Takumi Honda

Yuma Ichikawa, Keiji Kimura, Akihiro Yoshida et al.

Deploying foundation models is increasingly constrained by memory footprint, latency, and hardware costs. Post-training compression can mitigate these bottlenecks by reducing the precision of model parameters without significantly degrading performance; however, its practical implementation remains challenging as practitioners navigate a fragmented landscape of quantization algorithms, precision budgets, data-driven calibration strategies, and hardware-dependent execution regimes. We present OneComp, an open-source compression framework that transforms this expert workflow into a reproducible, resource-adaptive pipeline. Given a model identifier and available hardware, OneComp automatically inspects the model, plans mixed-precision assignments, and executes progressive quantization stages, ranging from layer-wise compression to block-wise refinement and global refinement. A key architectural choice is treating the first quantized checkpoint as a deployable pivot, ensuring that each subsequent stage improves the same model and that quality increases as more compute is invested. By converting state-of-the-art compression research into an extensible, open-source, hardware-aware pipeline, OneComp bridges the gap between algorithmic innovation and production-grade model deployment.

Kaiyan Zhao, Tsuguchika Tabaru, Kenichi Kobayashi et al.

Although recent quantized Large Language Models (LLMs), such as BitNet, have paved the way for significant reduction in memory usage during deployment with binary or ternary weights, training these models still demands substantial memory footprints. This is partly because high-precision (i.e., unquantized) weights required for straight-through estimation must be maintained throughout the whole training process. To address this, we explore directly updating the quantized low-precision weights without relying on straight-through estimation during backpropagation, aiming to save memory usage during training. Specifically, we employ a stochastic rounding technique to minimize the information loss caused by the use of low-bit weights throughout training. Experimental results on our LLaMA-structured models of various sizes indicate that (1) training with only low-precision weights is feasible even when they are constrained to ternary values; (2) extending the bit width to 8 bits achieves performance on par with BitNet b1.58; (3) our models remain robust to precision scaling and memory reduction, showing minimal performance degradation when moving from FP32 to lower-memory environments (BF16/FP8); and (4) our models also support inference using ternary weights, showcasing their flexibility in deployment.

Patrik Okanovic, Sameer Deshmukh, Grzegorz Kwasniewski et al.

The energy consumption of large-scale ML models is dominated by data movement, shuffling billions of parameters across memory hierarchies and data centers. Sparsification offers a principled way to mitigate these costs by pruning redundant weights and activations, thereby reducing data movement. Effective sparsification to prune redundant parameters is still challenging: existing methods incur significant accuracy degradation, performance overhead, or both. We introduce (Bl)ock (a)nd (S)parse (T)ransformers (BLaST), a general, robust, and reliable method for sparsification, applicable to linear layers in all settings. Our method iteratively sparsifies weight matrices into a block sparsity pattern suitable for efficient sparse matrix-matrix (SpMM) multiplication. BLaST achieves up to 95% sparsity in MLP weights with negligible accuracy loss (majority <2.25%). We show a 2.2x inference speedup for Llama 3.2 with 16 GPUs, and up to 4.45x reduction in inference memory footprint resulting in a 2.9x reduction in GPU setup and operating costs.

Steven Farrell, Murali Emani, Jacob Balma et al.

Scientific communities are increasingly adopting machine learning and deep learning models in their applications to accelerate scientific insights. High performance computing systems are pushing the frontiers of performance with a rich diversity of hardware resources and massive scale-out capabilities. There is a critical need to understand fair and effective benchmarking of machine learning applications that are representative of real-world scientific use cases. MLPerf is a community-driven standard to benchmark machine learning workloads, focusing on end-to-end performance metrics. In this paper, we introduce MLPerf HPC, a benchmark suite of large-scale scientific machine learning training applications driven by the MLCommons Association. We present the results from the first submission round, including a diverse set of some of the world's largest HPC systems. We develop a systematic framework for their joint analysis and compare them in terms of data staging, algorithmic convergence, and compute performance. As a result, we gain a quantitative understanding of optimizations on different subsystems such as staging and on-node loading of data, compute-unit utilization, and communication scheduling, enabling overall $>10 \times$ (end-to-end) performance improvements through system scaling. Notably, our analysis shows a scale-dependent interplay between the dataset size, a system's memory hierarchy, and training convergence that underlines the importance of near-compute storage. To overcome the data-parallel scalability challenge at large batch sizes, we discuss specific learning techniques and hybrid data-and-model parallelism that are effective on large systems. We conclude by characterizing each benchmark with respect to low-level memory, I/O, and network behavior to parameterize extended roofline performance models in future rounds.

Masafumi Yamazaki, Akihiko Kasagi, Akihiro Tabuchi et al.

There has been a strong demand for algorithms that can execute machine learning as faster as possible and the speed of deep learning has accelerated by 30 times only in the past two years. Distributed deep learning using the large mini-batch is a key technology to address the demand and is a great challenge as it is difficult to achieve high scalability on large clusters without compromising accuracy. In this paper, we introduce optimization methods which we applied to this challenge. We achieved the training time of 74.7 seconds using 2,048 GPUs on ABCI cluster applying these methods. The training throughput is over 1.73 million images/sec and the top-1 validation accuracy is 75.08%.