Running the test case on a 2-node H20 3E machine failed with internode.cu:1029 'unspecified launch failure'

I ran the test case on a 2-node H20 3E machine.
After running for a while, an error occurred. How can I troubleshoot and solve it?

- deep-ep version: 1.1.0+e6d61fc

- command line:
- `TORCH_USE_CUDA_DSA=1 CUDA_LAUNCH_BLOCKING=1 NVSHMEM_IB_ENABLE_IBGDA=1 NVSHMEM_IBGDA_NIC_HANDLER=gpu MASTER_ADDR=192.168.0.138 MASTER_PORT=7654 WORLD_SIZE=2 RANK=0 python3 tests/test_internode.py
-
- TORCH_USE_CUDA_DSA=1 CUDA_LAUNCH_BLOCKING=1 NVSHMEM_IB_ENABLE_IBGDA=1 NVSHMEM_IBGDA_NIC_HANDLER=gpu MASTER_ADDR=192.168.0.138 MASTER_PORT=7654 WORLD_SIZE=2 RANK=1 python3 tests/test_internode.py
- `

error is:

> xception raised from c10_cuda_check_implementation at /pytorch/c10/cuda/CUDAException.cpp:43 (most recent call first):
> frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string, std::allocator >) + 0x98 (0x7fef8af785e8 in /usr/local/lib/python3.10/dist-packages/torch/lib/libc10.so)
> frame #1: c10::detail::torchCheckFail(char const*, char const*, unsigned int, std::__cxx11::basic_string, std::allocator > const&) + 0xe0 (0x7fef8af0d4a2 in /usr/local/lib/python3.10/dist-packages/torch/lib/libc10.so)
> frame #2: c10::cuda::c10_cuda_check_implementation(int, char const*, char const*, int, bool) + 0x3c2 (0x7fef8b33f422 in /usr/local/lib/python3.10/dist-packages/torch/lib/libc10_cuda.so)
> frame #3: + 0x1e79f (0x7fef8b30779f in /usr/local/lib/python3.10/dist-packages/torch/lib/libc10_cuda.so)
> frame #4: + 0x20060 (0x7fef8b309060 in /usr/local/lib/python3.10/dist-packages/torch/lib/libc10_cuda.so)
> frame #5: + 0x2028c (0x7fef8b30928c in /usr/local/lib/python3.10/dist-packages/torch/lib/libc10_cuda.so)
> frame #6: + 0x457242 (0x7fef82857242 in /usr/local/lib/python3.10/dist-packages/torch/lib/libtorch_python.so)
> frame #7: c10::TensorImpl::~TensorImpl() + 0x9 (0x7fef8af52f39 in /usr/local/lib/python3.10/dist-packages/torch/lib/libc10.so)
> frame #8: + 0x7186d8 (0x7fef82b186d8 in /usr/local/lib/python3.10/dist-packages/torch/lib/libtorch_python.so)
> frame #9: + 0x718af1 (0x7fef82b18af1 in /usr/local/lib/python3.10/dist-packages/torch/lib/libtorch_python.so)
> frame #10: + 0x181390 (0x5597c1981390 in /usr/bin/python3)
> frame #11: + 0x1946c8 (0x5597c19946c8 in /usr/bin/python3)
> frame #12: + 0x1946dc (0x5597c19946dc in /usr/bin/python3)
> frame #13: + 0x1946dc (0x5597c19946dc in /usr/bin/python3)
> frame #14: + 0x1a061f (0x5597c19a061f in /usr/bin/python3)
> frame #15: + 0x15f9d6 (0x5597c195f9d6 in /usr/bin/python3)
> frame #16: + 0x2939a7 (0x5597c1a939a7 in /usr/bin/python3)
> frame #17: _PyEval_EvalFrameDefault + 0x54b7 (0x5597c197a727 in /usr/bin/python3)
> frame #18: _PyFunction_Vectorcall + 0x7c (0x5597c198b6ac in /usr/bin/python3)
> frame #19: _PyEval_EvalFrameDefault + 0x807 (0x5597c1975a77 in /usr/bin/python3)
> frame #20: _PyFunction_Vectorcall + 0x7c (0x5597c198b6ac in /usr/bin/python3)
> frame #21: _PyEval_EvalFrameDefault + 0x6c0 (0x5597c1975930 in /usr/bin/python3)
> frame #22: _PyFunction_Vectorcall + 0x7c (0x5597c198b6ac in /usr/bin/python3)
> frame #23: _PyEval_EvalFrameDefault + 0x1987 (0x5597c1976bf7 in /usr/bin/python3)
> frame #24: + 0x25a9d6 (0x5597c1a5a9d6 in /usr/bin/python3)
> frame #25: PyEval_EvalCode + 0x86 (0x5597c1a5a8a6 in /usr/bin/python3)
> frame #26: + 0x2812c8 (0x5597c1a812c8 in /usr/bin/python3)
> frame #27: + 0x27bacf (0x5597c1a7bacf in /usr/bin/python3)
> frame #28: PyRun_StringFlags + 0x81 (0x5597c1a756e1 in /usr/bin/python3)
> frame #29: PyRun_SimpleStringFlags + 0x41 (0x5597c1a75591 in /usr/bin/python3)
> frame #30: Py_RunMain + 0x3d0 (0x5597c1a748c0 in /usr/bin/python3)
> frame #31: Py_BytesMain + 0x2d (0x5597c1a4e8ed in /usr/bin/python3)
> frame #32: + 0x29d90 (0x7fef8bc29d90 in /usr/lib/x86_64-linux-gnu/libc.so.6)
> frame #33: __libc_start_main + 0x80 (0x7fef8bc29e40 in /usr/lib/x86_64-linux-gnu/libc.so.6)
> frame #34: _start + 0x25 (0x5597c1a4e7e5 in /usr/bin/python3)
>
> W0709 15:47:03.322000 886688 torch/multiprocessing/spawn.py:169] Terminating process 886753 via signal SIGTERM
> W0709 15:47:03.323000 886688 torch/multiprocessing/spawn.py:169] Terminating process 886754 via signal SIGTERM
> W0709 15:47:03.323000 886688 torch/multiprocessing/spawn.py:169] Terminating process 886755 via signal SIGTERM
> W0709 15:47:03.323000 886688 torch/multiprocessing/spawn.py:169] Terminating process 886756 via signal SIGTERM
> W0709 15:47:03.324000 886688 torch/multiprocessing/spawn.py:169] Terminating process 886757 via signal SIGTERM
> W0709 15:47:03.324000 886688 torch/multiprocessing/spawn.py:169] Terminating process 886758 via signal SIGTERM
> W0709 15:47:03.324000 886688 torch/multiprocessing/spawn.py:169] Terminating process 886760 via signal SIGTERM
> Traceback (most recent call last):
> File "/nvme3/lwq/deepep_install/DeepEP/tests/test_internode.py", line 281, in
> torch.multiprocessing.spawn(test_loop, args=(num_processes, args), nprocs=num_processes)
> File "/usr/local/lib/python3.10/dist-packages/torch/multiprocessing/spawn.py", line 340, in spawn
> return start_processes(fn, args, nprocs, join, daemon, start_method="spawn")
> File "/usr/local/lib/python3.10/dist-packages/torch/multiprocessing/spawn.py", line 296, in start_processes
> while not context.join():
> File "/usr/local/lib/python3.10/dist-packages/torch/multiprocessing/spawn.py", line 215, in join
> raise ProcessRaisedException(msg, error_index, failed_process.pid)
> torch.multiprocessing.spawn.ProcessRaisedException:
>
> -- Process 6 terminated with the following error:
> Traceback (most recent call last):
> File "/usr/local/lib/python3.10/dist-packages/torch/multiprocessing/spawn.py", line 90, in _wrap
> fn(i, *args)
> File "/nvme3/lwq/deepep_install/DeepEP/tests/test_internode.py", line 243, in test_loop
> test_main(args, i, local_rank, num_local_ranks, num_ranks, num_nodes, rank, buffer, group)
> File "/nvme3/lwq/deepep_install/DeepEP/tests/test_internode.py", line 188, in test_main
> t = bench(lambda: buffer.dispatch(**tune_args))[0]
> File "/nvme3/lwq/deepep_install/DeepEP/tests/utils.py", line 104, in bench
> fn()
> File "/nvme3/lwq/deepep_install/DeepEP/tests/test_internode.py", line 188, in
> t = bench(lambda: buffer.dispatch(**tune_args))[0]
> File "/usr/local/lib/python3.10/dist-packages/deep_ep-1.1.0+e6d61fc-py3.10-linux-x86_64.egg/deep_ep/buffer.py", line 331, in dispatch
> return self.internode_dispatch(x, handle, num_tokens_per_rank, num_tokens_per_rdma_rank, is_token_in_rank, num_tokens_per_expert,
> File "/usr/local/lib/python3.10/dist-packages/deep_ep-1.1.0+e6d61fc-py3.10-linux-x86_64.egg/deep_ep/buffer.py", line 430, in internode_dispatch
> recv_x, recv_x_scales, _, _, _, _, _, _, _, _, _, _, _, _, event = self.runtime.internode_dispatch(
> RuntimeError: Failed: CUDA error /nvme3/lwq/deepep_install/DeepEP/csrc/kernels/internode.cu:1029 'unspecified launch failure'

The logs before errors are as follows：

other info:
ibv_devinfo：
```
[root@opllm-h20-x6immzry-0000 lwq]# ibv_devinfo
hca_id: mlx5_bond_0
transport: InfiniBand (0)
fw_ver: 32.45.1020
node_guid: 7c8c:0903:00a6:82e6
sys_image_guid: 7c8c:0903:00a6:82e6
vendor_id: 0x02c9
vendor_part_id: 41692
hw_ver: 0x1
board_id: MT_0000001093
phys_port_cnt: 1
port: 1
state: PORT_ACTIVE (4)
max_mtu: 4096 (5)
active_mtu: 4096 (5)
sm_lid: 0
port_lid: 0
port_lmc: 0x00
link_layer: Ethernet

hca_id: mlx5_bond_1
transport: InfiniBand (0)
fw_ver: 32.45.1020
node_guid: 7c8c:0903:00a6:7056
sys_image_guid: 7c8c:0903:00a6:7056
vendor_id: 0x02c9
vendor_part_id: 41692
hw_ver: 0x1
board_id: MT_0000001093
phys_port_cnt: 1
port: 1
state: PORT_ACTIVE (4)
max_mtu: 4096 (5)
active_mtu: 4096 (5)
sm_lid: 0
port_lid: 0
port_lmc: 0x00
link_layer: Ethernet

hca_id: mlx5_bond_2
transport: InfiniBand (0)
fw_ver: 32.45.1020
node_guid: 7c8c:0903:00a6:7686
sys_image_guid: 7c8c:0903:00a6:7686
vendor_id: 0x02c9
vendor_part_id: 41692
hw_ver: 0x1
board_id: MT_0000001093
phys_port_cnt: 1
port: 1
state: PORT_ACTIVE (4)
max_mtu: 4096 (5)
active_mtu: 4096 (5)
sm_lid: 0
port_lid: 0
port_lmc: 0x00
link_layer: Ethernet

hca_id: mlx5_bond_3
transport: InfiniBand (0)
fw_ver: 32.45.1020
node_guid: 7c8c:0903:00a6:89f2
sys_image_guid: 7c8c:0903:00a6:89f2
vendor_id: 0x02c9
vendor_part_id: 41692
hw_ver: 0x1
board_id: MT_0000001093
phys_port_cnt: 1
port: 1
state: PORT_ACTIVE (4)
max_mtu: 4096 (5)
active_mtu: 4096 (5)
sm_lid: 0
port_lid: 0
port_lmc: 0x00
link_layer: Ethernet
```

`root@opllm-h20-x6immzry-0000:/nvme3/lwq/Big_and_EP# nvshmem-info -a
NVSHMEM v3.2.5

Build Information:
CUDA API 12080
CUDA Driver 12080
Build Timestamp Jul 7 2025 11:56:31
Build Variables
NVSHMEM_DEBUG=OFF NVSHMEM_DEVEL=OFF NVSHMEM_DEFAULT_PMI2=OFF
NVSHMEM_DEFAULT_PMIX=OFF NVSHMEM_DEFAULT_UCX=OFF
NVSHMEM_ENABLE_ALL_DEVICE_INLINING=OFF NVSHMEM_GPU_COLL_USE_LDST=OFF
NVSHMEM_IBGDA_SUPPORT=ON NVSHMEM_IBGDA_SUPPORT_GPUMEM_ONLY=OFF
NVSHMEM_IBDEVX_SUPPORT=OFF NVSHMEM_IBRC_SUPPORT=ON
NVSHMEM_MPI_SUPPORT=OFF NVSHMEM_NVTX=ON NVSHMEM_PMIX_SUPPORT=OFF
NVSHMEM_SHMEM_SUPPORT=OFF NVSHMEM_TEST_STATIC_LIB=OFF
NVSHMEM_TIMEOUT_DEVICE_POLLING=OFF NVSHMEM_TRACE=OFF NVSHMEM_UCX_SUPPORT=OFF
NVSHMEM_USE_DLMALLOC=OFF NVSHMEM_USE_NCCL=OFF NVSHMEM_USE_GDRCOPY=ON
NVSHMEM_VERBOSE=OFF CUDA_HOME=/usr/local/cuda GDRCOPY_HOME=/usr/local/gdrdrv
LIBFABRIC_HOME=/usr/local/libfabric MPI_HOME=/usr/local/ompi
NCCL_HOME=/usr/local/nccl NVSHMEM_PREFIX=/usr/local/nvshmem PMIX_HOME=/usr
SHMEM_HOME=/usr/local/ompi UCX_HOME=/usr/local/ucx

Standard options:
NVSHMEM_VERSION false (type: bool, default: false)
Print library version at startup
NVSHMEM_INFO false (type: bool, default: false)
Print environment variable options at startup
NVSHMEM_DISABLE_NVLS false (type: bool, default: false)
Disable NVLS SHARP resources for collectives, even if available for platform
NVSHMEM_SYMMETRIC_SIZE 1073741824 (type: size, default: 1073741824)
Specifies the size (in bytes) of the symmetric heap memory per PE. The
size is implementation-defined and must be at least as large as the integer
ceiling of the product of the numeric prefix and the scaling factor. The
character suffixes for the scaling factor are as follows:

* k or K multiplies by 2^10 (kibibytes)
* m or M multiplies by 2^20 (mebibytes)
* g or G multiplies by 2^30 (gibibytes)
* t or T multiplies by 2^40 (tebibytes)

For example, string '20m' is equivalent to the integer value 20971520, or 20
mebibytes. Similarly the string '3.1M' is equivalent to the integer value
3250586. Only one multiplier is recognized and any characters following the
multiplier are ignored, so '20kk' will not produce the same result as '20m'.
Usage of string '.5m' will yield the same result as the string '0.5m'.
An invalid value for NVSHMEM_SYMMETRIC_SIZE is an error, which the NVSHMEM
library shall report by either returning a nonzero value from
nvshmem_init_thread or causing program termination.
NVSHMEM_DEBUG "" (type: string, default: "")
Set to enable debugging messages.
Optional values: VERSION, WARN, INFO, ABORT, TRACE

Bootstrap options:
NVSHMEM_BOOTSTRAP "PMI" (type: string, default: "PMI")
Name of the default bootstrap that should be used to initialize NVSHMEM.
Allowed values: PMI, MPI, SHMEM, plugin, UID
NVSHMEM_BOOTSTRAP_PMI "PMI" (type: string, default: "PMI")
Name of the PMI bootstrap that should be used to initialize NVSHMEM.
Allowed values: PMI, PMI-2, PMIX
NVSHMEM_BOOTSTRAP_PLUGIN "" (type: string, default: "")
Absolute path to or name of the bootstrap plugin file to load when
NVSHMEM_BOOTSTRAP=plugin is specified
NVSHMEM_BOOTSTRAP_MPI_PLUGIN "nvshmem_bootstrap_mpi.so.3" (type: string, default: "nvshmem_bootstrap_mpi.so.3")
Absolute path to or name of the MPI bootstrap plugin file.
NVSHMEM will search for the plugin based on linux linker priorities. See man
dlopen
NVSHMEM_BOOTSTRAP_SHMEM_PLUGIN "nvshmem_bootstrap_shmem.so.3" (type: string, default: "nvshmem_bootstrap_shmem.so.3")
Absolute path to or name of the SHMEM bootstrap plugin file.
NVSHMEM will search for the plugin based on linux linker priorities. See man
dlopen
NVSHMEM_BOOTSTRAP_PMI_PLUGIN "nvshmem_bootstrap_pmi.so.3" (type: string, default: "nvshmem_bootstrap_pmi.so.3")
Absolute path to or name of the PMI bootstrap plugin file.
NVSHMEM will search for the plugin based on linux linker priorities. See man
dlopen
NVSHMEM_BOOTSTRAP_PMI2_PLUGIN "nvshmem_bootstrap_pmi2.so.3" (type: string, default: "nvshmem_bootstrap_pmi2.so.3")
Absolute path to or name of the PMI-2 bootstrap plugin file.
NVSHMEM will search for the plugin based on linux linker priorities. See man
dlopen
NVSHMEM_BOOTSTRAP_PMIX_PLUGIN "nvshmem_bootstrap_pmix.so.3" (type: string, default: "nvshmem_bootstrap_pmix.so.3")
Absolute path to or name of the PMIx bootstrap plugin file.
NVSHMEM will search for the plugin based on linux linker priorities. See man
dlopen
NVSHMEM_BOOTSTRAP_UID_PLUGIN "nvshmem_bootstrap_uid.so.3" (type: string, default: "nvshmem_bootstrap_uid.so.3")
Absolute path to or name of the UID bootstrap plugin file.
NVSHMEM will search for the plugin based on linux linker priorities. See man
dlopen

Additional options:
NVSHMEM_CUDA_PATH "" (type: string, default: "")
Path to directory containing libcuda.so (for use when not in default location)
NVSHMEM_DEBUG_ATTACH_DELAY 0 (type: int, default: 0)
Delay (in seconds) during the first call to NVSHMEM_INIT to allow for attaching
a debuggger (Default 0)
NVSHMEM_DEBUG_FILE "" (type: string, default: "")
Debugging output filename, may contain %h for hostname and %p for pid
NVSHMEM_MAX_TEAMS 32 (type: long, default: 32)
Maximum number of simultaneous teams allowed
NVSHMEM_MAX_MEMORY_PER_GPU 137438953472 (type: size, default: 137438953472)
Maximum memory per GPU
NVSHMEM_DISABLE_CUDA_VMM false (type: bool, default: false)
Disable use of CUDA VMM for P2P memory mapping. By default, CUDA VMM is enabled
on x86 and disabled on P9. CUDA VMM feature in NVSHMEM requires CUDA RT version
and CUDA Driver version to be greater than or equal to 11.3.
NVSHMEM_DISABLE_P2P false (type: bool, default: false)
Disable P2P connectivity of GPUs even when available
NVSHMEM_IGNORE_CUDA_MPS_ACTIVE_THREAD_PERCENTAGE false (type: bool, default: false)
When doing Multi-Process Per GPU (MPG) run, full API support is available only
if sum of CUDA_MPS_ACTIVE_THREAD_PERCENTAGE of processes running on a GPU is <=
100%. Through this variable, user can request NVSHMEM runtime to ignore the
active thread percentage and allow full MPG support. Users enable it at their
own risk as NVSHMEM might deadlock.
NVSHMEM_CUMEM_GRANULARITY 536870912 (type: size, default: 536870912)
Granularity for cuMemAlloc/cuMemCreate
NVSHMEM_PROXY_REQUEST_BATCH_MAX 32 (type: int, default: 32)
Maxmum number of requests that the proxy thread processes in a single iteration
of the progress loop.

Collectives options:
NVSHMEM_DISABLE_NCCL false (type: bool, default: false)
Disable use of NCCL for collective operations
NVSHMEM_BARRIER_DISSEM_KVAL 2 (type: int, default: 2)
Radix of the dissemination algorithm used for barriers
NVSHMEM_BARRIER_TG_DISSEM_KVAL 2 (type: int, default: 2)
Radix of the dissemination algorithm used for thread group barriers
NVSHMEM_FCOLLECT_LL_THRESHOLD 2048 (type: size, default: 2048)
Message size threshold up to which fcollect LL algo will be used

NVSHMEM_REDUCE_SCRATCH_SIZE 524288 (type: size, default: 524288)
Amount of symmetric heap memory (minimum 16B, multiple of 8B) reserved by
runtime for every team to implement reduce and reducescatter collectives

NVSHMEM_BCAST_ALGO 0 (type: int, default: 0)
Broadcast algorithm to be used.
* 0 - use default algorithm selection strategy

NVSHMEM_REDMAXLOC_ALGO 1 (type: int, default: 1)
Reduction algorithm to be used for MAXLOC operation.
* 1 - default, flag alltoall algorithm
* 2 - flat reduce + flat bcast
* 3 - topo-aware two-level reduce + topo-aware bcast

Transport options:
NVSHMEM_REMOTE_TRANSPORT "ibrc" (type: string, default: "ibrc")
Selected transport for remote operations: ibrc, ucx, libfabric, ibdevx, none
NVSHMEM_ENABLE_NIC_PE_MAPPING false (type: bool, default: false)
When not set or set to 0, a PE is assigned the NIC on the node that is closest
to it by distance. When set to 1, NVSHMEM either assigns NICs to PEs on a
round-robin basis or uses NVSHMEM_HCA_PE_MAPPING or NVSHMEM_HCA_LIST when they
are specified.
NVSHMEM_DISABLE_LOCAL_ONLY_PROXY false (type: bool, default: false)
When running on an NVLink-only configuaration (No-IB, No-UCX), completely
disable the proxy thread. This will disable device side global exit and device
side wait timeout polling (enabled by NVSHMEM_TIMEOUT_DEVICE_POLLING build-time
variable) because these are processed by the proxy thread.
NVSHMEM_IB_ENABLE_IBGDA false (type: bool, default: false)
Set to enable GPU-initiated communication transport.

Hidden options:
NVSHMEM_INFO_HIDDEN true (type: bool, default: false)
Print hidden environment variable options at startup
NVSHMEM_DISABLE_NVLS_SHARING true (type: bool, default: true)
Disable NVLS SHARP resource sharing for user-defined teams
NVSHMEM_HEAP_KIND "DEVICE" (type: string, default: "DEVICE")
Specify the memory kind used by the NVSHMEM symmetric heap.
Allowed values: VIDMEM, SYSMEM
NVSHMEM_ENABLE_RAIL_OPT false (type: bool, default: false)
Enable Rail Optimization when heap is in SYSMEM
NVSHMEM_BOOTSTRAP_TWO_STAGE false (type: bool, default: false)
Ignore CUDA device setting during initialization,forcing two-stage
initialization
NVSHMEM_DEBUG_SUBSYS "" (type: string, default: "")
Comma separated list of debugging message sources. Prefix with '^' to exclude.
Values: INIT, COLL, P2P, PROXY, TRANSPORT, MEM, BOOTSTRAP, TOPO, UTIL, ALL
NVSHMEM_ENABLE_ERROR_CHECKS false (type: bool, default: false)
Enable error checks
NVSHMEM_DISABLE_MNNVL false (type: bool, default: false)
Disable MNNVL connectivity for GPUs even when available
NVSHMEM_CUMEM_HANDLE_TYPE "FILE_DESCRIPTOR" (type: string, default: "FILE_DESCRIPTOR")
Handle type for cuMemCreate. Supported are - FABRIC or FILE_DESCRIPTOR
NVSHMEM_BYPASS_ACCESSIBILITY_CHECK false (type: bool, default: false)
Bypass peer GPU accessbility checks
NVSHMEM_FCOLLECT_NTHREADS 512 (type: int, default: 512)
Sets number of threads per block for fcollect collective.
By default, if no env is set, default value is min(max_occupancy per CTA, msg
size per PE).
If env is specified, value overrides the default irrespective of max occupancy
per CTA

NVSHMEM_REDUCESCATTER_NTHREADS 512 (type: int, default: 512)
Sets number of threads per block for reducescatter collective.
By default, if no env is set, default value is min(max_occupancy per CTA, msg
size per PE).
If env is specified, value overrides the default irrespective of max occupancy
per CTA

NVSHMEM_MAX_CTAS 1 (type: int, default: 1)
Sets number of blocks per grid for host onstream collective.
By default, if no env is set, default value to 1 CTA
If env is specified, value overrides the default value

NVSHMEM_REDUCE_RECEXCH_KVAL 2 (type: int, default: 2)
Radix of the recursive exchange reduction algorithm
NVSHMEM_FCOLLECT_LL128_THRESHOLD 0 (type: size, default: 0)
Message size threshold up to which the fcollect LL128 algo will be used.
LL128 will be used only when FCOLLECT_LL_THRESHOLD < size
NVSHMEM_FCOLLECT_NVLS_THRESHOLD 16777216 (type: size, default: 16777216)
Message size threshold up to which fcollect NVLS algo will be used

NVSHMEM_REDUCESCATTER_NVLS_THRESHOLD 16777216 (type: size, default: 16777216)
Message size threshold up to which reducescatter NVLS algo will be used

NVSHMEM_BCAST_TREE_KVAL 2 (type: int, default: 2)
Radix of the broadcast tree algorithm
NVSHMEM_FCOLLECT_ALGO 0 (type: int, default: 0)
Fcollect algorithm to be used.
* 0 - use default algorithm selection strategy

NVSHMEM_REDUCE_ALGO 0 (type: int, default: 0)
Allreduce algorithm to be used.
* 0/1 - use default algorithm selection strategy

NVSHMEM_REDUCE_NVLS_THRESHOLD 2048 (type: int, default: 2048)
Message size threshold up to which allreduce one-shot algo will be used

NVSHMEM_REDUCESCATTER_ALGO 0 (type: int, default: 0)
Reduce Scatter algorithm to be used.
* 0 - use default algorithm selection strategy

NVSHMEM_ASSERT_ATOMICS_SYNC false (type: bool, default: false)
Bypass flush on wait_until at target
NVSHMEM_BYPASS_FLUSH false (type: bool, default: false)
Bypass flush in proxy when enforcing consistency

NVTX options:
NVSHMEM_NVTX "off" (type: string, default: "off")
Set to enable NVTX instrumentation. Accepts a comma separated list of
instrumentation groups. By default the NVTX instrumentation is disabled.
init : library setup
alloc : memory management
launch : kernel launch routines
coll : collective communications
wait : blocking point-to-point synchronization
wait_on_stream : point-to-point synchronization (on stream)
test : non-blocking point-to-point synchronization
memorder : memory ordering (quiet, fence)
quiet_on_stream : nvshmemx_quiet_on_stream
atomic_fetch : fetching atomic memory operations
atomic_set : non-fetchong atomic memory operations
rma_blocking : blocking remote memory access operations
rma_nonblocking : non-blocking remote memory access operations
proxy : activity of the proxy thread
common : init,alloc,launch,coll,memorder,wait,atomic_fetch,rma_blocking,proxy
all : all groups
off : disable all NVTX instrumentation
`

Start with tests/test_internode.py, especially test_main around line 188 and test_loop around line 243, then inspect tests/utils.py:104 and csrc/kernels/internode.cu:1029. Reproduce the two-node command with the supplied environment and logs, narrowing the failure to the internode dispatch path. Done means the cause is identified and the internode test completes successfully on the reported setup.

Written by the indexing model from the issue text.