SplitZip: Ultra Fast Lossless KV Compression for Disaggregated LLM Serving
For large-scale LLM serving with prefill-decode disaggregation, SplitZip addresses the critical bottleneck of KV-cache transfer between workers, enabling faster and more efficient serving of long-input and agentic workloads.
SplitZip is a GPU-friendly lossless compressor for KV-cache transfer in disaggregated LLM serving, achieving 613.3 GB/s compression and 2181.8 GB/s decompression throughput, with up to 1.32× speedup for KV-cache transfer and 1.30× improvement in time-to-first-token.
Contemporary systems serving large language models (LLMs) have adopted prefill-decode disaggregation to better load-balance between the compute-bound prefill phase and the memory-bound decode phase. Under this design, prefill workers generate a KV cache that must be transferred to decode workers before token generation can begin. With these workers residing on different physical systems, this transfer becomes a significant bottleneck to serving LLMs at scale. This bottleneck gets exacerbated for long-input and agentic workloads, which typically require long inputs. Existing lossless codecs are not well suited to this setting as they primarily target offline weight compression, rely on CPU-side, or use variable-length coding that decompresses fast but compresses too slowly for online use. SplitZip is a GPU-friendly lossless compressor for KV-cache transfer. It exploits redundancy in floating-point exponents of KV activations, encoding the most frequent exponent values with fixed-length codes, and encoding (position, value) pairs and value of rare exponents in an escape stream. An offline calibrated top-16 exponent codebook enables online encoding, while the regular dense path and sparse escape correction make both encoding and decoding efficient on GPUs. On real BF16 activation tensors, SplitZip achieves 613.3 GB/s compression throughput and 2181.8 GB/s decompression throughput, substantially outperforming prior lossless compressors on the latency-critical codec path. End-to-end transfer experiments show up to 1.32$\times$ speedup for BF16 KV-cache transfer, 1.30$\times$ speedup for TTFT and 1.23$\times$ increase on Request Throughput.