Two GPU memory management approaches reduce deep learning model memory usage.
problem Limited GPU memory in deep learning systems.
method Two orthogonal approaches exploiting iterative training algorithm to optimize memory usage.
result Up to 34.2% reduction in memory usage without communication overhead.
New GPU algorithm boosts machine learning with larger datasets.
problem Limited GPU memory restricts training data size.
method Out-of-core GPU gradient boosting algorithm.
result Training larger datasets on GPUs without accuracy loss.
The Long-Short-Term-Memory Recurrent Neural Networks (LSTM RNNs) are a popular class of machine learning models for analyzing sequential data. Their training on modern GPUs, however, is limited by the GPU memory capacity. Our profiling results of the LSTM RNN-based Neural Machine Translation (NMT) model reveal that fea…
Paper proposes a GPU-based system for training massive deep learning models in ads systems.
problem Training massive deep learning models with terabyte-scale parameters in ads systems.
method Hierarchical GPU parameter server with 3-layer storage (GPU High-Bandwidth Memory, CPU main memory, SSD).
result 4-node hierarchical GPU parameter server trains a model 2X faster than a 150-node in-memory system.
New GPU kernels boost deep learning speed and memory efficiency.
problem Sparse deep learning matrices are not well-suited for existing sparse kernels.
method Identified favorable properties of sparse matrices from deep learning, developed high-performance GPU kernels for sparse matrix operations.
result 27% of single-precision peak performance on Nvidia V100 GPUs achieved with new kernels.
RFX accelerates and compresses Random Forests for large datasets.
problem Memory bottleneck in proximity matrices limits Random Forest analysis.
method QLORA compression, CPU TriBlock storage, GPU batch sizing, 3D MDS visualization.
result Proximity-based Random Forest analysis on larger datasets is feasible.
GPU optimization speeds up large-scale classification tasks.
problem Efficiently training large-scale classification models on GPUs.
method Judecious GPU-optimization principles applied to TRON algorithm.
result Significant speedups for logistic regression and SVM classification.
New algorithm reduces costs and latency for large language model inference.
problem Optimizing inference costs and latency for large language models with GPU constraints.
method Formulated as an online scheduling problem with endogenous memory growth, introduced fluid model and WAIT algorithms.
result Reduced costs and latency, especially in near-overloaded and overloaded regimes.
Introduces BMF for efficient matrix factorization of large data.
problem Efficiently factorizing large scale matrices with limited memory.
method Uses block matrix approach and factorization at a block level.
result Demonstrates faster convergence on large matrices.
Deep Neural Networks(DNNs) require huge GPU memory when training on modern image/video databases. Unfortunately, the GPU memory is physically finite, which limits the image resolutions and batch sizes that could be used in training for better DNN performance. Unlike solutions that require physically upgrade GPUs, the G…
KineticSim accelerates financial market simulations 3406x over CPU.
problem Simulating financial markets at scale with multi-agent models is bottlenecked by sequential processing and GPU kernel overhead.
method Formalized and implemented a reusable parallel design pattern for iterative multi-agent reductions in thread-block shared memory.
result Achieved a peak throughput of over 54.7 billion agent-events per second, delivering 3406x speedup over CPU.
KineticSim: A lightweight, high-performance execution engine for real-time market simulators
problem Simulating financial markets at scale with multi-agent models
method Reusable parallel design pattern: persistent, state-carrying clearing for iterative multi-agent reductions
result Reduces per-step critical-path depth from Theta(L+A) to Theta(log L + ceil(A/L))
Quasi-Monte Carlo speeds up option Greeks calculation on GPUs.
problem Efficiently calculating option Greeks for risk management.
method Quasi-Monte Carlo (QMC) combined with GPU acceleration for pathwise sensitivity calculation.
result Increased computational speed and efficiency in estimating option Greeks.
A deep learning strategy outperforms traditional methods in stocks portfolio management.
problem Optimizing stock portfolio performance using machine learning.
method Deep Deterministic Policy Gradient framework with neural networks.
result Compound annual return rate of 14.12% compared to 7 other strategies.
Study shows Transformer and Neural GPU are Turing complete without external memory.
problem Exploring computational power of modern neural network architectures.
method Analyzing computational properties of Transformer and Neural GPU.
result Transformer and Neural GPU are Turing complete without external memory.
Latent Dirichlet Allocation (LDA) is a popular tool for analyzing discrete count data such as text and images. Applications require LDA to handle both large datasets and a large number of topics. Though distributed CPU systems have been used, GPU-based systems have emerged as a promising alternative because of the high…
We describe the multi-GPU gradient boosting algorithm implemented in the XGBoost library (https://github.com/dmlc/xgboost). Our algorithm allows fast, scalable training on multi-GPU systems with all of the features of the XGBoost library. We employ data compression techniques to minimise the usage of scarce GPU memory …
New GPU algorithm speeds up Gaussian Process analysis.
problem Reducing computational complexity for large spatial datasets.
method Implemented three GPU methods for Vecchia Approximation.
result New GPU method outperforms existing methods.
SliceOut speeds up deep learning training without sacrificing accuracy.
problem Frequent model re-training and large model training workloads in deep learning.
method SliceOut uses dropout-inspired scheme to drop contiguous sets of units at random, leveraging GPU memory layout.
result 10-40% speedups and memory reduction with minimal accuracy loss.
Neural architecture search (NAS) has a great impact by automatically designing effective neural network architectures. However, the prohibitive computational demand of conventional NAS algorithms (e.g. 104 GPU hours) makes it difficult to \emph{directly} search the architectures on large-scale tasks (e.g. ImageNet).…
We propose a generic algorithmic building block to accelerate training of machine learning models on heterogeneous compute systems. Our scheme allows to efficiently employ compute accelerators such as GPUs and FPGAs for the training of large-scale machine learning models, when the training data exceeds their memory cap…
Novel framework identifies pump-specific deterioration rates using Bayesian hierarchical hazard modeling and causal discovery.
problem Challenges in asset management due to heterogeneous deterioration rates in pump equipment.
method Bayesian hierarchical hazard modeling with causal discovery, GPU-accelerated No-U-Turn Sampling (NUTS), and DirectLiNGAM.
result Identified striking heterogeneity in deterioration rates, with negative effects 400 times larger than positive effects.
DREAM model improves computational efficiency for non-linear effects in relational event models.
problem Efficiently modeling non-linear effects in dynamic relational networks.
method Introduces Deep Relational Event Additive Model (DREAM) using Neural Additive Models.
result Demonstrates superior computational efficiency compared to traditional REM approaches.
Deep neural network models used for medical image segmentation are large because they are trained with high-resolution three-dimensional (3D) images. Graphics processing units (GPUs) are widely used to accelerate the trainings. However, the memory on a GPU is not large enough to train the models. A popular approach to …
Compact DNNs increase memory footprint and reduce energy efficiency.
problem Designing compact deep neural networks (DNNs) for improved energy efficiency.
method Evaluation of recently proposed compact DNNs on a Tesla P100 GPU.
result Higher number of activations and memory footprint lead to reduced energy efficiency.
Stable Hadamard Memory improves reinforcement learning by efficiently managing memory.
problem Memory models struggle in partially observable reinforcement learning environments.
method Introduces a novel memory model using the Hadamard product for efficient memory management and updates.
result Significantly outperforms state-of-the-art memory-based methods on challenging benchmarks.
Asymmetry PRISM outperforms CPU and GPU solvers for institutional rebalancing.
problem Institutional rebalancing with deadline constraints
method Asymmetry PRISM
result Asymmetry PRISM-CPU is 4.5x to 24.1x faster than the fastest completed reference row in the same lane.
In this paper, we present a novel massively parallel algorithm for accelerating the decision tree building procedure on GPUs (Graphics Processing Units), which is a crucial step in Gradient Boosted Decision Tree (GBDT) and random forests training. Previous GPU based tree building algorithms are based on parallel multi-…
GraphGP: Scalable Gaussian Processes with Vecchia's Approximation
problem Naive Gaussian Process computation limits practical use
method GPU algorithm for Vecchia's approximation
result Linear time and memory requirements for nearly a billion parameters
cuRegOT accelerates GPU-based entropic OT solving.
problem Slow convergence and high computational cost of optimal transport on GPUs.
method High-performance GPU solver with algorithmic and architectural optimizations.
result Significant speedups over state-of-the-art solvers.
Cyclic Data Parallelism reduces memory usage and balances gradient communications.
problem Training large deep learning models requires efficient parallelism to scale.
method Cyclic Data Parallelism shifts micro-batches from simultaneous to sequential execution, balancing memory and gradient communications.
result Cyclic Data Parallelism reduces total memory usage and balances gradient communications.
NetFuse merges different DNN models with varying weights for faster inference.
problem Inference speed of DNN models with different weights cannot be improved using existing techniques.
method NetFuse merges models with the same architecture but different weights and inputs, replacing operations with more general ones.
result NetFuse can speed up DNN inference time up to 3.6x on a NVIDIA V100 GPU.
Proposes RBGP framework for efficient block sparse neural networks.
problem Efficiently exploit structured sparsity patterns for sparse neural networks on GPU.
method Uses Ramanujan Bipartite Graph Product to generate structured multi-level block sparse neural networks.
result Achieves 5-9x and 2-5x runtime gains over unstructured and block sparsity patterns respectively, while maintaining accuracy.
ORIGAMI accelerates ML algorithms by splitting compute tasks between in-memory and off-chip accelerators.
problem Memory bandwidth bottleneck in ML processing.
method Heterogeneous in-memory accelerators and off-chip compute platform, pattern-matching for compute patterns, computation-splitting compiler.
result ORIGAMI outperforms state-of-the-art accelerators in performance and energy-efficiency.
A new parallel algorithm speeds up Hawkes process estimation.
problem Slow maximum likelihood estimation for Hawkes processes.
method Parallel prefix scan for sparse transition matrices.
result Massive speedup with O(N/P) complexity. Banks must manage their trading books, not just value them. Pricing includes valuation adjustments collectively known as XVA (at least credit, funding, capital and tax), so management must also include XVA. In trading book management we focus on pricing, hedging, and allocation of prices or hedging costs to desks on an…
TonY simplifies distributed ML job management.
problem Managing distributed ML jobs is complex and resource-intensive.
method TonY is an open-source orchestrator for distributed ML jobs.
result TonY simplifies distributed ML job management.
MIOpen optimizes deep learning operators for GPUs, accelerating research.
problem Optimizing deep learning operators for efficient GPU execution.
method Highly optimized implementations, fusion, auto-tuning, bfloat16 support.
result Accelerates time to discovery in deep learning research.
DoRA improves adaptation efficiency for large models by factoring norms and fusing kernels.
problem High-rank DoRA is computationally expensive and infeasible on common GPUs.
method Factored norms and fused Triton kernels to reduce memory and speed up computation.
result Fused implementation is up to 2.0x faster for inference and 1.9x faster for gradient computation.
XGBoost implements AFT models for survival regression.
problem Survival regression for time-to-event data.
method Loss functions for AFT models in XGBoost.
result XGBoost's AFT model improves generalization and training speed.
This paper optimizes neural network training by packing multiple models on a single GPU.
problem Efficiently sharing limited training resources among multiple neural network models.
method Proposes a primitive called 'pack' to jointly train multiple models on a single GPU.
result Significant performance improvements for hyperparameter tuning, up to 40% for two models.
PruneTrain speeds up neural network training by dynamically pruning weights.
problem Efficiently training large neural networks with high compute and memory costs.
method Structured group-lasso regularization and reconfiguration techniques to reduce weights and model size.
result Achieved 39% reduction in end-to-end training time for ResNet50 on ImageNet.
Improves memory efficiency for meta-learning with large images.
problem High memory usage in meta-learning for few-shot classification.
method LITE: episodic training scheme that decomposes task gradients and back-propagates only a random subset of images.
result Achieves state-of-the-art accuracy on real-world and challenging benchmarks.
Paper explores reducing precision in SVM for faster text classification.
problem Efficiency in multi-class text classification training.
method Comparison of SVM trained with reduced precision (16-bit, half) vs original.
result Reduced precision training maintains text classification accuracy.
D-CBRS manages memory for continual learning by accounting for intra-class diversity.
problem Forgetting in continual learning, especially with class-imbalanced data.
method D-CBRS introduces a novel approach to store instances in memory, considering intra-class diversity.
result D-CBRS outperforms existing methods on data sets with intra-class diversity.
This paper proposes an alternative to E2E training for deep networks, reducing memory footprint.
problem High GPUs memory footprint in end-to-end training of deep networks.
method Locally supervised learning with information propagation loss to avoid information collapse.
result The proposed method achieves competitive performance with less than 40% memory footprint compared to E2E training.
Dual memory improves reinforcement learning efficiency.
problem Training inefficiency in reinforcement learning.
method Introduces a dual memory structure with a main and cache memory.
result Dual memory structure leads to higher scores in reinforcement learning environments.
NVIDIA cuDNN is a low-level library that provides GPU kernels frequently used in deep learning. Specifically, cuDNN implements several equivalent convolution algorithms, whose performance and memory footprint may vary considerably, depending on the layer dimensions. When an algorithm is automatically selected by cuDNN,…