Named Symbol not Found Error on B200

[osayamenja]

Hey,

Nice work on this project! I cannot run torchcomms with NCCLX as a backend on B200. I tried running the quick-start AllReduce example but it fails with the following error.

Traceback (most recent call last):
  File "/root/dcl/benchmark/example.py", line 42, in <module>
    main()
  File "/root/dcl/benchmark/example.py", line 21, in main
    tensor = torch.full(
             ^^^^^^^^^^^
torch.AcceleratorError: CUDA error: named symbol not found
Search for `cudaErrorSymbolNotFound' in https://docs.nvidia.com/cuda/cuda-runtime-api/group__CUDART__TYPES.html for more information.
CUDA kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect.
For debugging consider passing CUDA_LAUNCH_BLOCKING=1
Compile with `TORCH_USE_CUDA_DSA` to enable device-side assertions.

I installed torchcomms using pip install --pre torch torchcomms --index-url https://download.pytorch.org/whl/nightly/cu128

Note:

• Changing the backend to NCCL makes the example work.
• I have successfully used NCCLX on Ampere and Hopper, so my intuition is that the error is specific to the Blackwell.

Machine Details

nvidia-smi
Sun Nov  9 01:32:20 2025
+-----------------------------------------------------------------------------------------+
| NVIDIA-SMI 570.195.03             Driver Version: 570.195.03     CUDA Version: 12.8     |
|-----------------------------------------+------------------------+----------------------+
| GPU  Name                 Persistence-M | Bus-Id          Disp.A | Volatile Uncorr. ECC |
| Fan  Temp   Perf          Pwr:Usage/Cap |           Memory-Usage | GPU-Util  Compute M. |
|                                         |                        |               MIG M. |
|=========================================+========================+======================|
|   0  NVIDIA B200                    On  |   00000000:29:00.0 Off |                    0 |
| N/A   40C    P0            147W / 1000W |       0MiB / 183359MiB |      0%      Default |
|                                         |                        |             Disabled |
+-----------------------------------------+------------------------+----------------------+
|   1  NVIDIA B200                    On  |   00000000:3A:00.0 Off |                    0 |
| N/A   31C    P0            144W / 1000W |       0MiB / 183359MiB |      0%      Default |
|                                         |                        |             Disabled |
+-----------------------------------------+------------------------+----------------------+

+-----------------------------------------------------------------------------------------+
| Processes:                                                                              |
|  GPU   GI   CI              PID   Type   Process name                        GPU Memory |
|        ID   ID                                                               Usage      |
|=========================================================================================|
|  No running processes found                                                             |
+-----------------------------------------------------------------------------------------+

nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2025 NVIDIA Corporation
Built on Fri_Feb_21_20:23:50_PST_2025
Cuda compilation tools, release 12.8, V12.8.93
Build cuda_12.8.r12.8/compiler.35583870_0