Haibing Guan

Oversubscription is central to large-scale clouds: multiplexing virtual CPUs (vCPUs) over fewer physical CPUs (pCPUs) improves utilization and sales density, but meeting strict latency Service Level Objectives (SLOs) requires precise control over idle behavior. In 1:1 settings, idle-passthrough—e...

High-end virtual machines (VMs) have become essential in public clouds for supporting large-scale guest applications. While conventional pre-copy live migration cannot reliably migrate such high-end VMs due to convergence problems, post-copy migration offers a viable alternative. However, high-en...

Arbitrary-Precision Quantization (APQ), which uses asymmetric bit-widths for weights and activations (e.g., W4A8), is a prevalent technique for LLM inference because of its excellent accuracy-performance balance. APQ transforms the general matrix multiplications (GEMM), the core of LLM computatio...

Speculative decoding accelerates autoregressive generation by letting draft tokens bypass full verification, but conventional frameworks suffer from frequent false rejections, particularly when draft models produce semantically correct but lexically divergent outputs. In this paper, we present Ca...