Releases: NVIDIAGameWorks/NRI
Releases · NVIDIAGameWorks/NRI
v1.115
v1.115:
HIGHLIGHTS:
- NRI: added
MemoryLocation::DEVICE_UPLOAD
to get access toGPU_UPLOAD
or(Resizable) BAR
heap with silent fallback toHOST_UPLOAD
- Cmake: explicit specification of required backends
- Cmake: optional
Agility SDK
support (needed for D3D12 to get access to recent features, likeReBAR
) - fixes and improvements
DETAILS:
- NRI: added
MemoryLocation::DEVICE_UPLOAD
(ReBAR
) - NRI: fixed compilation on Android due to missing
_mm_pause
- VK: improved "memoryType" selection
- VK: improved memory heaps info reporting
- Cmake improvements
- updated README
v1.114
v1.114:
- NRI: removed
skipLiveObjectsReporting
because it doesn't make a lot of sense in D3D in case of multi device usage - NRI: exposed
nriReportLiveObjects
(D3D only, not needed for VK) - NRI: exposed
nriGetGraphicsAPIString
- NRI: exposed
nriGetFormatString
- NRI: improved atomic based "lock"
- D3D11: improved error reporting
- VK: fixed SwapChain semaphore usage related issue reported by VK SDK
v1.3.275
- added GitHub actions (1st attempt)
- Vulkan headers updated to
v1.3.275
- merged MRs from GitHub
- updated README
v1.113
v1.112
v1.111
v1.111:
- NRI: removed unnecessary BT709_G10_8BIT swap chain format
- NRI: removed
SetSwapChainHdrMetadata
as it's being deprecated by MS - NRI: exposed
GetDisplayDesc
in swap chain to retrieve current color settings of the display the window is on - NRI: removed
ResizeBuffers
, which was a WAR for a D3D11 issue, which is fixed now - NRI: removed
WindowSystemType
, just initialize one ofWindow
members
v1.110
v1.109
v1.109:
- SwapChain: removed exclusive fullscreen (was partially implemented, should be done in another way)
- D3D12: fixed bug in "ResizeBuffers" (introduced by recent PR)
- D3D11: fixed "locking" (implicit referencing) of resources by CommandBuffer
- D3D11/D3D12: fixed wrong (0 alpha) background color in SwapChain
- VK: fixed mismatched image layouts in some "vkCmdCopy*" calls (introduced by recent PR)
- VK: swap chain improvements
- VK: refactoring
v1.108
v1.107
v1.106
v1.106:
- merged PR: wait for event completion in fence
- merged PR: added check for outdated swapchain
- merged PR: use optimal layout for texture copying
- merged PR: added support for swapchain resize
- NRI:
uint16_t
=>Dim_t
in some forgotten places - D3D11/D3D12: added waitable swapchain support (implicit)
- D3D11/D3D12: swap chain improvements
- Cmake: fixed unconditional use of
CMAKE_MSVC_RUNTIME_LIBRARY