feat: Enable FP8 (E4M3/E5M2) in concat_mla_k for optimize long-context prefill performance and refactor type dispatch for BF16/FP16#3129

qiching wants to merge 2 commits intoflashinfer-ai:mainflashinfer-ai/flashinfer:mainfrom

qiching:fix/concat-mla-k-fp8-supportqiching/flashinfer:fix/concat-mla-k-fp8-supportCopy head branch name to clipboard