SGLang加入DSpark后，高并发场景下吞吐能提升20%，在DeepSeek-V4上实测383.7 tok/s，推理优化党可以试试。
SGLang现支持DSpark算法，通过置信度驱动的变长验证优化推测解码。与传统验证所有候选token不同，DSpark仅验证高置信度token，在批大小1-256的DeepSeek-V4-Flash上实现最佳吞吐延迟权衡，优于MTP和非推测解码方案。高并发下动态调度相比固定预算带来约20%吞吐提升。使用融合内核和零开销调度后，DeepSeek-V4-Pro在B300上达到383.7 tok/s。
SGLang now supports DSpark, enabling confidence-dr…
SGLang now supports DSpark, enabling confidence-driven, variable-length verification for speculative decoding 🎉
DSpark addresses a key bottleneck under load: instead of verifying every draft token, it verifies only where the draft model is confident, so the gains hold even as batch size scales.
We heavily optimized variable-length verification in SGLang. Across batch sizes 1 to 256, DSpark gives the best throughput/latency tradeoff on DeepSeek-V4-Flash, ahead of both MTP and non-spec.
At high concurrency, dynamic scheduling provides up to ~20% higher throughput compared to a fixed budget, while maintaining high verification quality across workloads.
With fused kernels and zero-overhead scheduling, DeepSeek-V4-Pro reaches 383.7 tok/s at B=1 on B300.
DSpark is now available in SGLang with support for Qwen3 and DeepSeek-V4. Thanks @deepseek_ai for open-sourcing!
Blog with full technical details and commands to run below 👇