<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Publications | Mingkai Zheng's Homepage</title><link>https://zhengmk.github.io/publication/</link><atom:link href="https://zhengmk.github.io/publication/index.xml" rel="self" type="application/rss+xml"/><description>Publications</description><generator>Wowchemy (https://wowchemy.com)</generator><language>en-us</language><lastBuildDate>Mon, 28 Sep 2026 00:00:00 +0000</lastBuildDate><image><url>https://zhengmk.github.io/media/icon_hu0158c3d4eaa2a6b730ca31a67c18838d_142122_512x512_fill_lanczos_center_2.png</url><title>Publications</title><link>https://zhengmk.github.io/publication/</link></image><item><title>Broken Symmetry in BF16 Attention: Why FlashAttention Gradients Blow Up Late in Training</title><link>https://zhengmk.github.io/publication/broken-symmetry-in-bf16-attention-why-flashattention-gradients-blow-up-late-in-training/</link><pubDate>Mon, 28 Sep 2026 00:00:00 +0000</pubDate><guid>https://zhengmk.github.io/publication/broken-symmetry-in-bf16-attention-why-flashattention-gradients-blow-up-late-in-training/</guid><description/></item><item><title>Diamond Agent: Agentic Control of Federated HPC Resources as a Service</title><link>https://zhengmk.github.io/publication/diamond-agent-agentic-control-of-federated-hpc-resources-as-a-service/</link><pubDate>Sat, 05 Sep 2026 00:00:00 +0000</pubDate><guid>https://zhengmk.github.io/publication/diamond-agent-agentic-control-of-federated-hpc-resources-as-a-service/</guid><description/></item><item><title>A Comparative Study on LLM-enabled HPC User Support</title><link>https://zhengmk.github.io/publication/a-comparative-study-on-llm-enabled-hpc-user-support/</link><pubDate>Sun, 26 Jul 2026 00:00:00 +0000</pubDate><guid>https://zhengmk.github.io/publication/a-comparative-study-on-llm-enabled-hpc-user-support/</guid><description/></item><item><title>PHOENIX: Resilient LLM Training with Hot-Swapping via Zero-Overhead Checkpoint</title><link>https://zhengmk.github.io/publication/phoenix-resilient-llm-training-with-hot-swapping-via-zero-overhead-checkpoint/</link><pubDate>Thu, 02 Jul 2026 00:00:00 +0000</pubDate><guid>https://zhengmk.github.io/publication/phoenix-resilient-llm-training-with-hot-swapping-via-zero-overhead-checkpoint/</guid><description/></item><item><title>SCAPE: Accurate and Efficient LLM Training with Extreme Sparse Communication</title><link>https://zhengmk.github.io/publication/scape-accurate-and-efficient-llm-training-with-extreme-sparse-communication/</link><pubDate>Thu, 02 Jul 2026 00:00:00 +0000</pubDate><guid>https://zhengmk.github.io/publication/scape-accurate-and-efficient-llm-training-with-extreme-sparse-communication/</guid><description/></item><item><title>Quantifying Performance Variability in GPU Clusters</title><link>https://zhengmk.github.io/publication/quantifying-performance-variability-in-gpu-clusters/</link><pubDate>Mon, 01 Jun 2026 00:00:00 +0000</pubDate><guid>https://zhengmk.github.io/publication/quantifying-performance-variability-in-gpu-clusters/</guid><description/></item><item><title>Efficient Fine-Grained GPU Performance Modeling for Distributed Deep Learning of LLM</title><link>https://zhengmk.github.io/publication/efficient-fine-grained-gpu-performance-modeling-for-distributed-deep-learning-of-llm/</link><pubDate>Wed, 17 Dec 2025 00:00:00 +0000</pubDate><guid>https://zhengmk.github.io/publication/efficient-fine-grained-gpu-performance-modeling-for-distributed-deep-learning-of-llm/</guid><description/></item><item><title>Radius: Range-based Gradient Sparsity for Large Foundation Model Pre-training</title><link>https://zhengmk.github.io/publication/radius-range-based-gradient-sparsity-for-large-foundation-model-pre-training/</link><pubDate>Mon, 12 May 2025 00:00:00 +0000</pubDate><guid>https://zhengmk.github.io/publication/radius-range-based-gradient-sparsity-for-large-foundation-model-pre-training/</guid><description/></item><item><title>COMPSO: Optimizing Gradient Compression for Distributed Training with Second-Order Optimizers</title><link>https://zhengmk.github.io/publication/compso-optimizing-gradient-compression-for-distributed-training-with-second-order-optimizers/</link><pubDate>Fri, 28 Feb 2025 00:00:00 +0000</pubDate><guid>https://zhengmk.github.io/publication/compso-optimizing-gradient-compression-for-distributed-training-with-second-order-optimizers/</guid><description/></item><item><title>Dynamic Response Improvement for DAB Converter with Constant Power Load under Extended-Phase-Shift Control Based on Trajectory Control</title><link>https://zhengmk.github.io/publication/dynamic-response-improvement-for-dab-converter-with-constant-power-load-under-extended-phase-shift-control-based-on-trajectory-control/</link><pubDate>Sun, 29 Nov 2020 10:26:26 +0000</pubDate><guid>https://zhengmk.github.io/publication/dynamic-response-improvement-for-dab-converter-with-constant-power-load-under-extended-phase-shift-control-based-on-trajectory-control/</guid><description/></item><item><title>Constant Power Load stabilization Based on Trajectory Control in Dual-Active-Bridge DC-DC Converter</title><link>https://zhengmk.github.io/publication/constant-power-load-stabilization-based-on-trajectory-control-in-dual-active-bridge-dc-dc-converter/</link><pubDate>Sun, 29 Nov 2020 10:06:01 +0000</pubDate><guid>https://zhengmk.github.io/publication/constant-power-load-stabilization-based-on-trajectory-control-in-dual-active-bridge-dc-dc-converter/</guid><description/></item><item><title>Open-Circuit Fault Diagnosis of Dual Active Bridge DC-DC Converter With Extended-Phase-Shift Control</title><link>https://zhengmk.github.io/publication/open-circuit-fault-diagnosis-of-dual-active-bridge-dc-dc-converter-with-extended-phase-shift-control/</link><pubDate>Wed, 13 Feb 2019 00:00:00 +0000</pubDate><guid>https://zhengmk.github.io/publication/open-circuit-fault-diagnosis-of-dual-active-bridge-dc-dc-converter-with-extended-phase-shift-control/</guid><description/></item></channel></rss>