Hi, I’m a beginner in cuda and recently I’ve been having some problems analyzing my programs using the nsight system:
I am using windows 11 and building my program in WSL2.
The graphics card I am using is:NVIDIA GeForce RTX 4060 Laptop GPU.
nvidia-smi got the following result:
I am using ubuntu 18.04 in WSL2 with cuda toolkit installed with 11.8.
I use Nsight system2023.4.1 to analyze my program.
But there is an error：CUDA device 0: Unified Memory cannot be traced on devices that don’t support peer-to-peer transfers.Please verify that SLI/NVLink is functioning properly.
And a warnning:：No NVTX events collected. Does the process use NVTX?
I queried the GPU status with the nvidia-smi topo -m command
GPU0 CPU Affinity NUMA Affinity GPU NUMA ID
GPU0 X N/A
Legend:
X = Self
SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI)
NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node
PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU)
PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge)
PIX = Connection traversing at most a single PCIe bridge
NV# = Connection traversing a bonded set of # NVLinks
Does this mean I can’t use the nsight system? If so that’s a real shame!
Is there any solution for this please?
If using ubuntu 20.04 or 22.04 and upgrading the cuda toolkit version to 12.3 to fit my cuda version and then using the latest nsight system to analyze my project.
Is such an approach feasible?