The demand for long-context processing in large language models (LLMs) continues to escalate alongside rapid advancements in their capabilities. However, the intermediate attention keys and values (KV cache) employed to avoid re-computations, also grow linearly with sequence length, far exceeding the memory capacity of consumer-grade GPUs. Consequently, many studies have proposed KV cache compression methods that evict unimportant tokens based on variant attention scoring strategies. These methods typically retain the KV pairs of the top-k scoring tokens under a fixed memory budget. However, they still face several limitations. First, they disregard the activation frequency of tokens, specifically the count of times tokens achieve top-k scores in the attention distribution of following tokens. The methods based on variant attention scores may incorrectly evict some high-activation-frequency yet low final-scoring tokens. Second, the activation frequency exhibits different distribution patterns across layers and tasks. Neglecting these differences negatively impacts model performance and task adaptability. Our analysis of the actual token activation frequency and its unique characteristics across layers and task types reveals potential opportunities to address these issues. In this paper, we propose HitKV, which employs hit rates to directly characterize token activation frequencies, enabling adaptive layer-aware and task-aware KV cache eviction under the uniform memory allocation strategies. Also, HitKV can be easily integrated into layer-specific memory allocation methods. Experimental results demonstrate that HitKV maintains model performance with preserving only 3% of the KV cache, achieves high-quality generation outputs in long-text generation tasks, and delivers 4× throughput improvement over baselines.

The block-based inference engine, powered by noncontiguous key-value (KV) cache management, has emerged as a new paradigm for large language model (LLM) inference due to its efficient memory utilization. However, in large-batch, longcontext workloads, the substantial demand for KV cache remains a major bottleneck in inference performance. Existing research leverages the sparsity of the attention mechanisms by removing non-critical tokens to limit the KV cache size. However, we observe that in block-based inference engines, current sparse methods require defragmentation after token removal to maintain tensor continuity, incurring significant overhead. Additionally, we find a correlation between request length and sparsity potential, yet existing methods apply a uniform sparsity strategy at batch level without dynamic adjustment. To address these, we propose DualSpar, a novel sparse KV cache framework with dual-granularity memory and adaptive sparsity strategy. First, it binds token importance to KV cache block granularity, achieving low-overhead pre-consolidation. Second, it incorporates system load and request length into sparsity decision, fully exploiting the sparse potential of different requests while reducing the queuing latency in large-batch processing. Evaluations show that DualSpar achieves up to a 3.16× throughput improvement, a 3.75× faster time-to-first-token (TTFT), and an 87.2% reduction in defragmentation overhead while maintaining high accuracy.

In the paper, we focus on exploring to construct a novel digital gas fields produce real-time scheduling system, which is based on the theory of intelligent autonomous decentralized system (IADS). The system combines the practical demand and the characteristics of the gas fields, and it aims at dealing with the real-time property, dynamic and complexity during gas fields produce scheduling. Besides embodying on-line intelligent expansion, intelligent fault tolerance and on-line intelligent maintenance of IADS particular properties, the scheme adequately attaches importance to the flexibility. The model & method based on intelligent information pull push (IIPP) and intelligent management (IM) is applied to the system, thus it is helpful to improve the performance of the scheduling system. In according with the current requirement of gas fields produce scheduling, the concrete solvent is put forward. The related system architecture and software structure is presented. An effective method of solving the real-time property was developed for use in scheduling of the gas field produce. We simulated some experiments, and the result demonstrates the validity of the system. At Last, we predict its promising research trend in the future of gas fields practice application.