RuntimeError: Unexpected optimization option guard_filter_fn This function isn't available in torch 2.7

### 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

Torch.compile node to work on previous versions of torch as before. Some GPU are locked to this version such as pascal and will be unable to upgrade.

### Actual Behavior

Torch 2.9 compiles, Torch 2.7 fails.

### Steps to Reproduce

Add torch.compile node and try to compile any models.

### Debug Logs

```powershell
RuntimeError: Unexpected optimization option guard_filter_fn, known options are ['TYPE_CHECKING', 'inplace_padding', 'can_inplace_pad_graph_input', 'enable_auto_functionalized_v2', 'debug', 'disable_progress', 'verbose_progress', 'fx_graph_cache', 'fx_graph_remote_cache', 'bundle_triton_into_fx_graph_cache', 'autotune_local_cache', 'autotune_remote_cache', 'bundled_autotune_remote_cache', 'force_disable_caches', 'sleep_sec_TESTING_ONLY', 'custom_op_default_layout_constraint', 'triton_kernel_default_layout_constraint', 'cpp_wrapper', 'online_softmax', 'dce', 'static_weight_shapes', 'size_asserts', 'nan_asserts', 'scalar_asserts', 'pick_loop_orders', 'inplace_buffers', 'allow_buffer_reuse', 'memory_planning', 'use_fast_math', 'memory_pool', 'benchmark_harness', 'epilogue_fusion', 'prologue_fusion', 'epilogue_fusion_first', 'pattern_matcher', 'b2b_gemm_pass', 'post_grad_custom_pre_pass', 'post_grad_custom_post_pass', 'joint_custom_pre_pass', 'joint_custom_post_pass', 'pre_grad_custom_pass', '_pre_fusion_custom_pass', 'split_cat_fx_passes', 'efficient_conv_bn_eval_fx_passes', 'is_predispatch', 'group_fusion', 'batch_fusion', 'pre_grad_fusion_options', 'post_grad_fusion_options', 'reorder_for_locality', 'dynamic_scale_rblock', 'force_fuse_int_mm_with_mul', 'use_mixed_mm', 'fx_passes_numeric_check', 'mixed_mm_choice', 'reorder_for_compute_comm_overlap', 'reorder_for_compute_comm_overlap_passes', 'reorder_for_peak_memory', 'estimate_op_runtime', 'intra_node_bw', 'inter_node_bw', 'use_experimental_benchmarker', 'max_autotune', 'max_autotune_pointwise', 'max_autotune_gemm', 'autotune_num_choices_displayed', 'graph_partition', 'force_same_precision', 'max_autotune_gemm_backends', 'max_autotune_conv_backends', 'max_autotune_gemm_search_space', 'autotune_fallback_to_aten', 'unbacked_symint_fallback', 'search_autotune_cache', 'save_args', 'autotune_in_subproc', 'max_autotune_subproc_result_timeout_seconds', 'max_autotune_subproc_graceful_timeout_seconds', 'max_autotune_subproc_terminate_timeout_seconds', 'autotune_multi_device', 'coordinate_descent_tuning', 'coordinate_descent_check_all_directions', 'coordinate_descent_search_radius', 'autoheuristic_collect', 'autoheuristic_use', 'autoheuristic_log_path', 'layout_opt_default', 'layout_optimization', 'force_layout_optimization', 'keep_output_stride', 'warn_mix_layout', 'realize_reads_threshold', 'realize_opcount_threshold', 'realize_acc_reads_threshold', 'fallback_random', 'implicit_fallbacks', 'aggressive_fusion', 'debug_fusion', 'benchmark_fusion', 'enabled_metric_tables', 'loop_ordering_after_fusion', 'score_fusion_memory_threshold', 'benchmark_epilogue_fusion', 'max_epilogue_benchmarked_choices', 'max_fusion_size', 'max_pointwise_cat_inputs', 'force_pointwise_cat', 'unroll_reductions_threshold', 'comment_origin', 'conv_1x1_as_mm', 'split_reductions', 'benchmark_kernel', 'constant_and_index_propagation', 'always_keep_tensor_constants', 'assert_indirect_indexing', 'compute_all_bounds', 'combo_kernels', 'benchmark_combo_kernel', 'combo_kernels_autotune', 'combo_kernel_allow_mixed_sizes', 'combo_kernel_foreach_dynamic_shapes', 'joint_graph_constant_folding', 'debug_index_asserts', 'emulate_precision_casts', 'is_nightly_or_source', 'developer_warnings', 'optimize_scatter_upon_const_tensor', 'add_pre_grad_passes', 'remove_pre_grad_passes', 'worker_start_method', '_fuse_ddp_communication', '_fuse_ddp_bucket_size', '_fuse_ddp_communication_passes', '_micro_pipeline_tp', '_collective.auto_select', '_collective.one_shot_all_reduce_threshold_bytes', 'compile_threads', 'global_cache_dir', 'kernel_name_max_ops', 'shape_padding', 'comprehensive_padding', 'pad_channels_last', 'disable_padding_cpu', 'padding_alignment_bytes', 'padding_stride_threshold', 'pad_outputs', 'bw_outputs_user_visible', 'force_shape_pad', 'permute_fusion', 'profiler_mark_wrapper_call', 'generate_intermediate_hooks', 'debug_ir_traceback', '_raise_error_for_testing', '_profile_var', 'profile_bandwidth', 'profile_bandwidth_regex', 'profile_bandwidth_output', 'profile_bandwidth_with_do_bench_using_profiling', 'disable_cpp_codegen', 'freezing', 'freezing_discard_parameters', 'decompose_mem_bound_mm', 'assume_aligned_inputs', 'unsafe_ignore_unsupported_triton_autotune_args', 'check_stack_no_cycles_TESTING_ONLY', 'always_complex_memory_overlap_TESTING_ONLY', 'enable_linear_binary_folding', 'annotate_training', 'cpp.threads', 'cpp.no_redundant_loops', 'cpp.dynamic_threads', 'cpp.simdlen', 'cpp.min_chunk_size', 'cpp.cxx', 'cpp.enable_kernel_profile', 'cpp.weight_prepack', 'cpp.inject_relu_bug_TESTING_ONLY', 'cpp.inject_log1p_bug_TESTING_ONLY', 'cpp.vec_isa_ok', 'cpp.descriptive_names', 'cpp.max_horizontal_fusion_size', 'cpp.fallback_scatter_reduce_sum', 'cpp.enable_unsafe_math_opt_flag', 'cpp.enable_floating_point_contract_flag', 'cpp.enable_tiling_heuristics', 'cpp.enable_grouped_gemm_template', 'cpp.gemm_max_k_slices', 'cpp.gemm_cache_blocking', 'cpp.gemm_thread_factors', 'cpp.enable_loop_tail_vec', 'cpp.enable_concat_linear', 'triton.cudagraphs', 'triton.cudagraph_trees', 'triton.cudagraph_skip_dynamic_graphs', 'triton.slow_path_cudagraph_asserts', 'triton.cudagraph_trees_history_recording', 'triton.cudagraph_support_input_mutation', 'triton.cudagraph_unexpected_rerecord_limit', 'triton.cudagraph_dynamic_shape_warn_limit', 'triton.force_cudagraph_sync', 'triton.force_cudagraphs_warmup', 'triton.fast_path_cudagraph_asserts', 'triton.skip_cudagraph_warmup', 'triton.debug_sync_graph', 'triton.debug_sync_kernel', 'triton.dense_indexing', 'triton.max_tiles', 'triton.prefer_nd_tiling', 'triton.autotune_pointwise', 'triton.autotune_cublasLt', 'triton.autotune_at_compile_time', 'triton.tile_reductions', 'triton.tiling_prevents_pointwise_fusion', 'triton.tiling_prevents_reduction_fusion', 'triton.unique_kernel_names', 'triton.unique_user_kernel_names', 'triton.descriptive_names', 'triton.persistent_reductions', 'triton.cooperative_reductions', 'triton.force_cooperative_reductions', 'triton.multi_kernel', 'triton.divisible_by_16', 'triton.min_split_scan_rblock', 'triton.store_cubin', 'triton.spill_threshold', 'triton.use_block_ptr', 'triton.inject_relu_bug_TESTING_ONLY', 'triton.codegen_upcast_to_fp32', 'triton.enable_persistent_tma_matmul', 'triton.skip_l1_cache', 'triton.disallow_failing_autotune_kernels_TESTING_ONLY', 'aot_inductor.output_path', 'aot_inductor.debug_compile', 'aot_inductor.compile_wrapper_with_O0', 'aot_inductor.debug_intermediate_value_printer', 'aot_inductor.filtered_kernel_names', 'aot_inductor.serialized_in_spec', 'aot_inductor.serialized_out_spec', 'aot_inductor.use_runtime_constant_folding', 'aot_inductor.force_mmap_weights', 'aot_inductor.package', 'aot_inductor.package_cpp_only', 'aot_inductor.metadata', 'aot_inductor.raise_error_on_ignored_optimization', 'aot_inductor.dump_aoti_minifier', 'aot_inductor.repro_level', 'aot_inductor.presets', 'aot_inductor.allow_stack_allocation', 'aot_inductor.use_minimal_arrayref_interface', 'aot_inductor.package_constants_in_so', 'cuda.arch', 'cuda.version', 'cuda.compile_opt_level', 'cuda.enable_cuda_lto', 'cuda.enable_ptxas_info', 'cuda.enable_debug_info', 'cuda.use_fast_math', 'cuda.cutlass_dir', 'cuda.cutlass_max_profiling_configs', 'cuda.cutlass_max_profiling_swizzle_options', 'cuda.cuda_cxx', 'cuda.cutlass_backend_min_gemm_size', 'cuda.generate_test_runner', 'cuda.cutlass_op_allowlist_regex', 'cuda.cutlass_op_denylist_regex', 'cuda.cutlass_instantiation_level', 'rocm.arch', 'rocm.ck_supported_arch', 'rocm.compile_opt_level', 'rocm.is_debug', 'rocm.save_temps', 'rocm.use_fast_math', 'rocm.flush_denormals', 'rocm.print_kernel_resource_usage', 'rocm.rocm_home', 'rocm.ck_dir', 'rocm.generate_test_runner', 'rocm.n_max_profiling_configs', 'rocm.use_preselected_instances', 'rocm.kBatch_sweep', 'rocm.split_k_threshold', 'cpu_backend', 'cuda_backend', 'halide.cpu_target', 'halide.gpu_target', 'halide.scheduler_cuda', 'halide.scheduler_cpu', 'halide.asserts', 'halide.debug', 'halide.scan_kernels', 'trace.enabled', 'trace.save_real_tensors', 'trace.debug_dir', 'trace.debug_log', 'trace.info_log', 'trace.fx_graph', 'trace.fx_graph_transformed', 'trace.ir_pre_fusion', 'trace.ir_post_fusion', 'trace.output_code', 'trace.graph_diagram', 'trace.draw_orig_fx_graph', 'trace.dot_graph_shape', 'trace.log_url_for_graph_xform', 'trace.compile_profile', 'trace.upload_tar', 'trace.log_autotuning_results', 'trace.log_inductor_triton_kernel_to_post_grad_node_info', '_save_config_ignore', '_cache_config_ignore_prefix', 'external_matmul', 'test_configs.force_extern_kernel_in_multi_template', 'test_configs.max_mm_configs', 'test_configs.runtime_triton_dtype_assert', 'test_configs.autotune_choice_name_regex', 'test_configs.autotune_choice_desc_regex', 'test_configs.graphsafe_rng_func_ignores_fallback_random']
```

### Other

Honestly there is something weird with mixed-precision casting. On 2.7 I can compile fp8_scaled and mixed using patched triton. On torch 2.9 it passes FP8 directly to triton without casting anything causing compilation failure for ampere and turning. This is a new bug I just ran into after pulling and yea _I_ can just comment it out but what about others?