Revert "Add subgroup topk kernel for XPU (part1 of #3369) (#3371)"

[CuiYifeng]

This reverts commit 8eaa591, except new overloads of sycl_kernel_submit.
The motivation is that #3371 caused build timeout in stock CD.

[Copilot]

Pull request overview

This PR reverts the subgroup top-k XPU optimization that caused build timeouts with multiple AOT targets, restoring the existing segmented top-k path while removing the dedicated subgroup kernel files.

Review context: read .github/skills/xpu-ops-pr-review/SKILL.md.

Changes:

• Removes the subgroup top-k kernel implementation and declaration.
• Removes the optimized-path dispatch from topk_kernel.
• Restores unconditional use of segmented_group_select_pairs followed by optional sorting.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated no comments.

File	Description
src/ATen/native/xpu/sycl/TensorTopKSbtopkKernel.h	Deletes the subgroup top-k API declarations.
src/ATen/native/xpu/sycl/TensorTopKSbtopkKernel.cpp	Deletes the subgroup top-k kernel implementation and launch logic.
src/ATen/native/xpu/sycl/TensorTopKKernel.cpp	Restores the original top-k execution path without subgroup dispatch.

[jianyizh]

Please write clear in description that it's reverted because of long build time in CD

[CuiYifeng]

Please write clear in description that it's reverted because of long build time in CD

Updated.

[guangyey]

I saw the build passed and landed it immediately since the change only removed code. We need to update commit pin as soon as possible to unblock pytorch/pytorch#173895 PyTorch CI.

[chuanqi129]

@copilot New UT failures detected in op_ut tests (7 total). Please check these new failures and analyze whether they are caused by the PR changes.

Important: Do NOT update the plan in the PR description directly. Use reply comments to update the status.

Job log: https://github.com/intel/torch-xpu-ops/actions/runs/26078819694
Full new failure report: https://github.com/intel/torch-xpu-ops/actions/runs/26078819694/artifacts/7081996199

New failures:

Category	Class name	Test name	Status	Message
op_ut	third_party.torch-xpu-ops.test.xpu.dynamo.test_regional_inductor_xpu.RegionalInductorInvokeSubgraphTests	test_max_autotune_no_cudagraphs_serialize_False	failed	AttributeError: module 'torch._inductor' has no attribute 'compile_fx'
op_ut	third_party.torch-xpu-ops.test.xpu.functorch.test_control_flow_xpu.TestControlFlowTraced	test_while_loop_autograd_simple	failed	AssertionError: 'clas[152 chars] zeros: "i64[]" = torch.ops.aten.zeros.def[3514 chars]1)\n' != 'clas[152 chars] clone: "f32[3, 3]" = torch.ops.aten.clone[3636 chars]1)\n'
op_ut		third_party.torch-xpu-ops.test.xpu.functorch.test_ops_xpu	error	ImportError: cannot import name 'skipOps' from 'common_utils' (/__w/torch-xpu-ops/torch-xpu-ops/pytorch/third_party/torch-xpu-ops/test/xpu/../../../../test/functorch/common_utils.py)
op_ut		third_party.torch-xpu-ops.test.xpu.functorch.test_vmap_xpu	error	ImportError: cannot import name 'skipOps' from 'common_utils' (/__w/torch-xpu-ops/torch-xpu-ops/pytorch/third_party/torch-xpu-ops/test/xpu/../../../../test/functorch/common_utils.py)
op_ut	third_party.torch-xpu-ops.test.xpu.test_custom_ops_xpu.TestCustomOpAPI	test_no_grad_skips_autograd	failed	AssertionError: Scalars are not equal!

... and 2 more failures. See full list in artifact.

Instructions:

1. Download the full new failure report artifact for complete error messages
2. Analyze whether the failures are caused by PR changes
3. If caused by PR changes, fix those failures on this PR branch directly
4. Add failing test cases with pytest commands in the PR description for the reproducer job to use
5. Use reply comments to report analysis status - do NOT modify the PR description plan directly

[github-actions]

Performance outliers, please check!

• 🔴 [-1, 80%), should be regression

Category	Model	Target vs. Baseline [Eager]	Target vs. Baseline [Inductor]
timm_models_bfloat16_training	ghostnet_100	0.856848	0.564996
timm_models_bfloat16_training	mobilenetv3_large_100	0.783607	0.569622
timm_models_bfloat16_training	mobilevit_s	0.718829	0.579479
timm_models_bfloat16_training	mobilenetv2_100	0.835302	0.584723
timm_models_bfloat16_training	tf_efficientnet_b0	0.804629	0.619609
timm_models_bfloat16_training	vit_base_patch16_siglip_256	0.695374	0.744301
timm_models_bfloat16_training	deit_base_distilled_patch16_224	0.752327	0.772345
timm_models_bfloat16_training	dm_nfnet_f0	0.639916	0.782825
timm_models_bfloat16_training	nfnet_l0	0.738771	0.783153
torchbench_bfloat16_training	mobilenet_v2	1.063866	0.791071
timm_models_bfloat16_training	adv_inception_v3	0.791371	0.796311
timm_models_bfloat16_training	beit_base_patch16_224	0.735183	0.813880
timm_models_bfloat16_training	visformer_small	0.719781	0.822906

• 🟡 [80%, 90%), may be fluctuations

Category	Model	Target vs. Baseline [Eager]	Target vs. Baseline [Inductor]
timm_models_bfloat16_training	swin_base_patch4_window7_224	0.828552	0.804603
timm_models_bfloat16_training	inception_v3	0.836975	0.830448
timm_models_bfloat16_training	repvgg_a2	0.823781	0.837616
timm_models_bfloat16_training	convnextv2_nano.fcmae_ft_in22k_in1k	0.812822	0.854369
torchbench_bfloat16_training	resnext50_32x4d	0.968512	0.862016
torchbench_bfloat16_training	mnasnet1_0	1.071825	0.863574
timm_models_bfloat16_training	deit_tiny_patch16_224.fb_in1k	0.884297	0.890124