The hardware resources about each vGPU in a VM

Virtual GPU Forums General Discussion
1

1 What are the independent hardware resources(e.g. L1cache, L2cache, DRAM space, engine) for each vGPU in a VM?
2 What hardware resources of the GPU are shared by all vGPUs?
3 Can Nsight analyse the performance of each vGPU?

PS: the vGPU is not MIG instance.

2

1 What are the independent hardware resources(e.g. L1cache, L2cache, DRAM space, engine) for each vGPU in a VM?
->GPU memory (fixed reserved for the VM)
2 What hardware resources of the GPU are shared by all vGPUs?
->GPU Rendering and Encode/Decode ->Scheduler with timeslicing
3 Can Nsight analyse the performance of each vGPU?
->No

3

Thank you for your reply.

1 What are the independent hardware resources(e.g. L1cache, L2cache, DRAM space, engine) for each vGPU in a VM?
->GPU memory (fixed reserved for the VM)
→ GPU memory you mentioned belongs to DRAM, L1 cache and L2 cache are shared by all vGPUs, right?

3 Can Nsight analyse the performance of each vGPU?
->No
→ But this link Virtual GPU Software User Guide - NVIDIA Docs mentions the following contents, how should i understand it?

NVIDIA CUDA Toolkit Development Tool Support

NVIDIA vGPU supports the following NVIDIA CUDA Toolkit development tools on some GPUs:

  • Debuggers:
    • CUDA-GDB
    • Compute Sanitizer
  • Profilers:
    • The Activity, Callback, and Profiling APIs of the CUDA Profiling Tools Interface (CUPTI)Other CUPTI APIs, such as the Event and Metric APIs, are not supported.
    • NVIDIA Nsight™ Compute
    • NVIDIA Nsight Systems
    • NVIDIA Nsight plugin
    • NVIDIA Nsight Visual Studio plugin