You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hello there!
I've managed to integrate FSR3.1 with the latest SDK using FFX API. I get into an issue while providing our depth buffer into the format VK_FORMAT_D32_SFLOAT_S8_UINT. The ffx Vulkan backend tries to create a view into VK_FORMAT_D32_SFLOAT resulting into a GPU lost.
Is there a real restriction to D32_SFLOAT format (and coping our depth buffer to a D32 depth target)?
I've tried to hardcode our texture format into ffx_vk.cpp, it doesn't crash but still not a solution. Strangely, the resourceDescription has a format field but it is fill with unknowned format. It could be an actual solution to propagate the texture format provided from the dispatch description to this backendResource->resourceDescription.
The text was updated successfully, but these errors were encountered:
Hello there!
I've managed to integrate FSR3.1 with the latest SDK using FFX API. I get into an issue while providing our depth buffer into the format VK_FORMAT_D32_SFLOAT_S8_UINT. The ffx Vulkan backend tries to create a view into VK_FORMAT_D32_SFLOAT resulting into a GPU lost.
Is there a real restriction to D32_SFLOAT format (and coping our depth buffer to a D32 depth target)?
I've tried to hardcode our texture format into ffx_vk.cpp, it doesn't crash but still not a solution. Strangely, the resourceDescription has a format field but it is fill with unknowned format. It could be an actual solution to propagate the texture format provided from the dispatch description to this backendResource->resourceDescription.
The text was updated successfully, but these errors were encountered: