New failures in test_scaled_matmul_cuda_xpu.py

### 🐛 Describe the bug

Cases:
op_ut,third_party.torch-xpu-ops.test.xpu.test_scaled_matmul_cuda_xpu.TestFP8MatmulXPU,test_scaled_addmm_contraction_dim_fake_False_inplace_False_contraction_dim0_supported_True_xpu
op_ut,third_party.torch-xpu-ops.test.xpu.test_scaled_matmul_cuda_xpu.TestFP8MatmulXPU,test_scaled_addmm_contraction_dim_fake_False_inplace_False_contraction_dim1_supported_True_xpu
op_ut,third_party.torch-xpu-ops.test.xpu.test_scaled_matmul_cuda_xpu.TestFP8MatmulXPU,test_scaled_addmm_contraction_dim_fake_False_inplace_True_contraction_dim0_supported_True_xpu
op_ut,third_party.torch-xpu-ops.test.xpu.test_scaled_matmul_cuda_xpu.TestFP8MatmulXPU,test_scaled_addmm_contraction_dim_fake_False_inplace_True_contraction_dim1_supported_True_xpu
op_ut,third_party.torch-xpu-ops.test.xpu.test_scaled_matmul_cuda_xpu.TestFP8MatmulXPU,test_scaled_addmm_cudagraph_xpu
op_ut,third_party.torch-xpu-ops.test.xpu.test_scaled_matmul_cuda_xpu.TestFP8MatmulXPU,test_scaled_addmm_empty_inplace_False_m_0_n_32_k_16_xpu
op_ut,third_party.torch-xpu-ops.test.xpu.test_scaled_matmul_cuda_xpu.TestFP8MatmulXPU,test_scaled_addmm_empty_inplace_False_m_32_n_0_k_16_xpu
op_ut,third_party.torch-xpu-ops.test.xpu.test_scaled_matmul_cuda_xpu.TestFP8MatmulXPU,test_scaled_addmm_empty_inplace_False_m_32_n_32_k_0_xpu
op_ut,third_party.torch-xpu-ops.test.xpu.test_scaled_matmul_cuda_xpu.TestFP8MatmulXPU,test_scaled_addmm_empty_inplace_True_m_0_n_32_k_16_xpu
op_ut,third_party.torch-xpu-ops.test.xpu.test_scaled_matmul_cuda_xpu.TestFP8MatmulXPU,test_scaled_addmm_empty_inplace_True_m_32_n_0_k_16_xpu
op_ut,third_party.torch-xpu-ops.test.xpu.test_scaled_matmul_cuda_xpu.TestFP8MatmulXPU,test_scaled_addmm_empty_inplace_True_m_32_n_32_k_0_xpu
op_ut,third_party.torch-xpu-ops.test.xpu.test_scaled_matmul_cuda_xpu.TestFP8MatmulXPU,test_scaled_addmm_fullgraph_xpu
op_ut,third_party.torch-xpu-ops.test.xpu.test_scaled_matmul_cuda_xpu.TestFP8MatmulXPU,test_scaled_addmm_mxfp8_bfloat16_xpu
op_ut,third_party.torch-xpu-ops.test.xpu.test_scaled_matmul_cuda_xpu.TestFP8MatmulXPU,test_scaled_addmm_mxfp8_float16_xpu
op_ut,third_party.torch-xpu-ops.test.xpu.test_scaled_matmul_cuda_xpu.TestFP8MatmulXPU,test_scaled_addmm_mxfp8_float32_xpu
op_ut,third_party.torch-xpu-ops.test.xpu.test_scaled_matmul_cuda_xpu.TestFP8MatmulXPU,test_scaled_addmm_nvfp4_two_level_False_bfloat16_xpu
op_ut,third_party.torch-xpu-ops.test.xpu.test_scaled_matmul_cuda_xpu.TestFP8MatmulXPU,test_scaled_addmm_nvfp4_two_level_False_float16_xpu
op_ut,third_party.torch-xpu-ops.test.xpu.test_scaled_matmul_cuda_xpu.TestFP8MatmulXPU,test_scaled_addmm_nvfp4_two_level_False_float32_xpu
op_ut,third_party.torch-xpu-ops.test.xpu.test_scaled_matmul_cuda_xpu.TestFP8MatmulXPU,test_scaled_addmm_nvfp4_two_level_True_bfloat16_xpu
op_ut,third_party.torch-xpu-ops.test.xpu.test_scaled_matmul_cuda_xpu.TestFP8MatmulXPU,test_scaled_addmm_nvfp4_two_level_True_float16_xpu
op_ut,third_party.torch-xpu-ops.test.xpu.test_scaled_matmul_cuda_xpu.TestFP8MatmulXPU,test_scaled_addmm_nvfp4_two_level_True_float32_xpu
op_ut,third_party.torch-xpu-ops.test.xpu.test_scaled_matmul_cuda_xpu.TestFP8MatmulXPU,test_scaled_addmm_scalar_semantics_inplace_False_xpu
op_ut,third_party.torch-xpu-ops.test.xpu.test_scaled_matmul_cuda_xpu.TestFP8MatmulXPU,test_scaled_addmm_scalar_semantics_inplace_True_xpu
op_ut,third_party.torch-xpu-ops.test.xpu.test_scaled_matmul_cuda_xpu.TestFP8MatmulXPU,test_scaled_addmm_tensorwise_bfloat16_m_1_xpu
op_ut,third_party.torch-xpu-ops.test.xpu.test_scaled_matmul_cuda_xpu.TestFP8MatmulXPU,test_scaled_addmm_tensorwise_bfloat16_m_64_xpu
op_ut,third_party.torch-xpu-ops.test.xpu.test_scaled_matmul_cuda_xpu.TestFP8MatmulXPU,test_scaled_addmm_tensorwise_float16_m_1_xpu
op_ut,third_party.torch-xpu-ops.test.xpu.test_scaled_matmul_cuda_xpu.TestFP8MatmulXPU,test_scaled_addmm_tensorwise_float16_m_64_xpu
op_ut,third_party.torch-xpu-ops.test.xpu.test_scaled_matmul_cuda_xpu.TestFP8MatmulXPU,test_scaled_addmm_tensorwise_float32_m_1_xpu
op_ut,third_party.torch-xpu-ops.test.xpu.test_scaled_matmul_cuda_xpu.TestFP8MatmulXPU,test_scaled_addmm_tensorwise_float32_m_64_xpu
op_ut,third_party.torch-xpu-ops.test.xpu.test_scaled_matmul_cuda_xpu.TestFP8MatmulXPU,test_scaled_addmm_validation_xpu
op_ut,third_party.torch-xpu-ops.test.xpu.test_scaled_matmul_cuda_xpu.TestFP8MatmulXPU,test_scaled_addmm_wgrad_accumulation_xpu

## ErrorLog
### NotImplementedError: The operator 'aten::_scaled_addmm.out' is not currently implemented for the XPU device.
```bash
_______________ TestFP8MatmulXPU.test_scaled_addmm_fullgraph_xpu _______________
[gw6] linux -- Python 3.10.21 /__w/torch-xpu-ops/torch-xpu-ops/.venv/bin/python
Traceback (most recent call last):
File "/__w/torch-xpu-ops/torch-xpu-ops/pytorch/third_party/torch-xpu-ops/test/xpu/../../../../test/test_scaled_matmul_cuda.py", line 1586, in test_scaled_addmm_fullgraph
expected = fn(input, *operands)
File "/__w/torch-xpu-ops/torch-xpu-ops/pytorch/third_party/torch-xpu-ops/test/xpu/../../../../test/test_scaled_matmul_cuda.py", line 1581, in fn
return scaled_addmm(
File "/__w/torch-xpu-ops/torch-xpu-ops/.venv/lib/python3.10/site-packages/torch/nn/functional.py", line 7325, in scaled_addmm
return torch._scaled_addmm(
NotImplementedError: The operator 'aten::_scaled_addmm.out' is not currently implemented for the XPU device. Please open a feature on https://github.com/intel/torch-xpu-ops/issues. You can set the environment variable `PYTORCH_ENABLE_XPU_FALLBACK=1` to use the CPU implementation as a fallback for XPU unimplemented operators. WARNING: this will bring unexpected performance compared with running natively on XPU.
```
### ValueError: For XPU MX/NVFP4 gemm, scale_a swizzle entries must all be NO_SWIZZLE
```bash
____________ TestFP8MatmulXPU.test_scaled_addmm_mxfp8_bfloat16_xpu _____________
[gw6] linux -- Python 3.10.21 /__w/torch-xpu-ops/torch-xpu-ops/.venv/bin/python
Traceback (most recent call last):
File "/__w/torch-xpu-ops/torch-xpu-ops/pytorch/third_party/torch-xpu-ops/test/xpu/../../../../test/test_scaled_matmul_cuda.py", line 1480, in test_scaled_addmm_mxfp8
actual = scaled_addmm(input, *args, **kwargs)
File "/__w/torch-xpu-ops/torch-xpu-ops/.venv/lib/python3.10/site-packages/torch/nn/functional.py", line 7325, in scaled_addmm
return torch._scaled_addmm(
ValueError: For XPU MX/NVFP4 gemm, scale_a swizzle entries must all be NO_SWIZZLE
```
## Pytorch Version
[8374e0f](https://github.com/pytorch/pytorch/commit/8374e0fbb472a88cb5be7fcf0f14d5b3f8ec08ac)
## Torch-xpu-ops
c513b6ad85ffe36858bfbf39fb1103d19cecd8e2

### Versions

Detail
Collecting environment information...
PyTorch version: 2.15.0a0+git8374e0f
Is debug build: False
CUDA used to build PyTorch: None
ROCm SDK used to build PyTorch: N/A
HIP used to build PyTorch: N/A

OS: Ubuntu 26.04 LTS (x86_64)
GCC version: (Ubuntu 13.4.0-10ubuntu1) 13.4.0
Clang version: Could not collect
CMake version: version 3.31.6
Libc version: glibc-2.43

Python version: 3.10.21 (main, Sep 1 2026, 14:16:49) [Clang 22.1.3 ] (64-bit runtime)
Python platform: Linux-7.0.0-14-generic-x86_64-with-glibc2.43
Is CUDA available: False
CUDA runtime version: No CUDA
CUDA_MODULE_LOADING set to: N/A
GPU models and configuration: No CUDA
Nvidia driver version: No CUDA
cuDNN version: No CUDA
Is XPU available: True
XPU used to build PyTorch: 20260100
Intel GPU driver version:
* intel-opencl-icd: 26.18.38308.1-1~26.04~ppa1
* libze1: 1.28.2-2
Intel GPU models onboard:
* Intel(R) Arc(TM) Pro B60 Graphics
* Intel(R) Arc(TM) Pro B60 Graphics
* Intel(R) Arc(TM) Pro B60 Graphics
* Intel(R) Arc(TM) Pro B60 Graphics
* Intel(R) Arc(TM) Pro B60 Graphics
* Intel(R) Arc(TM) Pro B60 Graphics
* Intel(R) Arc(TM) Pro B60 Graphics
* Intel(R) Arc(TM) Pro B60 Graphics
Intel GPU models detected:
* [0] _XpuDeviceProperties(name='Intel(R) Arc(TM) Pro B60 Graphics', platform_name='Intel(R) oneAPI Unified Runtime over Level-Zero V2', type='gpu', device_id=0xE211, uuid=868011e2-0000-0000-1700-000000000000, driver_version='1.15.38308+1', total_memory=24480MB, local_mem_size=128KB, last_level_cache_size=18432KB, max_compute_units=160, memory_clock_rate=2400MHz, memory_bus_width=64-bit, gpu_eu_count=160, gpu_subslice_count=20, max_work_group_size=1024, max_num_sub_groups=64, sub_group_sizes=[16 32], has_fp16=1, has_fp64=1, has_atomic64=1, is_integrated_gpu=0)
* [1] _XpuDeviceProperties(name='Intel(R) Arc(TM) Pro B60 Graphics', platform_name='Intel(R) oneAPI Unified Runtime over Level-Zero V2', type='gpu', device_id=0xE211, uuid=868011e2-0000-0000-2c00-000000000000, driver_version='1.15.38308+1', total_memory=24480MB, local_mem_size=128KB, last_level_cache_size=18432KB, max_compute_units=160, memory_clock_rate=2400MHz, memory_bus_width=64-bit, gpu_eu_count=160, gpu_subslice_count=20, max_work_group_size=1024, max_num_sub_groups=64, sub_group_sizes=[16 32], has_fp16=1, has_fp64=1, has_atomic64=1, is_integrated_gpu=0)
* [2] _XpuDeviceProperties(name='Intel(R) Arc(TM) Pro B60 Graphics', platform_name='Intel(R) oneAPI Unified Runtime over Level-Zero V2', type='gpu', device_id=0xE211, uuid=868011e2-0000-0000-3d00-000000000000, driver_version='1.15.38308+1', total_memory=24480MB, local_mem_size=128KB, last_level_cache_size=18432KB, max_compute_units=160, memory_clock_rate=2400MHz, memory_bus_width=64-bit, gpu_eu_count=160, gpu_subslice_count=20, max_work_group_size=1024, max_num_sub_groups=64, sub_group_sizes=[16 32], has_fp16=1, has_fp64=1, has_atomic64=1, is_integrated_gpu=0)
* [3] _XpuDeviceProperties(name='Intel(R) Arc(TM) Pro B60 Graphics', platform_name='Intel(R) oneAPI Unified Runtime over Level-Zero V2', type='gpu', device_id=0xE211, uuid=868011e2-0000-0000-4e00-000000000000, driver_version='1.15.38308+1', total_memory=24480MB, local_mem_size=128KB, last_level_cache_size=18432KB, max_compute_units=160, memory_clock_rate=2400MHz, memory_bus_width=64-bit, gpu_eu_count=160, gpu_subslice_count=20, max_work_group_size=1024, max_num_sub_groups=64, sub_group_sizes=[16 32], has_fp16=1, has_fp64=1, has_atomic64=1, is_integrated_gpu=0)
* [4] _XpuDeviceProperties(name='Intel(R) Arc(TM) Pro B60 Graphics', platform_name='Intel(R) oneAPI Unified Runtime over Level-Zero V2', type='gpu', device_id=0xE211, uuid=868011e2-0000-0000-9700-000000000000, driver_version='1.15.38308+1', total_memory=24480MB, local_mem_size=128KB, last_level_cache_size=18432KB, max_compute_units=160, memory_clock_rate=2400MHz, memory_bus_width=64-bit, gpu_eu_count=160, gpu_subslice_count=20, max_work_group_size=1024, max_num_sub_groups=64, sub_group_sizes=[16 32], has_fp16=1, has_fp64=1, has_atomic64=1, is_integrated_gpu=0)
* [5] _XpuDeviceProperties(name='Intel(R) Arc(TM) Pro B60 Graphics', platform_name='Intel(R) oneAPI Unified Runtime over Level-Zero V2', type='gpu', device_id=0xE211, uuid=868011e2-0000-0000-a900-000000000000, driver_version='1.15.38308+1', total_memory=24480MB, local_mem_size=128KB, last_level_cache_size=18432KB, max_compute_units=160, memory_clock_rate=2400MHz, memory_bus_width=64-bit, gpu_eu_count=160, gpu_subslice_count=20, max_work_group_size=1024, max_num_sub_groups=64, sub_group_sizes=[16 32], has_fp16=1, has_fp64=1, has_atomic64=1, is_integrated_gpu=0)
* [6] _XpuDeviceProperties(name='Intel(R) Arc(TM) Pro B60 Graphics', platform_name='Intel(R) oneAPI Unified Runtime over Level-Zero V2', type='gpu', device_id=0xE211, uuid=868011e2-0000-0000-ba00-000000000000, driver_version='1.15.38308+1', total_memory=24480MB, local_mem_size=128KB, last_level_cache_size=18432KB, max_compute_units=160, memory_clock_rate=2400MHz, memory_bus_width=64-bit, gpu_eu_count=160, gpu_subslice_count=20, max_work_group_size=1024, max_num_sub_groups=64, sub_group_sizes=[16 32], has_fp16=1, has_fp64=1, has_atomic64=1, is_integrated_gpu=0)
* [7] _XpuDeviceProperties(name='Intel(R) Arc(TM) Pro B60 Graphics', platform_name='Intel(R) oneAPI Unified Runtime over Level-Zero V2', type='gpu', device_id=0xE211, uuid=868011e2-0000-0000-cb00-000000000000, driver_version='1.15.38308+1', total_memory=24480MB, local_mem_size=128KB, last_level_cache_size=18432KB, max_compute_units=160, memory_clock_rate=2400MHz, memory_bus_width=64-bit, gpu_eu_count=160, gpu_subslice_count=20, max_work_group_size=1024, max_num_sub_groups=64, sub_group_sizes=[16 32], has_fp16=1, has_fp64=1, has_atomic64=1, is_integrated_gpu=0)
HIP runtime version: N/A
MIOpen runtime version: N/A
Is XNNPACK available: False
Caching allocator config: N/A

Vulnerability Vmscape: Mitigation; IBPB before exit to userspace

Versions of relevant libraries:
[pip3] dpcpp-cpp-rt==2026.1.0
[pip3] impi-rt==2021.18.1
[pip3] intel-cmplr-lib-rt==2026.1.0
[pip3] intel-cmplr-lib-ur==2026.1.0
[pip3] intel-cmplr-lic-rt==2026.1.0
[pip3] intel-opencl-rt==2026.1.0
[pip3] intel-openmp==2026.1.0
[pip3] intel-pti==1.0.1
[pip3] intel-sycl-rt==2026.1.0
[pip3] mkl==2026.1.0
[pip3] mypy==1.16.0
[pip3] mypy_extensions==1.1.0
[pip3] numpy==1.24.4
[pip3] nvidia-cuda-cupti==13.4.58
[pip3] oneccl==2022.1.1
[pip3] oneccl-devel==2022.1.1
[pip3] onemkl-license==2026.1.0
[pip3] onemkl-sycl-blas==2026.1.0
[pip3] onemkl-sycl-dft==2026.1.0
[pip3] onemkl-sycl-lapack==2026.1.0
[pip3] onemkl-sycl-rng==2026.1.0
[pip3] onemkl-sycl-sparse==2026.1.0
[pip3] onnx==1.21.0
[pip3] onnx-ir==0.1.16
[pip3] onnxscript==0.6.2
[pip3] optree==0.13.0
[pip3] tbb==2023.1.0
[pip3] tcmlib==1.5.0
[pip3] torch==2.15.0a0+git8374e0f
[pip3] torchao==0.19.0.dev20260911+xpu
[pip3] torchaudio==2.11.0a0+b85c99c
[pip3] torchvision==0.30.0a0+ac8d215
[pip3] triton-xpu==3.8.0+git1e2d42a0
[pip3] umf==1.1.0
[conda] No relevant packages

Start by running the listed TestFP8MatmulXPU failures and read test_scaled_matmul_cuda.py around lines 1480 and 1581-1586. Trace the XPU handling for aten::_scaled_addmm.out and the MX/NVFP4 scale validation. Done means the listed tests pass natively on XPU without enabling CPU fallback.

Written by the indexing model from the issue text.