Related papers: A Generic Checkpoint-Restart Mechanism for Virtual…
A distributed system consisting of a huge number of computational entities is prone to faults, because faults in a few nodes cause the entire system to fail. Consequently, fault tolerance of distributed systems is a critical issue.…
Scaling supercomputers comes with an increase in failure rates due to the increasing number of hardware components. In standard practice, applications are made resilient through checkpointing data and restarting execution after a failure…
The success of Google's Pregel framework in distributed graph processing has inspired a surging interest in developing Pregel-like platforms featuring a user-friendly "think like a vertex" programming model. Existing Pregel-like systems…
The ever-increasing complexity of HW/SW systems presents a persistent challenge, particularly in safety-critical domains like automotive, where extensive testing is imperative. However, the availability of hardware often lags behind,…
As LLMs and foundation models scale, checkpoint/restore has become a critical pattern for training and inference. With 3D parallelism (tensor, pipeline, data), checkpointing involves many processes, each managing numerous tensors of varying…
Code initialization -- the step of loading code, executing static code, filling caches, and forming re-used connections -- tends to dominate cold-start time in serverless compute systems such as AWS Lambda. Post-initialization memory…
Snapshot testing has emerged as a critical technique for UI validation in modern software development, yet it suffers from substantial maintenance overhead due to frequent UI changes causing test failures that require manual inspection to…
MPI accomplishes portable, standardized message-passing between processes by exposing a standard API that hides the implementation of the underlying mechanism for message passing. Until now, checkpointing an MPI program required knowledge…
Creating new materials, discovering new drugs, and simulating systems are essential processes for research and innovation and require substantial computational power. While many applications can be split into many smaller independent tasks,…
Virtualization, after having found widespread adoption in the server and desktop arena, is poised to change the architecture of embedded systems as well. The benefits afforded by virtualization - enhanced isolation, manageability,…
Fuzzing is one of the most popular and widely used techniques to find vulnerabilities in any application. Fuzzers are fast enough, but they still spend a good portion of time to restart a crashed application and then fuzz it from the…
The rise of million-token, agent-based applications has placed unprecedented demands on large language model (LLM) inference services. The long-running nature of these tasks increases their susceptibility to hardware and software faults,…
Deep learning (DL) applications are increasingly being deployed on HPC systems, to leverage the massive parallelism and computing power of those systems for DL model training. While significant effort has been put to facilitate distributed…
Efficient construction of checkpoints/snapshots is a critical tool for training and diagnosing deep learning models. In this paper, we propose a lossy compression scheme for checkpoint constructions (called LC-Checkpoint). LC-Checkpoint…
We focus on the problem of checkpointing (or taking a snapshot) in fully replicated eventually consistent distributed databases. In particular, we consider the problem of taking Distributed Transaction-Consistent Snapshots (DTCS). A typical…
Serverless computing has seen rapid adoption due to its high scalability and flexible, pay-as-you-go billing model. In serverless, developers structure their services as a collection of functions, sporadically invoked by various events like…
Training LLMs on decentralized nodes or on-spot instances, lowers the training cost and enables model democratization. The inevitable challenge here is the transient churns of nodes due to failures and the operator's scheduling policies,…
To efficiently scale large model (LM) training, researchers transition from data parallelism (DP) to hybrid parallelism (HP) on GPU clusters, which frequently experience hardware and software failures. Existing works introduce in-memory…
Model checkpoints are critical Deep Learning (DL) artifacts that enable fault tolerance for training and downstream applications, such as inference. However, writing checkpoints to persistent storage, and other I/O aspects of DL training,…
Resilience is a major design goal for HPC. Checkpoint is the most common method to enable resilient HPC. Checkpoint periodically saves critical data objects to non-volatile storage to enable data persistence. However, using checkpoint, we…