arXiv subjects

Torsten Hoefler

Publications and source records attributed to Torsten Hoefler.

At least 181 records · Page 10Linked to original sources

arXiv2023

GPTQ: Accurate Post-Training Quantization for Generative Pre-trained Transformers

Generative Pre-trained Transformer models, known as GPT or OPT, set themselves apart through breakthrough performance across complex language modelling tasks, but also by their extremely high computational and storage costs. Specifically, due to their massive size, even inference for large, highly-accurate GPT models may require multiple performant GPUs, which limits the usability of such models. While there is emerging work on relieving this pressure via model compression, the applicability and performance of existing compression techniques is limited by the scale and complexity of GPT models. In this paper, we address this challenge, and propose GPTQ, a new one-shot weight quantization method based on approximate second-order information, that is both highly-accurate and highly-efficient. Specifically, GPTQ can quantize GPT models with 175 billion parameters in approximately four GPU hours, reducing the bitwidth down to 3 or 4 bits per weight, with negligible accuracy degradation relative to the uncompressed baseline. Our method more than doubles the compression gains relative to previously-proposed one-shot quantization methods, preserving accuracy, allowing us for the first time to execute an 175 billion-parameter model inside a single GPU for generative inference. Moreover, we also show that our method can still provide reasonable accuracy in the extreme quantization regime, in which weights are quantized to 2-bit or even ternary quantization levels. We show experimentally that these improvements can be leveraged for end-to-end inference speedups over FP16, of around 3.25x when using high-end GPUs (NVIDIA A100) and 4.5x when using more cost-effective ones (NVIDIA A6000). The implementation is available at https://github.com/IST-DASLab/gptq.

cs.LG↗

arXiv2023

Performance Embeddings: A Similarity-based Approach to Automatic Performance Optimization

Performance optimization is an increasingly challenging but often repetitive task. While each platform has its quirks, the underlying code transformations rely on data movement and computational characteristics that recur across applications. This paper proposes to leverage those similarities by constructing an embedding space for subprograms. The continuous space captures both static and dynamic properties of loop nests via symbolic code analysis and performance profiling, respectively. Performance embeddings enable direct knowledge transfer of performance tuning between applications, which can result from autotuning or tailored improvements. We demonstrate this transfer tuning approach on case studies in deep neural networks, dense and sparse linear algebra compositions, and numerical weather prediction stencils. Transfer tuning reduces the search complexity by up to four orders of magnitude and outperforms the MKL library in sparse-dense matrix multiplication. The results exhibit clear correspondences between program characteristics and optimizations, outperforming prior specialized state-of-the-art approaches and generalizing beyond their capabilities.

cs.SE↗

arXiv2023

Approximate Reversible Circuits for NISQ-Era Quantum Computers

The synthesis approaches for quantum circuits typically aim at minimizing the number of lines or gates. Given the tight restrictions on those logical resources in physical implementations, we propose to view the problem fundamentally different: Given noisy gates and a fixed number of lines, how can we use them to perform a computation as precisely as possible? In this paper we show approximate circuits can be deployed for computations with limited resources. Performing experiments on a QC simulator, we show that under the influence of noise, approximate circuits can have lower error-rates than exact circuits.

cs.ET↗

arXiv2023

The spatial computer: A model for energy-efficient parallel computation

We present a new parallel model of computation suitable for spatial architectures, for which the energy used for communication heavily depends on the distance of the communicating processors. In our model, processors have locations on a conceptual two-dimensional grid, and their distance therein determines their communication cost. In particular, we introduce the energy cost of a spatial computation, which measures the total distance traveled by all messages, and study the depth of communication, which measures the largest number of hops of a chain of messages. We show matching energy lower- and upper bounds for many foundational problems, including sorting, median selection, and matrix multiplication. Our model does not depend on any parameters other than the input shape and size, simplifying algorithm analysis. We also show how to simulate PRAM algorithms in our model and how to obtain results for a more complex model that introduces the size of the local memories of the processors as a parameter.

cs.DS↗

arXiv2023

A Theory of I/O-Efficient Sparse Neural Network Inference

As the accuracy of machine learning models increases at a fast rate, so does their demand for energy and compute resources. On a low level, the major part of these resources is consumed by data movement between different memory units. Modern hardware architectures contain a form of fast memory (e.g., cache, registers), which is small, and a slow memory (e.g., DRAM), which is larger but expensive to access. We can only process data that is stored in fast memory, which incurs data movement (input/output-operations, or I/Os) between the two units. In this paper, we provide a rigorous theoretical analysis of the I/Os needed in sparse feedforward neural network (FFNN) inference. We establish bounds that determine the optimal number of I/Os up to a factor of 2 and present a method that uses a number of I/Os within that range. Much of the I/O-complexity is determined by a few high-level properties of the FFNN (number of inputs, outputs, neurons, and connections), but if we want to get closer to the exact lower bound, the instance-specific sparsity patterns need to be considered. Departing from the 2-optimal computation strategy, we show how to reduce the number of I/Os further with simulated annealing. Complementing this result, we provide an algorithm that constructively generates networks with maximum I/O-efficiency for inference. We test the algorithms and empirically verify our theoretical and algorithmic contributions. In our experiments on real hardware we observe speedups of up to 45$\times$ relative to the standard way of performing inference.

cs.DC↗

arXiv2022

Python FPGA Programming with Data-Centric Multi-Level Design

Although high-level synthesis (HLS) tools have significantly improved programmer productivity over hardware description languages, developing for FPGAs remains tedious and error prone. Programmers must learn and implement a large set of vendor-specific syntax, patterns, and tricks to optimize (or even successfully compile) their applications, while dealing with ever-changing toolflows from the FPGA vendors. We propose a new way to develop, optimize, and compile FPGA programs. The Data-Centric parallel programming (DaCe) framework allows applications to be defined by their dataflow and control flow through the Stateful DataFlow multiGraph (SDFG) representation, capturing the abstract program characteristics, and exposing a plethora of optimization opportunities. In this work, we show how extending SDFGs with multi-level Library Nodes incorporates both domain-specific and platform-specific optimizations into the design flow, enabling knowledge transfer across application domains and FPGA vendors. We present the HLS-based FPGA code generation backend of DaCe, and show how SDFGs are code generated for either FPGA vendor, emitting efficient HLS code that is structured and annotated to implement the desired architecture.

cs.DC↗

arXiv2022

Efficient RDMA Communication Protocols

Developers of networked systems often work with low-level RDMA libraries to tailor network modules to take full advantage of offload capabilities offered by RDMA-capable network controllers. Because of the huge design space of networked data access protocols and variability in capabilities of RDMA infrastructure, developers tend to reinvent and reimplement common data exchange protocols, wasting months of development yet missing various performance and system capabilities. In this work, we summarise and categorize RDMA data exchange protocols and elaborate on what features they can offer to networked systems and what implications they have on their memory and network management.

cs.NI↗

arXiv2022

Neural Graph Databases

Graph databases (GDBs) enable processing and analysis of unstructured, complex, rich, and usually vast graph datasets. Despite the large significance of GDBs in both academia and industry, little effort has been made into integrating them with the predictive power of graph neural networks (GNNs). In this work, we show how to seamlessly combine nearly any GNN model with the computational capabilities of GDBs. For this, we observe that the majority of these systems are based on, or support, a graph data model called the Labeled Property Graph (LPG), where vertices and edges can have arbitrarily complex sets of labels and properties. We then develop LPG2vec, an encoder that transforms an arbitrary LPG dataset into a representation that can be directly used with a broad class of GNNs, including convolutional, attentional, message-passing, and even higher-order or spectral models. In our evaluation, we show that the rich information represented as LPG labels and properties is properly preserved by LPG2vec, and it increases the accuracy of predictions regardless of the targeted learning task or the used GNN model, by up to 34% compared to graphs with no LPG labels/properties. In general, LPG2vec enables combining predictive power of the most powerful GNNs with the full scope of information encoded in the LPG model, paving the way for neural graph databases, a class of systems where the vast complexity of maintained data will benefit from modern and future graph machine learning methods.

cs.LG↗

arXiv2022

Spatial Mixture-of-Experts

Many data have an underlying dependence on spatial location; it may be weather on the Earth, a simulation on a mesh, or a registered image. Yet this feature is rarely taken advantage of, and violates common assumptions made by many neural network layers, such as translation equivariance. Further, many works that do incorporate locality fail to capture fine-grained structure. To address this, we introduce the Spatial Mixture-of-Experts (SMoE) layer, a sparsely-gated layer that learns spatial structure in the input domain and routes experts at a fine-grained level to utilize it. We also develop new techniques to train SMoEs, including a self-supervised routing loss and damping expert errors. Finally, we show strong results for SMoEs on numerous tasks, and set new state-of-the-art results for medium-range weather prediction and post-processing ensemble weather forecasts.

cs.LG↗

arXiv2022

ProbGraph: High-Performance and High-Accuracy Graph Mining with Probabilistic Set Representations

Important graph mining problems such as Clustering are computationally demanding. To significantly accelerate these problems, we propose ProbGraph: a graph representation that enables simple and fast approximate parallel graph mining with strong theoretical guarantees on work, depth, and result accuracy. The key idea is to represent sets of vertices using probabilistic set representations such as Bloom filters. These representations are much faster to process than the original vertex sets thanks to vectorizability and small size. We use these representations as building blocks in important parallel graph mining algorithms such as Clique Counting or Clustering. When enhanced with ProbGraph, these algorithms significantly outperform tuned parallel exact baselines (up to nearly 50x on 32 cores) while ensuring accuracy of more than 90% for many input graph datasets. Our novel bounds and algorithms based on probabilistic set representations with desirable statistical properties are of separate interest for the data analytics community.

cs.DC↗

arXiv2022

Assessing requirements to scale to practical quantum advantage

While quantum computers promise to solve some scientifically and commercially valuable problems thought intractable for classical machines, delivering on this promise will require a large-scale quantum machine. Understanding the impact of architecture design choices for a scaled quantum stack for specific applications, prior to full realization of the quantum system, is an important open challenge. To this end, we develop a framework for quantum resource estimation, abstracting the layers of the stack, to estimate resources required across these layers for large-scale quantum applications. Using a tool that implements this framework, we assess three scaled quantum applications and find that hundreds of thousands to millions of physical qubits are needed to achieve practical quantum advantage. We identify three qubit parameters, namely size, speed, and controllability, that are critical at scale to rendering these applications practical. A goal of our work is to accelerate progress towards practical quantum advantage by enabling the broader community to explore design choices across the stack, from algorithms to qubits.

quant-ph↗

arXiv2022

ENS-10: A Dataset For Post-Processing Ensemble Weather Forecasts

Post-processing ensemble prediction systems can improve the reliability of weather forecasting, especially for extreme event prediction. In recent years, different machine learning models have been developed to improve the quality of weather post-processing. However, these models require a comprehensive dataset of weather simulations to produce high-accuracy results, which comes at a high computational cost to generate. This paper introduces the ENS-10 dataset, consisting of ten ensemble members spanning 20 years (1998-2017). The ensemble members are generated by perturbing numerical weather simulations to capture the chaotic behavior of the Earth. To represent the three-dimensional state of the atmosphere, ENS-10 provides the most relevant atmospheric variables at 11 distinct pressure levels and the surface at 0.5-degree resolution for forecast lead times T=0, 24, and 48 hours (two data points per week). We propose the ENS-10 prediction correction task for improving the forecast quality at a 48-hour lead time through ensemble post-processing. We provide a set of baselines and compare their skill at correcting the predictions of three important atmospheric variables. Moreover, we measure the baselines' skill at improving predictions of extreme weather events using our dataset. The ENS-10 dataset is available under the Creative Commons Attribution 4.0 International (CC BY 4.0) license.

cs.LG↗

arXiv2022

Noise in the Clouds: Influence of Network Performance Variability on Application Scalability

Cloud computing represents an appealing opportunity for cost-effective deployment of HPC workloads on the best-fitting hardware. However, although cloud and on-premise HPC systems offer similar computational resources, their network architecture and performance may differ significantly. For example, these systems use fundamentally different network transport and routing protocols, which may introduce network noise that can eventually limit the application scaling. This work analyzes network performance, scalability, and cost of running HPC workloads on cloud systems. First, we consider latency, bandwidth, and collective communication patterns in detailed small-scale measurements, and then we simulate network performance at a larger scale. We validate our approach on four popular cloud providers and three on-premise HPC systems, showing that network (and also OS) noise can significantly impact performance and cost both at small and large scale.

cs.DC↗

arXiv2022

HammingMesh: A Network Topology for Large-Scale Deep Learning

Numerous microarchitectural optimizations unlocked tremendous processing power for deep neural networks that in turn fueled the AI revolution. With the exhaustion of such optimizations, the growth of modern AI is now gated by the performance of training systems, especially their data movement. Instead of focusing on single accelerators, we investigate data-movement characteristics of large-scale training at full system scale. Based on our workload analysis, we design HammingMesh, a novel network topology that provides high bandwidth at low cost with high job scheduling flexibility. Specifically, HammingMesh can support full bandwidth and isolation to deep learning training jobs with two dimensions of parallelism. Furthermore, it also supports high global bandwidth for generic traffic. Thus, HammingMesh will power future large-scale deep learning systems with extreme bandwidth requirements.

cs.DC↗

arXiv2022

Temporal Vectorization: A Compiler Approach to Automatic Multi-Pumping

The multi-pumping resource sharing technique can overcome the limitations commonly found in single-clocked FPGA designs by allowing hardware components to operate at a higher clock frequency than the surrounding system. However, this optimization cannot be expressed in high levels of abstraction, such as HLS, requiring the use of hand-optimized RTL. In this paper we show how to leverage multiple clock domains for computational subdomains on reconfigurable devices through data movement analysis on high-level programs. We offer a novel view on multi-pumping as a compiler optimization - a superclass of traditional vectorization. As multiple data elements are fed and consumed, the computations are packed temporally rather than spatially. The optimization is applied automatically using an intermediate representation that maps high-level code to HLS. Internally, the optimization injects modules into the generated designs, incorporating RTL for fine-grained control over the clock domains. We obtain a reduction of resource consumption by up to 50% on critical components and 23% on average. For scalable designs, this can enable further parallelism, increasing overall performance.

cs.DC↗

arXiv2022

SAGE: Software-based Attestation for GPU Execution

With the application of machine learning to security-critical and sensitive domains, there is a growing need for integrity and privacy in computation using accelerators, such as GPUs. Unfortunately, the support for trusted execution on GPUs is currently very limited - trusted execution on accelerators is particularly challenging since the attestation mechanism should not reduce performance. Although hardware support for trusted execution on GPUs is emerging, we study purely software-based approaches for trusted GPU execution. A software-only approach offers distinct advantages: (1) complement hardware-based approaches, enhancing security especially when vulnerabilities in the hardware implementation degrade security, (2) operate on GPUs without hardware support for trusted execution, and (3) achieve security without reliance on secrets embedded in the hardware, which can be extracted as history has shown. In this work, we present SAGE, a software-based attestation mechanism for GPU execution. SAGE enables secure code execution on NVIDIA GPUs of the Ampere architecture (A100), providing properties of code integrity and secrecy, computation integrity, as well as data integrity and secrecy - all in the presence of malicious code running on the GPU and CPU. Our evaluation demonstrates that SAGE is already practical today for executing code in a trustworthy way on GPUs without specific hardware support.

cs.CR↗

arXiv2022

A Data-Centric Optimization Framework for Machine Learning

Rapid progress in deep learning is leading to a diverse set of quickly changing models, with a dramatically growing demand for compute. However, as frameworks specialize performance optimization to patterns in popular networks, they implicitly constrain novel and diverse models that drive progress in research. We empower deep learning researchers by defining a flexible and user-customizable pipeline for optimizing training of arbitrary deep neural networks, based on data movement minimization. The pipeline begins with standard networks in PyTorch or ONNX and transforms computation through progressive lowering. We define four levels of general-purpose transformations, from local intra-operator optimizations to global data movement reduction. These operate on a data-centric graph intermediate representation that expresses computation and data movement at all levels of abstraction, including expanding basic operators such as convolutions to their underlying computations. Central to the design is the interactive and introspectable nature of the pipeline. Every part is extensible through a Python API, and can be tuned interactively using a GUI. We demonstrate competitive performance or speedups on ten different networks, with interactive optimizations discovering new opportunities in EfficientNet.

cs.LG↗

arXiv2022

Productive Performance Engineering for Weather and Climate Modeling with Python

Earth system models are developed with a tight coupling to target hardware, often containing specialized code predicated on processor characteristics. This coupling stems from using imperative languages that hard-code computation schedules and layout. We present a detailed account of optimizing the Finite Volume Cubed-Sphere Dynamical Core (FV3), improving productivity and performance. By using a declarative Python-embedded stencil domain-specific language and data-centric optimization, we abstract hardware-specific details and define a semi-automated workflow for analyzing and optimizing weather and climate applications. The workflow utilizes both local and full-program optimization, as well as user-guided fine-tuning. To prune the infeasible global optimization space, we automatically utilize repeating code motifs via a novel transfer tuning approach. On the Piz Daint supercomputer, we scale to 2,400 GPUs, achieving speedups of up to 3.92x over the tuned production implementation at a fraction of the original code.

cs.DC↗