[ET-VK] Add ANY_STORAGE support to softmax and log_softmax

[SS-JIA]

Stack from ghstack (oldest at bottom):

• [ET-VK] Buckify parakeet #18062
• [ET-VK][qlinear] Add bias support to q4gsw and dq8ca_q4gsw quantized linear ops #18061
• [ET-VK] Add conv1d shaders and dispatch for 1D convolution #18060
• [ET-VK] Implement missing Vulkan operators for Parakeet TDT model #18059
• [ET-VK] Add ANY_STORAGE support to argmax and argmin #18058
• [ET-VK] Add ANY_STORAGE support to embedding #18057
• [ET-VK] Add ANY_STORAGE support to repeat #18056
• [ET-VK] Add ANY_STORAGE support to native_layer_norm #18055
• -> [ET-VK] Add ANY_STORAGE support to softmax and log_softmax #18054
• [ET-VK] Add ANY_STORAGE support to expand_copy #18053
• [ET-VK] Add ANY_STORAGE support to arange and full #18052
• [ET-VK] Generalize constant_pad_nd to support any storage type and packed dimension #18051
• [ET-VK] Add symint infrastructure to VulkanBackend and ComputeGraph #18050
• [ET-VK] Add Vulkan backend support for Parakeet runner and export #18049
• [ET-VK][ez] Fix IndexError in Vulkan partitioner DtypeSetList/TensorRepSetList #18048

Add a buffer shader variant for softmax using workgroup-based reduction with
shared memory. The buffer shader dispatches one workgroup per "row" along the
reduction dimension, with NWORKERS=4 threads per workgroup cooperating on the
max and sum reductions before writing outputs.

In Softmax.cpp, the existing texture dispatch is renamed add_softmax_texture_node
and new add_softmax_buffer_node / add_softmax_node dispatch functions are added.
op_registry.py is updated from ANY_TEXTURE to ANY_STORAGE so the partitioner
accepts buffer-storage inputs. Test cases now also exercise buffer storage.

Differential Revision: D95970171

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/18054

• 📄 Preview Python docs built from this PR

Note: Links to docs will display an error until the docs builds have been completed.

❌ 2 New Failures, 2 Cancelled Jobs

As of commit 9ca763e with merge base f09bd55 ():

NEW FAILURES - The following jobs have failed:

• Test CUDA Builds / test-models-cuda (conv1d) / linux-job (gh)
  RuntimeError: Command docker exec -t dcc39c96650fa652f92c530106834038eaa90eb86204bec912764d156361c8f6 /exec failed with exit code 1
• Test Metal Backend / export-model-metal-artifact (mistralai, Voxtral-Mini-4B-Realtime-2602, quantized-int4-metal) / macos-job (gh)
  RuntimeError: Command bash /Users/ec2-user/runner/_work/_temp/exec_script failed with exit code 1

CANCELLED JOBS - The following jobs were cancelled. Please retry:

• Check Labels / Check labels (gh)
• pull / unittest / macos / macos-job (gh)
  ##[error]The operation was canceled.

This comment was automatically generated by Dr. CI and updates every 15 minutes.

[github-actions]

This PR needs a release notes: label

If your change should be included in the release notes (i.e. would users of this library care about this change?), please use a label starting with release notes:. This helps us keep track and include your important work in the next release notes.

To add a label, you can comment to pytorchbot, for example
@pytorchbot label "release notes: none"

For more information, see
https://github.com/pytorch/pytorch/wiki/PyTorch-AutoLabel-Bot#why-categorize-for-release-notes-and-how-does-it-work.