Current stream processing systems provide exactly-once semantics using checkpointing or a combination of logging and checkpointing. These approaches can introduce high overhead, significantly increasing the latency for normal stream processing because maintaining exactly-once semantics requires coordination across distributed nodes and streams to capture a globally consistent state. We observe that modern distributed shared logs offer a promising solution for maintaining exactly-once semantics with a small overhead. We propose Impeller, a stream processing system that uses a distributed shared log for data storage and exactly-once processing. To maintain exactly-once semantics, Impeller includes a novel and efficient progress marking protocol based on string tags and selective reads in a shared log. The key idea is to leverage the log's record-tagging feature to atomically mark progress across all streams. The experiments over the NEXMark benchmark show that Impeller achieves 1.3× to 5.4× lower p50 latency, or 1.3× to 5.0× higher saturation throughput than Kafka Streams.

The discovery of the *lazy neuron phenomenon* (Li et al., 2022), where fewer than 10% of the feedforward networks (FFN) parameters in trained Transformers are activated per token, has spurred significant interests in *activation sparsity* for enhancing large model efficiency. While notable progress has been made in translating such sparsity to wall-time benefits across CPUs, GPUs, and TPUs, modern Transformers have moved away from the ReLU activation function crucial to this phenomenon. Existing efforts on re-introducing activation sparsity, e.g., by reverting to ReLU or applying top-k masking, often degrade model quality, increase parameter count, or complicate training. Sparse attention, the application of sparse activation to the attention mechanism, often face similar challenges. This paper introduces the Spark Transformer, a novel architecture that achieves high activation sparsity in both FFN and the attention mechanism while maintaining model quality, parameter count, and standard training procedures. Our method realizes sparsity via top-$k$ masking for explicit control over sparsity level. Crucially, we introduce *statistical top-k*, a hardware-accelerator-friendly, linear-time approximate algorithm that avoids costly sorting and mitigates significant training slowdown from standard top-k operators. Furthermore, Spark Transformer reallocates existing FFN parameters and attention key embeddings to form a low-cost predictor for identifying activated entries. This design not only mitigates quality loss from enforced sparsity, but also enhances wall-time benefit. Pretrained with the Gemma-2 recipe, Spark Transformer demonstrates competitive performance on standard benchmarks while exhibiting significant sparsity: only 8\% of FFN neurons are activated, and each token attends to a maximum of 256 tokens. This translates to a 2.5x reduction in FLOPs, leading to decoding wall-time speedups of up to 1.79x on CPU and 1.40xon GPU.

Bokiis a new serverless runtime that exports a shared log API to serverless functions. Boki shared logs enable stateful serverless applications to manage their state with durability, consistency, and fault tolerance. Boki shared logs achieve high throughput and low latency. The key enabler is the metalog , a novel mechanism that allows Boki to address ordering, consistency and fault tolerance independently. The metalog orders shared log records with high throughput and it provides read consistency while allowing service providers to optimize the write and read path of the shared log in different ways. To demonstrate the value of shared logs for stateful serverless applications, we build Boki support libraries that implement fault-tolerant workflows, durable object storage, and message queues. Our evaluation shows that shared logs can speed up important serverless workloads by up to 4.2×.

Ease of use and transparent access to elastic resources have attracted many applications away from traditional platforms toward serverless functions. Many of these applications, such as machine learning, could benefit significantly from GPU acceleration. Unfortunately, GPUs remain inaccessible from serverless functions in modern production settings. We present DGSF, a platform that transparently enables serverless functions to use GPUs through general purpose APIs such as CUDA. DGSF solves provisioning and utilization challenges with disaggregation, serving the needs of a potentially large number of functions through virtual GPUs backed by a small pool of physical GPUs on dedicated servers. Disaggregation allows the provider to decouple GPU provisioning from other resources, and enables significant benefits through consolidation. We describe how DGSF solves GPU disaggregation challenges including supporting API transparency, hiding the latency of communication with remote GPUs, and load-balancing access to heavily shared GPUs. Evaluation of our prototype on six workloads shows that DGSF's API remoting optimizations can improve the runtime of a function by up to 50% relative to unoptimized DGSF. Such optimizations, which aggressively remove GPU runtime and object management latency from the critical path, can enable functions running over DGSF to have a lower end-to-end time than when running on a GPU natively. By enabling GPU sharing, DGSF can reduce function queueing latency by up to 53%. We use DGSF to augment AWS Lambda with GPU support, showing similar benefits.

The microservice architecture is a popular software engineering approach for building flexible, large-scale online services. Serverless functions, or function as a service (FaaS), provide a simple programming model of stateless functions which are a natural substrate for implementing the stateless RPC handlers of microservices, as an alternative to containerized RPC servers. However, current serverless platforms have millisecond-scale runtime overheads, making them unable to meet the strict sub-millisecond latency targets required by existing interactive microservices.

Boki is a new serverless runtime that exports a shared log API to serverless functions. Boki shared logs enable stateful serverless applications to manage their state with durability, consistency, and fault tolerance. Boki shared logs achieve high throughput and low latency. The key enabler is the metalog, a novel mechanism that allows Boki to address ordering, consistency and fault tolerance independently. The metalog orders shared log records with high throughput and it provides read consistency while allowing service providers to optimize the write and read path of the shared log in different ways. To demonstrate the value of shared logs for stateful serverless applications, we build Boki support libraries that implement fault-tolerant workflows, durable object storage, and message queues. Our evaluation shows that shared logs can speed up important serverless workloads by up to 4.7x.

System security has historically relied on hardware-provided isolation primitives. However, Meltdown [36] and Spectre [30] demonstrate that basic user/kernel isolation could be bypassed in every widely deployed ISA for decades; they are a caution to system designers who accept hardware isolation guarantees as an article of faith. Hardware isolation is fallible and should be considered fallible by software systems.

BACKGROUND: Histopathology image analysis is a gold standard for cancer recognition and diagnosis. Automatic analysis of histopathology images can help pathologists diagnose tumor and cancer subtypes, alleviating the workload of pathologists. There are two basic types of tasks in digital histopathology image analysis: image classification and image segmentation. Typical problems with histopathology images that hamper automatic analysis include complex clinical representations, limited quantities of training images in a dataset, and the extremely large size of singular images (usually up to gigapixels). The property of extremely large size for a single image also makes a histopathology image dataset be considered large-scale, even if the number of images in the dataset is limited. RESULTS: In this paper, we propose leveraging deep convolutional neural network (CNN) activation features to perform classification, segmentation and visualization in large-scale tissue histopathology images. Our framework transfers features extracted from CNNs trained by a large natural image database, ImageNet, to histopathology images. We also explore the characteristics of CNN features by visualizing the response of individual neuron components in the last hidden layer. Some of these characteristics reveal biological insights that have been verified by pathologists. According to our experiments, the framework proposed has shown state-of-the-art performance on a brain tumor dataset from the MICCAI 2014 Brain Tumor Digital Pathology Challenge and a colon cancer histopathology image dataset. CONCLUSIONS: The framework proposed is a simple, efficient and effective system for histopathology image automatic analysis. We successfully transfer ImageNet knowledge as deep convolutional activation features to the classification and segmentation of histopathology images with little training data. CNN features are significantly more powerful than expert-designed features.

In this paper, we develop a new weakly supervised learning algorithm to learn to segment cancerous regions in histopathology images. This paper is under a multiple instance learning (MIL) framework with a new formulation, deep weak supervision (DWS); we also propose an effective way to introduce constraints to our neural networks to assist the learning process. The contributions of our algorithm are threefold: 1) we build an end-to-end learning system that segments cancerous regions with fully convolutional networks (FCNs) in which image-to-image weakly-supervised learning is performed; 2) we develop a DWS formulation to exploit multi-scale learning under weak supervision within FCNs; and 3) constraints about positive instances are introduced in our approach to effectively explore additional weakly supervised information that is easy to obtain and enjoy a significant boost to the learning process. The proposed algorithm, abbreviated as DWS-MIL, is easy to implement and can be trained efficiently. Our system demonstrates the state-of-the-art results on large-scale histopathology image data sets and can be applied to various applications in medical imaging beyond histopathology images, such as MRI, CT, and ultrasound images.

We propose a simple, efficient and effective method using deep convolutional activation features (CNNs) to achieve stat- of-the-art classification and segmentation for the MICCAI 2014 Brain Tumor Digital Pathology Challenge. Common traits of such medical image challenges are characterized by large image dimensions (up to the gigabyte size of an image), a limited amount of training data, and significant clinical feature representations. To tackle these challenges, we transfer the features extracted from CNNs trained with a very large general image database to the medical image challenge. In this paper, we used CNN activations trained by ImageNet to extract features (4096 neurons, 13.3% active). In addition, feature selection, feature pooling, and data augmentation are used in our work. Our system obtained 97.5% accuracy on classification and 84% accuracy on segmentation, demonstrating a significant performance gain over other participating teams.