Fix topk inference performance reduce

[lambert0312]

Motivation

When the following logic is added to topk.py, the inference performance will be significantly affected:

sglang/python/sglang/srt/layers/moe/topk.py

Lines 267 to 269 in 6632489

	torch.compile(
	_mask_topk_ids_padded_region, dynamic=True, backend=get_compiler_backend()
	)(topk_ids, num_token_non_padded)

Run command:

python3 -m sglang.launch_server --model-path /path/to/DeepSeek-V3-0324 --trust-remote-code --host 0.0.0.0 --port 30000 --attention-backend flashinfer --n-share-experts-fusion 16 --tp 16 --dist-init-addr IP:20000 --nnodes 2 --node-rank 0

Ref: #6175

Modifications

Add num_token_non_padded judgment logic. If it is None, directly return the previous result.

Checklist

• Format your code according to the Code Formatting with Pre-Commit.
• Add unit tests as outlined in the Running Unit Tests.
• Update documentation / docstrings / example tutorials as needed, according to Writing Documentation.
• Provide throughput / latency benchmark results and accuracy evaluation results as needed, according to Benchmark and Profiling and Accuracy Results.
• For reviewers: If you haven't made any contributions to this PR and are only assisting with merging the main branch, please remove yourself as a co-author when merging the PR.
• Please feel free to join our Slack channel at https://slack.sglang.ai to discuss your PR.