arXiv subjects

Qinqin Chen

Publications and source records attributed to Qinqin Chen.

3 recordsLinked to original sources

arXiv2026

X-Stage: Modeling Post-Issue Backpressure in GPU Communication--Computation Fusion

Fine-grained, device-initiated communication allows fused GPU kernels to issue remote stores directly from their compute pipelines, a pattern increasingly used in expert parallelism (EP), tensor parallelism (TP), and Ulysses-style sequence parallelism (UP). Existing designs reason about where communication is issued and when remote data becomes ready, but lack a quantitative model of the sender-side interval after a remote store is accepted and before it becomes visible at the destination. This interval determines whether communication remains decoupled from computation or backpressures it. We identify X-Stage, a software-visible post-issue stage with finite decoupling. Downstream pressure can dissipate while the issuer resumes useful work, whereas sustained injection consumes X-Stage headroom and eventually stalls the compute pipeline. We characterize this behavior and build a calibrated model that predicts whether remote-store arrivals accumulate backpressure or recover during intervening computation. Guided by the model, we reshape bursty arrivals when they would exhaust X-Stage headroom and exploit natural compute windows when headroom can recover concurrently. Evaluation across representative EP, TP, and UP workloads shows up to 1.62x fused-kernel, 1.75x end-to-end, and 1.43x sender-visible speedup, respectively. Microbenchmarks further validate the model's predictions of backlog accumulation, recovery, and sender-side backpressure.

cs.DC↗

arXiv2026

ScalingAttention: Discovering Intrinsic Sparse Attention Topology for Video Diffusion Transformers

While Diffusion Transformers (DiTs) have revolutionized high-fidelity video generation, their reliance on 3D full attention creates a quadratic computational bottleneck. Existing sparse methods face a dilemma: dynamic pruning suffers from prohibitive runtime overhead and memory fragmentation, while static heuristics fail to capture fine-grained dependencies. In this work, we propose ScalingAttention, a training-free framework grounded in a key inductive bias: while individual activations are input-dependent, the high-mass attention regions for each head rapidly converge to a stable, prompt-agnostic Intrinsic Sparse Topology. This topology is weight-encoded, scale-invariant, and efficient to extract. ScalingAttention decouples topology discovery from sparsity control via: (1) WEST (Weight-Encoded Sparse Topology), which extracts a robust block-sparse prior mask offline to eliminate runtime search; (2) FAST (Fidelity-Aware Sensitivity Tuning), which adaptively tunes head-wise sparsity based on diffusion fidelity requirements. To ensure practical acceleration, we co-design a hardware-aligned bit-wise block-sparse kernel. Experiments on Wan2.1 show up to 1.90X end-to-end speedup with superior fidelity, establishing a new Pareto frontier over state-of-the-art baselines.

cs.CV↗

arXiv2015

A Closed-Loop UL Power Control Scheme for Interference Mitigation in Dynamic TD-LTE Systems

The TD-LTE system is envisaged to adopt dynamic time division duplexing (TDD) transmissions for small cells to adapt their communication service to the fast variation of downlink (DL) and uplink (UL) traffic demands. However, different DL/UL directions for the same subframe in adjacent cells will result in new destructive interference components, i.e., eNB-to-eNB and UE-to-UE, with levels that can significantly differ from one subframe to another. In this paper, a feasible UL power control mechanism is proposed to manage eNB-to-eNB interference, where different UL power control parameters are set based on different interference level. We consider the geometric location information and the subframe set selection process about adjacent eNBs when the interference level is estimated. The performance of the proposed scheme is evaluated through system level simulations and it is shown that the scheme can achieve preferable improvement in terms of UL average and 5%-ile packet throughputs compared with the original scheme without power control. Also, the UE-to-UE interference is not worse when the UE transmit power become higher.

cs.NI↗