I wanna connect 2 of my RTX 2070 Super via NVLink bridge, but it doesn’t work.
My config:
- Motherboard: Gigabyte Z390 GAMING SLI
- Intel Core i9-9900K CPU
- Two GPU NVIDIA RTX 2070 SUPER
- Ubuntu 18.04
- Nvidia driver: 440.33.01
- CUDA 10.2, CUDNN 7.6.5
The command nvidia-smi nvlink -s
gives out:
GPU 0: GeForce RTX 2070 SUPER (UUID: GPU-913c9fff-8680-7958-692b-8c637e145a26)
Link 0: <inactive>
GPU 1: GeForce RTX 2070 SUPER (UUID: GPU-045ff14d-996c-5c48-2314-a9184260de02)
Link 0: <inactive>
The command nvidia-smi topo -m
gives out:
GPU0 GPU1 CPU Affinity
GPU0 X PHB 0-15
GPU1 PHB X 0-15
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
The test nvidia simpleP2P gives out:
Checking for multiple GPUs...
CUDA-capable device count: 2
Checking GPU(s) for support of peer to peer memory access...
> Peer access from GeForce RTX 2070 SUPER (GPU0) -> GeForce RTX 2070 SUPER (GPU1) : No
> Peer access from GeForce RTX 2070 SUPER (GPU1) -> GeForce RTX 2070 SUPER (GPU0) : No
Two or more GPUs with Peer-to-Peer access capability are required for /usr/local/cuda/samples/bin/x86_64/linux/release/simpleP2P.
Peer to Peer access is not available amongst GPUs in the system, waiving test.
I found 2 similar questions on the forum, unanswered:
- can't enable nvlink on ubuntu with two RTX2070S, but windows can
- No Peer-To-Peer-Access and slow CUDA execution with NVLink Bridge present on Ubuntu 20.04
Please, could anyone help me to solve this issue?