Optional (Pix): Use TrackWrite to notify Pix about modifications to the resource.Use Map to get a CPU pointer to the resource and upload CPU data to the resource.(Or D3D12_HEAP_TYPE_CUSTOM, then use GetCustomHeapProperties to get the page property and memory pool) Create a resource using D3D12_HEAP_TYPE_GPU_UPLOAD, with D3D12_CPU_PAGE_PROPERTY_UNKNOWN and D3D12_MEMORY_POOL_UNKNOWN.iGPUs do not have this requirement, since GPU upload heaps are going to use L0 memory for UMA.

For dGPUs, resizable bar need to be turned on in BIOS. GPU upload heaps can be used on both iGPUs and dGPUs. To check if GPU upload heaps are supported on the current device, check D3D12_FEATURE_DATA_D3D12_OPTIONS16:: GPUUploadHeapSupported. Please note: this direct access is being released as an implementation preview requiring developer mode and performance may not be indicative of the final release. With the VRAM being managed by Windows, D3D now exposes the heap memory access directly to the CPU! This allows both the CPU and GPU to directly access the memory simultaneously, removing the need to copy data from the CPU to the GPU increasing performance in certain scenarios. Most modern GPUs have introduced VRAM resizable base address register (BAR) enabling Windows to manage the GPU VRAM in WDDM 2.0 or later. Historically a GPU’s VRAM was inaccessible to the CPU, forcing programs to have to copy large amounts of data to the GPU via the PCI bus. Head to Agility SDK Downloads to grab it.

We’re pleased to announce that we have a new preview Agility SDK out today, letting developers experiment with two new features: GPU Upload Heaps and Non-normalized Sampling!

Directx 11 runtime download