Nsight Perf SDK Release Notes - v2021.1 Early Access 2

Prerequisites

NVIDIA GeForce, Quadro, or RTX GPU listed below:

NVIDIA Architecture	Chip Names	GeForce Families	Quadro and NVIDIA RTX Families
NVIDIA Ampere GA10x Architecture	GA102, GA104, GA106	RTX 3090, RTX 3080, RTX 3070, RTX 3060	RTX A6000, A40
NVIDIA Turing TU10x Architecture	TU102, TU104, TU106	Titan RTX, RTX 2080, RTX 2070, RTX 2060	RTX 8000, RTX 6000, RTX 5000, RTX 4000, RTX 3000
NVIDIA Turing TU11x Architecture	TU116, TU117	GTX 1660, GTX 1650	T2000, T1000
NVIDIA Volta V100i Architecture	GV100	Titan V	Quadro GV100

By default, the driver disallows GPU profiling. To enable permissions, follow the instructions at https://developer.nvidia.com/ERR_NVGPUCTRPERM, or try the workarounds described in the Getting Started Guide.

vkQueueWaitIdle is required before present when profiling Vulkan applications. (3201058)
Vulkan profiling on desktop linux requires X-windows to be running. (3289536)

Nsight Perf SDK does not support SLI systems. (PW-2935)
Each GPU is limited to one profiling session at a time, across the entire system. However, multiple GPUs can be independently profiled concurrently. This is a driver limitation.
D3D12 Range Profiler is not compatible with the D3D12 debug layer. (PW-925)
D3D12 Range Profiler does not support PushRange and PopRange commands on Bundles (Command Lists created with D3D12_COMMAND_LIST_TYPE_BUNDLE).
D3D11 Range Profiler does not support PushRange and PopRange calls on a Deffered Context.
Vulkan Range Profiler is not compatible with Vulkan layers that perform object-wrapping.
Vulkan Range Profiler does not support PushRange and PopRange calls on secondary command buffers (VkCommandBuffer created with VK_COMMAND_BUFFER_LEVEL_SECONDARY.
Vulkan Range Profiler does not support PushRange and PopRange calls on a VkCommandBuffer that is recording with VK_COMMAND_BUFFER_USAGE_SIMULTANEOUS_USE_BIT.
NVPW_Device_SetClockSetting() performs a system-global clock change, that is not undone at process exit. The *SetDeviceClockState() family of NvPerfUtility functions have the same behavior, as they are convenience functions around NVPW_Device_SetClockSetting().
Not all counters and metrics are available to the Range Profiler APIs, due to driver limitations. The CounterAvailability APIs exist to reveal which metrics are available.

API Versions

API Versions