@Hikari_07_jp: I got DeepSeek-V4-Flash MTP speculative decoding actually working on 2× RTX PRO 6000 +38% single-stream throughput. It …

X AI KOLs Timeline Tools

Summary

Achieved DeepSeek-V4-Flash MTP speculative decoding on 2× RTX PRO 6000 with a 38% throughput increase by fixing a mis-routed quantization format issue.

I got DeepSeek-V4-Flash MTP speculative decoding actually working on 2× RTX PRO 6000 +38% single-stream throughput. It was declared “broken on SM120” The kernels weren’t the problem. It was one mis-routed quantization format in the loader ←on 45tok/s off 98tok/s→ https://t.co/MYvNntPC9S
Original Article
View Cached Full Text

Cached at: 06/25/26, 07:14 AM

I got DeepSeek-V4-Flash MTP speculative decoding actually working on 2× RTX PRO 6000 +38% single-stream throughput. It was declared “broken on SM120” The kernels weren’t the problem. It was one mis-routed quantization format in the loader ←on 45tok/s off 98tok/s→ https://t.co/MYvNntPC9S

Similar Articles