Speculative decoding improves LLM inference by generating and verifying multiple tokens in parallel, but existing systems suffer from suboptimal performance due to a mismatch between dynamic speculation and static runtime assumptions. We present Yggdrasil, a co-designed system that enables latency-optimal speculative decoding through context-aware tree drafting and compiler-friendly execution. Yggdrasil introduces an equal-growth tree structure for static graph compatibility, a latency-aware optimization objective for draft selection, and stage-based scheduling to reduce overhead. Yggdrasil supports unmodified LLMs and achieves up to $3.98\times$ speedup over state-of-the-art baselines across multiple hardware setups.

Model pruning, which eliminates redundant parameters and reduces computational complexity, emerges as a viable strategy for efficient deep neural network (DNN) deployment. Owing to the irregular memory access and computation patterns in the sparse DNN models after pruning, existing arts have suggested various structured sparse patterns to enhance sparse DNN performance. In this work, we propose a unique perspective of understanding existing sparse pattern design as computation-skipping after tiling the tensor computation into multi-level hierarchies. This unified perspective opens up a new design space of multi-level sparse tiling to maximize the sparsity benefits of DNNs, as opposed to the single-level choice in current practices. We present Fractal, an auto-tuning system for sparse patterns that identifies the optimal multi-level sparse tiling pattern. We introduce PatternIR, a novel high-level intermediate representation (IR), to express a diverse range of multi-level sparse patterns. By leveraging insights from prior dense operator optimizations, we translate PatternIR into low-level compiler IRs, facilitating further operator optimization and code generation. Our evaluations demonstrate that Fractal yields substantial speedups of up to on average 3.16× on CUDA Core, 2.52× on TensorCore of GPUs compared to the state-of-art dense baseline under 75% sparsity while upholding minimal accuracy degradation compared to prior sparse operator libraries.

This work presents a spectrally resolved Raman lidar (SRRL) for simultaneous measurement of volume backscattering coefficient spectra (backscatter spectra for short) of atmospheric three-phase water and fluorescent aerosols. The SRRL emits 354.8-nm laser light and records both N2Raman echoes around 386.7 nm and spectral signals in a specially modified spectral range of ~393–424 nm. By defining normalized spectra, a unique spectra decomposition approach is developed for successive retrieval of normalized spectra components of fluorescent aerosols, water vapor, ice water, and liquid water. Furthermore, the backscatter spectra of each component are obtained by referencing the N2Raman signals. Three typical measurement cases are provided. For the clear-day case, the lidar-measured normalized water vapor Raman spectra are found to be nearly invariant in shape and can serve as background spectra reference for decomposition of mixed-phase water Raman spectra. Besides, fluorescence backscatter intensities are usually weak enough to be neglected in clear areas. For the fluorescence case, the fluorescence backscatter intensities become much stronger to prevent accurate measurements of water vapor. For the mixed-phase water virga case, the decomposed backscatter spectra of condensed ice and liquid water indicate the coexistence of ice water and less amount of liquid water in the falling water virga. In conclusion, this kind of SRRL is believed to have provided the potential for simultaneous and accurate profiling of atmospheric water vapor and cloud liquid/ice water and opened up new perspectives for studies of cloud–aerosol interaction.