amd gfx1151 throw error when use `--enable-triton-backend` run krea2 template

### Custom Node Testing

- [x] I have tried disabling custom nodes and the issue persists (see [how to disable custom nodes](https://docs.comfy.org/troubleshooting/custom-node-issues#step-1%3A-test-with-all-custom-nodes-disabled) if you need help)

### Expected Behavior

without --enable-triton-backend workflow can run success,after add throw error

### Actual Behavior

throw error

### Steps to Reproduce

install
```
pip install --index-url https://rocm.nightlies.amd.com/whl-multi-arch/ \
"torch[device-gfx1151]" \
"torchvision[device-gfx1151]" \
torchaudio
```
run
python main.py --disable-mmap --enable-manager --listen 0.0.0.0 --enable-triton-backend

select `krea2 template workflow`
run

### Debug Logs

```powershell
(.venv) chen@amd395 ~/ComfyUI ((v0.28.0)) [0|SIGINT]> AMD_SERIALIZE_KERNEL=3 python main.py --disable-mmap --enable-manager --listen 0.0.0.0 --enable-triton-backend --disable-all-custom-nodes
[INFO] setup plugin alembic.autogenerate.schemas
[INFO] setup plugin alembic.autogenerate.tables
[INFO] setup plugin alembic.autogenerate.types
[INFO] setup plugin alembic.autogenerate.constraints
[INFO] setup plugin alembic.autogenerate.defaults
[INFO] setup plugin alembic.autogenerate.comments
[START] Security scan
[INFO] [ComfyUI-Manager] Using uv as Python module for pip operations.
[DONE] Security scan
** ComfyUI startup time: 2026-07-23 00:07:26.456
** Platform: Linux
** Python version: 3.14.4 (main, Jun 18 2026, 14:25:02) [GCC 15.2.0]
** Python executable: /home/chen/ComfyUI/.venv/bin/python
** ComfyUI Path: /home/chen/ComfyUI
** ComfyUI Base Folder Path: /home/chen/ComfyUI
** User directory: /home/chen/ComfyUI/user
** ComfyUI-Manager config path: /home/chen/ComfyUI/user/__manager/config.ini
** Log path: /home/chen/ComfyUI/user/comfyui.log
[INFO] [PRE] ComfyUI-Manager
[INFO] Found triton 3.7.1. Enabling comfy-kitchen triton backend.
[INFO] Found comfy_kitchen backend cuda: {'available': True, 'disabled': True, 'unavailable_reason': None, 'capabilities': ['adaln', 'apply_rope', 'apply_rope1', 'apply_rope_split_half', 'apply_rope_split_half1', 'convrot_w4a4_linear', 'dequantize_convrot_w4a4_weight', 'dequantize_int8_convrot_weight', 'dequantize_int8_convrot_weight_dtype', 'dequantize_int8_simple', 'dequantize_int8_simple_dtype', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'gemv_awq_w4a16', 'prepare_int4_weight_for_int8_linear', 'quantize_and_rotate_rowwise', 'quantize_convrot_w4a4_weight', 'quantize_int8_convrot_weight', 'quantize_int8_rowwise', 'quantize_int8_tensorwise', 'quantize_mxfp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8', 'quantize_svdquant_w4a4', 'rms_rope', 'rms_rope1', 'rms_rope_split_half', 'rms_rope_split_half1', 'scaled_mm_svdquant_w4a4', 'stochastic_rounding_fp8']}
[INFO] Found comfy_kitchen backend eager: {'available': True, 'disabled': False, 'unavailable_reason': None, 'capabilities': ['adaln', 'apply_rope', 'apply_rope1', 'apply_rope_split_half', 'apply_rope_split_half1', 'convrot_w4a4_linear', 'dequantize_convrot_w4a4_weight', 'dequantize_int8_convrot_weight', 'dequantize_int8_convrot_weight_dtype', 'dequantize_int8_simple', 'dequantize_int8_simple_dtype', 'dequantize_mxfp8', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'gemv_awq_w4a16', 'int8_linear', 'prepare_int4_weight_for_int8_linear', 'quantize_and_rotate_rowwise', 'quantize_convrot_w4a4_weight', 'quantize_int8_convrot_weight', 'quantize_int8_rowwise', 'quantize_int8_tensorwise', 'quantize_mxfp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8', 'quantize_svdquant_w4a4', 'rms_rope', 'rms_rope1', 'rms_rope_split_half', 'rms_rope_split_half1', 'scaled_mm_mxfp8', 'scaled_mm_nvfp4', 'scaled_mm_svdquant_w4a4', 'stochastic_rounding_fp8']}
[INFO] Found comfy_kitchen backend triton: {'available': True, 'disabled': False, 'unavailable_reason': None, 'capabilities': ['adaln', 'apply_rope', 'apply_rope1', 'apply_rope_split_half', 'apply_rope_split_half1', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'int8_linear', 'quantize_and_rotate_rowwise', 'quantize_int8_rowwise', 'quantize_mxfp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8']}
[INFO] Checkpoint files will always be loaded safely.
[INFO] Total VRAM 122880 MB, total RAM 128072 MB
[INFO] pytorch version: 2.12.0+rocm7.15.0a20260721
[INFO] Set: torch.backends.cudnn.enabled = False for better AMD performance.
[INFO] AMD arch: gfx1151
[INFO] ROCm version: (7, 15)
[INFO] Set vram state to: NORMAL_VRAM
[INFO] Device: cuda:0 AMD Radeon 8060S Graphics : native
[INFO] Using async weight offloading with 2 streams
[INFO] Enabled pinned memory 115264.0
[INFO] Using pytorch attention
[INFO] Python version: 3.14.4 (main, Jun 18 2026, 14:25:02) [GCC 15.2.0]
[INFO] ComfyUI version: 0.28.0
[INFO] comfy-aimdo version: 0.4.10
[INFO] comfy-kitchen version: 0.2.20
[INFO] comfyui-frontend-package version: 1.45.21
[INFO] comfyui-workflow-templates version: 0.11.9
[INFO] comfyui-embedded-docs version: 0.5.8
[INFO] comfy-kitchen version: 0.2.20
[INFO] comfy-aimdo version: 0.4.10
[INFO] [Prompt Server] web root: /home/chen/ComfyUI/.venv/lib/python3.14/site-packages/comfyui_frontend_package/static
[INFO] Asset seeder disabled
[INFO] [START] ComfyUI-Manager
[ComfyUI-Manager] Using GitPython backend
[INFO] [ComfyUI-Manager] network_mode: personal_cloud
[WARNING] [ComfyUI-Manager] The matrix sharing feature has been disabled because the `matrix-nio` dependency is not installed.
To use this feature, please run the following command:
/home/chen/ComfyUI/.venv/bin/python -m pip install matrix-nio

[INFO] No OpenGL_accelerate module loaded: No module named 'OpenGL_accelerate'
[INFO] Skipping loading of custom nodes
[INFO] Context impl SQLiteImpl.
[INFO] Will assume non-transactional DDL.
[INFO] Using RAM pressure cache.
[INFO] Starting server

[INFO] To see the GUI go to: http://0.0.0.0:8188
[INFO] got prompt
[INFO] Using split attention in VAE
[INFO] Using split attention in VAE
[INFO] VAE load device: cuda:0, offload device: cpu, dtype: torch.bfloat16
[INFO] Found quantization metadata version 1
[INFO] Using MixedPrecisionOps for text encoder
[INFO] Requested to load Krea2TEModel_
[INFO] loaded completely; 4999.47 MB loaded, full load: True
[INFO] CLIP/text encoder model load device: cuda:0, offload device: cpu, current: cuda:0, dtype: torch.float16
Generating tokens: 70%|██████████████████████████████████████████████████████████████████████████████████████████▏ | 358/512 [00:50<00:21, 7.13it/s]
[INFO] Found quantization metadata version 1
[INFO] Detected mixed precision quantization
[INFO] Using mixed precision operations
[INFO] Native ops: convrot_w4a4, int8_tensorwise , emulated ops: float8_e5m2, float8_e4m3fn, nvfp4, mxfp8
[INFO] model weight dtype torch.bfloat16, manual cast: torch.bfloat16
[INFO] model_type FLUX
[INFO] Requested to load Krea2
[INFO] loaded completely; 116042.32 MB usable, 12532.86 MB loaded, full load: True
0%| | 0/8 [00:00 returned a result with an exception set
[ERROR] RuntimeError: Triton Error [HIP]: Code: 719, Messsage: unspecified launch failure

The above exception was the direct cause of the following exception:

Traceback (most recent call last):
File "/home/chen/ComfyUI/execution.py", line 543, in execute
output_data, output_ui, has_subgraph, has_pending_tasks = await get_output_data(prompt_id, unique_id, obj, input_data_all, execution_block_cb=execution_block_cb, pre_execute_cb=pre_execute_cb, v3_data=v3_data)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/chen/ComfyUI/execution.py", line 342, in get_output_data
return_values = await _async_map_node_over_list(prompt_id, unique_id, obj, input_data_all, obj.FUNCTION, allow_interrupt=True, execution_block_cb=execution_block_cb, pre_execute_cb=pre_execute_cb, v3_data=v3_data)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/chen/ComfyUI/execution.py", line 316, in _async_map_node_over_list
await process_inputs(input_dict, i)
File "/home/chen/ComfyUI/execution.py", line 304, in process_inputs
result = f(**inputs)
File "/home/chen/ComfyUI/nodes.py", line 1607, in sample
return common_ksampler(model, seed, steps, cfg, sampler_name, scheduler, positive, negative, latent_image, denoise=denoise)
File "/home/chen/ComfyUI/nodes.py", line 1571, in common_ksampler
samples = comfy.sample.sample(model, noise, steps, cfg, sampler_name, scheduler, positive, negative, latent_image,
denoise=denoise, disable_noise=disable_noise, start_step=start_step, last_step=last_step,
force_full_denoise=force_full_denoise, noise_mask=noise_mask, callback=callback, disable_pbar=disable_pbar, seed=seed)
File "/home/chen/ComfyUI/comfy/sample.py", line 74, in sample
samples = sampler.sample(noise, positive, negative, cfg=cfg, latent_image=latent_image, start_step=start_step, last_step=last_step, force_full_denoise=force_full_denoise, denoise_mask=noise_mask, sigmas=sigmas, callback=callback, disable_pbar=disable_pbar, seed=seed)
File "/home/chen/ComfyUI/comfy/samplers.py", line 1444, in sample
return sample(self.model, noise, positive, negative, cfg, self.device, sampler, sigmas, self.model_options, latent_image=latent_image, denoise_mask=denoise_mask, callback=callback, disable_pbar=disable_pbar, seed=seed)
File "/home/chen/ComfyUI/comfy/samplers.py", line 1334, in sample
return cfg_guider.sample(noise, latent_image, sampler, sigmas, denoise_mask, callback, disable_pbar, seed)
~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/chen/ComfyUI/comfy/samplers.py", line 1316, in sample
output = executor.execute(noise, latent_image, sampler, sigmas, denoise_mask, callback, disable_pbar, seed, latent_shapes=latent_shapes)
File "/home/chen/ComfyUI/comfy/patcher_extension.py", line 113, in execute
return self.original(*args, **kwargs)
~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^
File "/home/chen/ComfyUI/comfy/samplers.py", line 1254, in outer_sample
output = self.inner_sample(noise, latent_image, device, sampler, sigmas, denoise_mask, callback, disable_pbar, seed, latent_shapes=latent_shapes)
File "/home/chen/ComfyUI/comfy/samplers.py", line 1229, in inner_sample
samples = executor.execute(self, sigmas, extra_args, callback, noise, latent_image, denoise_mask, disable_pbar)
File "/home/chen/ComfyUI/comfy/patcher_extension.py", line 113, in execute
return self.original(*args, **kwargs)
~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^
File "/home/chen/ComfyUI/comfy/samplers.py", line 999, in sample
samples = self.sampler_function(model_k, noise, sigmas, extra_args=extra_args, callback=k_callback, disable=disable_pbar, **self.extra_options)
File "/home/chen/ComfyUI/.venv/lib/python3.14/site-packages/torch/utils/_contextlib.py", line 124, in decorate_context
return func(*args, **kwargs)
File "/home/chen/ComfyUI/comfy/k_diffusion/sampling.py", line 205, in sample_euler
denoised = model(x, sigma_hat * s_in, **extra_args)
File "/home/chen/ComfyUI/comfy/samplers.py", line 639, in __call__
out = self.inner_model(x, sigma, model_options=model_options, seed=seed)
File "/home/chen/ComfyUI/comfy/samplers.py", line 1202, in __call__
return self.outer_predict_noise(*args, **kwargs)
~~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^
File "/home/chen/ComfyUI/comfy/samplers.py", line 1209, in outer_predict_noise
).execute(x, timestep, model_options, seed)
~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/chen/ComfyUI/comfy/patcher_extension.py", line 113, in execute
return self.original(*args, **kwargs)
~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^
File "/home/chen/ComfyUI/comfy/samplers.py", line 1212, in predict_noise
return sampling_function(self.inner_model, x, timestep, self.conds.get("negative", None), self.conds.get("positive", None), self.cfg, model_options=model_options, seed=seed)
File "/home/chen/ComfyUI/comfy/samplers.py", line 619, in sampling_function
out = calc_cond_batch(model, conds, x, timestep, model_options)
File "/home/chen/ComfyUI/comfy/samplers.py", line 210, in calc_cond_batch
return _calc_cond_batch_outer(model, conds, x_in, timestep, model_options)
File "/home/chen/ComfyUI/comfy/samplers.py", line 218, in _calc_cond_batch_outer
return executor.execute(model, conds, x_in, timestep, model_options)
~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/chen/ComfyUI/comfy/patcher_extension.py", line 113, in execute
return self.original(*args, **kwargs)
~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^
File "/home/chen/ComfyUI/comfy/samplers.py", line 334, in _calc_cond_batch
output = model.apply_model(input_x, timestep_, **c).chunk(batch_chunks)
~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/chen/ComfyUI/comfy/model_base.py", line 192, in apply_model
return comfy.patcher_extension.WrapperExecutor.new_class_executor(
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
...<2 lines>...
comfy.patcher_extension.get_all_wrappers(comfy.patcher_extension.WrappersMP.APPLY_MODEL, transformer_options)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
).execute(x, t, c_concat, c_crossattn, control, transformer_options, **kwargs)
~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/chen/ComfyUI/comfy/patcher_extension.py", line 113, in execute
return self.original(*args, **kwargs)
~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^
File "/home/chen/ComfyUI/comfy/model_base.py", line 236, in _apply_model
model_output = self.diffusion_model(xc, t, context=context, control=control, transformer_options=transformer_options, **extra_conds)
File "/home/chen/ComfyUI/.venv/lib/python3.14/site-packages/torch/nn/modules/module.py", line 1778, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^
File "/home/chen/ComfyUI/.venv/lib/python3.14/site-packages/torch/nn/modules/module.py", line 1789, in _call_impl
return forward_call(*args, **kwargs)
File "/home/chen/ComfyUI/comfy/ldm/krea2/model.py", line 225, in forward
return comfy.patcher_extension.WrapperExecutor.new_class_executor(
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
...<2 lines>...
comfy.patcher_extension.get_all_wrappers(comfy.patcher_extension.WrappersMP.DIFFUSION_MODEL, transformer_options),
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
).execute(x, timesteps, context, attention_mask, transformer_options, **kwargs)
~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/chen/ComfyUI/comfy/patcher_extension.py", line 113, in execute
return self.original(*args, **kwargs)
~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^
File "/home/chen/ComfyUI/comfy/ldm/krea2/model.py", line 270, in _forward
combined = block(combined, tvec, freqs, None, transformer_options=transformer_options)
File "/home/chen/ComfyUI/.venv/lib/python3.14/site-packages/torch/nn/modules/module.py", line 1778, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^
File "/home/chen/ComfyUI/.venv/lib/python3.14/site-packages/torch/nn/modules/module.py", line 1789, in _call_impl
return forward_call(*args, **kwargs)
File "/home/chen/ComfyUI/comfy/ldm/krea2/model.py", line 163, in forward
x = x + pregate * self.attn((1 + prescale) * self.prenorm(x) + preshift, freqs, mask, transformer_options=transformer_options)
~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/chen/ComfyUI/.venv/lib/python3.14/site-packages/torch/nn/modules/module.py", line 1778, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^
File "/home/chen/ComfyUI/.venv/lib/python3.14/site-packages/torch/nn/modules/module.py", line 1789, in _call_impl
return forward_call(*args, **kwargs)
File "/home/chen/ComfyUI/comfy/ldm/krea2/model.py", line 82, in forward
q, k = apply_rope(q, k, freqs)
~~~~~~~~~~^^^^^^^^^^^^^
File "/home/chen/ComfyUI/comfy/ldm/flux/math.py", line 51, in apply_rope
return comfy.quant_ops.ck.apply_rope(xq, xk, freqs_cis)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^
File "/home/chen/ComfyUI/.venv/lib/python3.14/site-packages/comfy_kitchen/__init__.py", line 434, in apply_rope
return torch.ops.comfy_kitchen.apply_rope(xq, xk, freqs_cis)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^
File "/home/chen/ComfyUI/.venv/lib/python3.14/site-packages/torch/_ops.py", line 1275, in __call__
return self._op(*args, **kwargs)
~~~~~~~~^^^^^^^^^^^^^^^^^
File "/home/chen/ComfyUI/.venv/lib/python3.14/site-packages/torch/_library/custom_ops.py", line 375, in backend_impl
result = self._backend_fns[device_type](*args, **kwargs)
File "/home/chen/ComfyUI/.venv/lib/python3.14/site-packages/torch/_compile.py", line 54, in inner
return disable_fn(*args, **kwargs)
File "/home/chen/ComfyUI/.venv/lib/python3.14/site-packages/torch/_dynamo/eval_frame.py", line 1298, in _fn
return fn(*args, **kwargs)
File "/home/chen/ComfyUI/.venv/lib/python3.14/site-packages/torch/_library/custom_ops.py", line 410, in wrapped_fn
return fn(*args, **kwargs)
File "/home/chen/ComfyUI/.venv/lib/python3.14/site-packages/comfy_kitchen/backends/eager/rope.py", line 121, in _op_apply_rope
return impl(**kwargs)
File "/home/chen/ComfyUI/.venv/lib/python3.14/site-packages/comfy_kitchen/backends/triton/rope.py", line 230, in apply_rope
return _apply_rope(xq, freqs_cis, xk)
File "/home/chen/ComfyUI/.venv/lib/python3.14/site-packages/comfy_kitchen/backends/triton/rope.py", line 191, in _apply_rope
apply_rope_kernel[grid](
~~~~~~~~~~~~~~~~~~~~~~~^
x1,
^^^
...<23 lines>...
split_half=split_half,
^^^^^^^^^^^^^^^^^^^^^^
)
^
File "/home/chen/ComfyUI/.venv/lib/python3.14/site-packages/triton/runtime/jit.py", line 370, in
return lambda *args, **kwargs: self.run(grid=grid, warmup=False, *args, **kwargs)
~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/chen/ComfyUI/.venv/lib/python3.14/site-packages/triton/runtime/jit.py", line 761, in run
kernel.run(grid_0, grid_1, grid_2, stream, kernel.function, kernel.packed_metadata, launch_metadata,
~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
knobs.runtime.launch_enter_hook, knobs.runtime.launch_exit_hook, *bound_args.values())
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/chen/ComfyUI/.venv/lib/python3.14/site-packages/triton/backends/amd/driver.py", line 428, in __call__
self.launch(self.launch_cooperative_grid, gridX, gridY, gridZ, stream, function, profile_scratch,
~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
kernel_metadata, launch_metadata, launch_enter_hook, launch_exit_hook, self.warp_size,
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
self.arg_annotations, self.kernel_signature, args)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/chen/ComfyUI/.venv/lib/python3.14/site-packages/triton/knobs.py", line 423, in __call__
for call in self.calls if not self.reversed else reversed(self.calls):
~~~~~~~~^^^^^^^^^^^^
SystemError: returned a result with an exception set

[INFO] Prompt executed in 58.20 seconds
/home/chen/ComfyUI/.venv/lib/python3.14/site-packages/torch/cuda/__init__.py:1168: UserWarning: Ignoring invalid value for boolean flag AMD_SERIALIZE_KERNEL: 3valid values are 0 or 1. (Triggered internally at /__w/rockrel/rockrel/external-builds/pytorch/pytorch/c10/util/env.cpp:89.)
return torch._C._cuda_getDevice()
Exception in thread Thread-2 (prompt_worker):
Traceback (most recent call last):
File "/usr/lib/python3.14/threading.py", line 1082, in _bootstrap_inner
self._context.run(self.run)
~~~~~~~~~~~~~~~~~^^^^^^^^^^
File "/usr/lib/python3.14/threading.py", line 1024, in run
self._target(*self._args, **self._kwargs)
~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/chen/ComfyUI/main.py", line 404, in prompt_worker
comfy.model_management.soft_empty_cache()
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^
File "/home/chen/ComfyUI/comfy/model_management.py", line 1975, in soft_empty_cache
torch.cuda.synchronize()
~~~~~~~~~~~~~~~~~~~~~~^^
File "/home/chen/ComfyUI/.venv/lib/python3.14/site-packages/torch/cuda/__init__.py", line 1180, in synchronize
with torch.cuda.device(device):
~~~~~~~~~~~~~~~~~^^^^^^^^
File "/home/chen/ComfyUI/.venv/lib/python3.14/site-packages/torch/cuda/__init__.py", line 613, in __init__
self.idx = _get_device_index(device, optional=True)
~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^
File "/home/chen/ComfyUI/.venv/lib/python3.14/site-packages/torch/cuda/_utils.py", line 590, in _get_device_index
return _torch_get_device_index(device, optional, allow_cpu)
File "/home/chen/ComfyUI/.venv/lib/python3.14/site-packages/torch/_utils.py", line 888, in _get_device_index
device_idx = _get_current_device_index()
File "/home/chen/ComfyUI/.venv/lib/python3.14/site-packages/torch/_utils.py", line 825, in _get_current_device_index
return _get_device_attr(lambda m: m.current_device())
File "/home/chen/ComfyUI/.venv/lib/python3.14/site-packages/torch/_utils.py", line 810, in _get_device_attr
return get_member(torch.cuda)
File "/home/chen/ComfyUI/.venv/lib/python3.14/site-packages/torch/_utils.py", line 825, in
return _get_device_attr(lambda m: m.current_device())
~~~~~~~~~~~~~~~~^^
File "/home/chen/ComfyUI/.venv/lib/python3.14/site-packages/torch/cuda/__init__.py", line 1168, in current_device
return torch._C._cuda_getDevice()
~~~~~~~~~~~~~~~~~~~~~~~~^^
torch.AcceleratorError: CUDA error: unspecified launch failure
Search for `hipErrorLaunchFailure' in https://rocm.docs.amd.com/projects/HIP/en/latest/index.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 AMD_SERIALIZE_KERNEL=3
Device-side assertion tracking was not enabled by user.
```

### Other

pip list
```
(.venv) chen@amd395 ~/ComfyUI ((v0.28.0)) [0|SIGINT]> pip list
Package Version
------------------------------------------ -------------------------------------
aiohappyeyeballs 2.6.1
aiohttp 3.13.5
aiosignal 1.4.0
albucore 0.0.24
albumentations 2.0.8
alembic 1.18.4
amd-torch-device-gfx1151 2.12.0+rocm7.15.0a20260721
amd-torch-device-gfx115x 2.12.0+rocm7.15.0a20260721
amd-torchvision-device-gfx1151 0.27.0+rocm7.15.0a20260721
annotated-doc 0.0.4
annotated-types 0.7.0
anyio 4.13.0
attrs 26.1.0
av 17.0.1
beautifulsoup4 4.15.0
blake3 1.0.8
certifi 2026.4.22
cffi 2.0.0
chardet 7.4.3
charset-normalizer 3.4.7
click 8.3.3
cmake 4.4.0
comfy-aimdo 0.4.10
comfy-angle 0.1.0
comfy-kitchen 0.2.20
comfyui-embedded-docs 0.5.8
comfyui_frontend_package 1.45.21
comfyui-manager 4.2.1
comfyui_workflow_templates 0.11.9
comfyui-workflow-templates-core 0.3.272
comfyui-workflow-templates-json 0.1.6
comfyui-workflow-templates-media-api 0.3.84
comfyui-workflow-templates-media-assets-01 0.1.3
comfyui-workflow-templates-media-image 0.3.160
comfyui-workflow-templates-media-other 0.3.229
comfyui-workflow-templates-media-video 0.3.101
contourpy 1.3.3
cryptography 47.0.0
cstr 0.1.0
cycler 0.12.1
diffusers 0.27.2
easydict 1.13
einops 0.8.2
fairscale 0.4.13
ffmpy 0.3.0
filelock 3.29.0
fonttools 4.63.0
frozenlist 1.8.0
fsspec 2026.3.0
gdown 6.1.0
gitdb 4.0.12
GitPython 3.1.47
glfw 2.10.0
greenlet 3.4.0
h11 0.16.0
hf-xet 1.4.3
httpcore 1.0.9
httpx 0.28.1
huggingface_hub 1.12.0
idna 3.13
ImageIO 2.37.4
img2texture 1.0.6
importlib_metadata 9.0.0
Jinja2 3.1.6
joblib 1.5.3
jsonschema 4.26.0
jsonschema-specifications 2025.9.1
kiwisolver 1.5.0
kornia 0.8.2
kornia_rs 0.1.10
lark 1.3.1
lazy-loader 0.5
llvmlite 0.48.0
Mako 1.3.11
markdown-it-py 4.0.0
MarkupSafe 3.0.3
matplotlib 3.11.1
mdurl 0.1.2
mpmath 1.3.0
multidict 6.7.1
narwhals 2.24.0
networkx 3.6.1
numba 0.66.0
numpy 2.4.3
opencv-contrib-python 5.0.0.93
opencv-python 5.0.0.93
opencv-python-headless 5.0.0.93
packaging 26.2
pilgram 2.0.0
pillow 12.2.0
pip 25.1.1
platformdirs 4.10.1
pooch 1.9.0
propcache 0.4.1
psutil 7.2.2
pycparser 3.0
pydantic 2.13.3
pydantic_core 2.46.3
pydantic-settings 2.14.0
PyGithub 2.9.1
Pygments 2.20.0
PyJWT 2.12.1
PyMatting 1.1.15
PyNaCl 1.6.2
PyOpenGL 3.1.10
pyparsing 3.3.2
PySocks 1.7.1
python-dateutil 2.9.0.post0
python-dotenv 1.2.2
PyYAML 6.0.3
referencing 0.37.0
regex 2026.4.4
rembg 2.0.77
requests 2.33.1
rich 15.0.0
rocm 7.15.0a20260721
rocm-bootstrap 0.1.0
rocm-sdk-core 7.15.0a20260721
rocm-sdk-device-gfx1151 7.15.0a20260721
rocm-sdk-libraries 7.15.0a20260721
rpds-py 2026.6.3
safetensors 0.7.0
scikit-image 0.26.0
scikit-learn 1.9.0
scipy 1.17.1
sentencepiece 0.2.1
setuptools 80.9.0
shellingham 1.5.4
simpleeval 1.0.7
simsimd 6.5.16
six 1.17.0
smmap 5.0.3
soupsieve 2.9
spandrel 0.4.2
SQLAlchemy 2.0.49
stringzilla 4.6.2
sympy 1.14.0
threadpoolctl 3.6.0
tifffile 2026.7.14
timm 1.0.28
tokenizers 0.22.2
toml 0.10.2
torch 2.12.0+rocm7.15.0a20260721
torchaudio 2.11.0+rocm7.15.0a20260721
torchsde 0.2.6
torchvision 0.27.0+rocm7.15.0a20260721
tqdm 4.67.3
trampoline 0.1.2
transformers 5.6.2
transparent-background 1.3.4
triton 3.7.1+git0263a6a6.rocm7.15.0a20260721
typer 0.24.2
typing_extensions 4.15.0
typing-inspection 0.4.2
urllib3 2.6.3
uv 0.11.7
wget 3.2
yarl 1.23.0
zipp 3.23.1
```

Reproduce the Krea2 template with the documented Python installation and main.py command, comparing runs with and without --enable-triton-backend. Start at nodes.py and comfy/samplers.py in the traceback, then follow the reported apply_rope_kernel memory fault. Done means the workflow runs with the Triton backend enabled without the HSA memory fault or launch failure.

Written by the indexing model from the issue text.