feat: Add DBCache release news (#338)

Author: DefTruth

README.md

@@ -50,6 +50,12 @@
 ## 📖 News 🔥🔥
 <div id="news"></div>
 
+- [2025-06-16]: [DBCache](https://github.com/vipshop/DBCache) is release! A **Training-free UNet-style** Cache Acceleration for Diffusion Transformers! Take a try!  
+
+<div align='center'>
+  <img src='https://github.com/user-attachments/assets/a5ec4320-d2f9-4254-888a-170b2d9e3784' height="170px">
+</div>
+
 - [2025-01-08]: [📚Split Q + Fully QKV Fine-grained Tiling](#mma-tiling-qkv) has been refactored into 🤖[ffpa-attn](https://github.com/xlite-dev/ffpa-attn.git): 📚FFPA - Yet another Faster Flash Prefill Attention with O(1)🎉SRAM complexity for headdim > 256, **1.8x~3x**🎉faster than SDPA EA: [📈L20 ~1.9x↑🎉](https://github.com/xlite-dev/ffpa-attn?tab=readme-ov-file#L1-bench-l20), [📈 A30 ~1.8x↑🎉](https://github.com/xlite-dev/ffpa-attn?tab=readme-ov-file#L1-bench-a30), [📈3080 ~2.9x↑🎉](https://github.com/xlite-dev/ffpa-attn?tab=readme-ov-file#L1-bench-3080), [📈4090 ~2.1x↑🎉](https://github.com/xlite-dev/ffpa-attn?tab=readme-ov-file#L1-bench-4090).
 
 <div align='center'>