback for GPU geometry read-back worked reliably on driver 596.49 (32.0.15.9649) and regressed on the 610 branch (first observed on 610.47, still present on 610.74 … --- ## [VKD3D-Proton descriptor heap path causes VK\_ERROR\_DEVICE\_LOST on driver 610 in The First Descendant](https://forums.developer.nvidia.com/t/vkd3d-proton-descriptor-heap-path-causes-vk-error-device-lost-on-driver-610-in-the-first-descendant/374747)
**Author:** [@daniyaljahandide8806](https://forums.developer.nvidia.com/u/daniyaljahandide8806)\ **Replies:** 2\ **Last updated:** [July 6, 2026, 9:24am UTC](https://forums.developer.nvidia.com/t/vkd3d-proton-descriptor-heap-path-causes-vk-error-device-lost-on-driver-610-in-the-first-descendant/374747 "2026-07-06T09:24:14Z")
System: GPU: RTX 2060 SUPER Driver: 610.xx (regression compared to 595) OS: Linux Proton: GE-Proton 11 VKD3D-Proton: 3.1.0 Game: The First Descendant (UE5 DX12 title) Issue: When forcing descriptor heap path in VKD3D… [Next page](https://forums.developer.nvidia.com/c/gaming-and-visualization-technologies/apis/vulkan/205.md?page=1)