Teaspiller was a US-based web application for customers to find accountants and hire them to do their taxes and accounting online. In 2013 the company was acquired by Intuit, Inc and added to its TurboTax product line. The Teaspiller employees and code were all acquired and the product was renamed as "TurboTax CPA select". It enabled accountants to work remotely with clients (share files, send secure messages, schedule appointments), as well as find new clients looking for their specific skills through a complex search algorithm. This was done through extended profiles containing licensing information, professional histories, user ratings, peer endorsements, association memberships, and practice areas. The service had been called an H&R Block killer by Business Insider as it helped customers find accountants to prepare tax returns online. As of 2011 it had 20,000 US accountants listed on the site. The application was built using the Django framework. == History == Teaspiller was built by Vemdara, LLC, a web company based in New York and founded in 2009 by Amit Vemuri (a former VP at Travelocity). The web application was launched in 2010. In 2013 the company was acquired by Intuit as part of their TurboTax product line and renamed as "TurboTax CPA select".
Depth peeling
In computer graphics, depth peeling is an exact multipass method of order-independent transparency that extracts transparent fragments into depth layers and composites those layers in depth order. Depth peeling has the advantage of being able to generate correct results even for complex images containing intersecting transparent objects. == Method == Depth peeling works by rendering the image multiple times. Depth peeling uses two Z buffers, one that works conventionally, and one that is not modified, and sets the minimum distance at which a fragment can be drawn without being discarded. For each pass, the previous pass' conventional Z-buffer is used as the minimal Z-buffer, so each pass removes already-captured nearer fragments and draws the next depth layer behind them. The resulting images can then be composited in depth order to form a single image. A major drawback of classical depth peeling is performance: it requires one geometry pass per peeled layer, so scenes with high depth complexity require many passes that each re-rasterize the transparent geometry. Later variants reduce the number of passes by peeling multiple layers or both front and back layers in a pass. Dual depth peeling reduces the geometry-pass count from N to N/2+1 by peeling one layer from the front and one from the back in each pass, while multi-layer depth peeling peels several layers per pass and reported up to an 8x speed-up in RGBA8 settings.
Cloud-based quantum computing
Cloud-based quantum computing refers to the remote access of quantum computing resources—such as quantum emulators, simulators, or processors—via the internet. Cloud access enables users to develop, test, and execute quantum algorithms without the need for direct interaction with specialized hardware, facilitating broader participation in quantum software development and experimentation. In 2016, IBM launched the IBM Quantum Experience, one of the first publicly accessible quantum processors connected to the cloud. In early 2017, researchers at Rigetti Computing demonstrated programmable quantum cloud access through their software platform Forest, which included the pyQuil Python library. Since the early-2020s, cloud-based quantum computing has grown significantly, with multiple providers offering access to a variety of quantum hardware modalities, including superconducting qubits, trapped ions, neutral atoms, and photonic systems. Major platforms such as Amazon Braket, Azure Quantum, and qBraid aggregate quantum devices from hardware developers like IonQ, Rigetti Computing, QuEra, Pasqal, Oxford Quantum Circuits, and IBM Quantum. These platforms provide unified interfaces for users to write and execute quantum algorithms across diverse backends, often supporting open-source SDKs such as Qiskit, Cirq, and PennyLane. The proliferation of cloud-based access has played a key role in accelerating quantum education, algorithm research, and early-stage application development by lowering the barrier to experimentation with real quantum hardware. Cloud-based quantum computing has expanded access to quantum hardware and tools beyond traditional research laboratories. These platforms support educational initiatives, algorithm development, and early-stage commercial applications. == Applications == Cloud-based quantum computing is used across education, research, and software development, offering remote access to quantum systems without the need for on-site infrastructure. === Education === Quantum cloud platforms have become valuable tools in education, allowing students and instructors to engage with real quantum processors through user-friendly interfaces. Educators use these platforms to teach foundational concepts in quantum mechanics and quantum computing, as well as to demonstrate and implement quantum algorithms in a classroom or laboratory setting. === Scientific Research === Cloud-based access to quantum hardware has enabled researchers to conduct experiments in quantum information, test quantum algorithms, and compare quantum hardware platforms. Experiments such as testing Bell's theorem or evaluating quantum teleportation protocols have been performed on publicly available quantum processors. === Software Development and Prototyping === Developers use cloud-based platforms to prototype quantum software applications across fields such as optimization, machine learning, and chemistry. These platforms offer SDKs and APIs that integrate classical and quantum workflows, enabling experimentation with quantum algorithms in real-world or simulated environments. === Public Engagement and Games === Quantum cloud tools have also been used to create educational games and interactive applications aimed at increasing public understanding of quantum concepts. These efforts help bridge the gap between theoretical content and intuitive learning. == Existing platforms == qBraid Lab by qBraid is a cloud-based platform for quantum computing. It provides software tools for researchers and developers in quantum, as well as access to quantum hardware. qBraid provides cloud based access to Microsoft Azure Quantum and Amazon Braket devices including IQM, QuEra, Pasqal, Rigetti, IonQ, QIR simulators, Amazon Braket simulators, and the NEC Vector Annealer, as of August 2025. qBraid's base version is free, where unlimited hardware and simulator access is available with the purchase of credits. Quandela Cloud by Quandela is the platform to access first cloud-accessible European photonic quantum computer. The computer is interfaced using the Perceval scripting language, with tutorials and documentation available online for free. Xanadu Quantum Cloud by Xanadu is a platform with cloud-based access to three fully programmable photonic quantum computers. Forest by Rigetti Computing is a tool suite for cloud-based quantum computing. It includes a programming language, development tools and example algorithms. LIQUi> by Microsoft is a software architecture and tool suite for quantum computing. It includes a programming language, example optimization and scheduling algorithms, and quantum simulators. Q#, a quantum programming language by Microsoft on the .NET Framework seen as a successor to LIQUi|>. IBM Quantum Platform by IBM, providing access to quantum hardware as well as HPC simulators. These can be accessed programmatically using the Python-based Qiskit framework, or via graphical interface with the IBM Q Experience GUI. Both are based on the OpenQASM standard for representing quantum operations. There is also a tutorial and online community. Quantum in the Cloud by The University of Bristol, which consists of a quantum simulator and a four qubit optical quantum system. Quantum Playground by Google is an educational resource which features a simulator with a simple interface, and a scripting language and 3D quantum state visualization. Quantum in the Cloud is an experimental quantum cloud platform for access to a four-qubit nuclear magnetic resonance-NMRCloudQ computer, managed by Tsinghua University. Quantum Inspire by Qutech is the first platform in Europe providing cloud-based quantum computing to two hardware chips. Next to a 5-qubit transmon processor, Quantum Inspire is the first platform in the world to provide online access to a fully programmable 2-qubit electron spin quantum processor. Amazon Braket is a cloud-based quantum computing platform hosted by AWS which, as of June 2025, provides access to quantum computers built by IonQ, Rigetti, IQM, and QuEra. Braket also provides a quantum algorithm development environment and simulator. Forge by QC Ware is a cloud-based quantum computing platform that provides access to D-Wave hardware, as well as Google and IBM simulators. The platform offers a 30-day free trial, including one minute of quantum computing time. Quantum-as-a-Service by Scaleway is a cloud-based platform created in 2022 to access to real quantum hardware from IQM Quantum Computers, Alpine Quantum Technologies, Quandela and Pasqal. It also include access to GPU-powered emulators such as Aer, Qsim and Quandela proprietary emulation.
Google Cloud Dataflow
Google Cloud Dataflow is a fully managed service for executing Apache Beam pipelines within the Google Cloud Platform ecosystem. Dataflow provides a fully managed service for executing Apache Beam pipelines, offering features like autoscaling, dynamic work rebalancing, and a managed execution environment. Dataflow is suitable for large-scale, continuous data processing jobs, and is one of the major components of Google's big data architecture on the Google Cloud Platform. At its core, Dataflow's architecture is designed to abstract away infrastructure management, allowing developers to focus purely on the logic of their data processing tasks. When a pipeline written using the Apache Beam SDK is submitted, Dataflow translates this high-level definition into an optimized job graph. The service then provisions and manages a fleet of Google Compute Engine workers to execute this graph in a highly parallelized and fault-tolerant manner. This serverless approach, combined with intelligent autoscaling of both the number of workers (horizontal) and the resources per worker (vertical), ensures that jobs have the precise amount of computational power needed at any given time, optimizing both performance and cost. The service's deep integration with the Google Cloud ecosystem makes it a powerful tool for a variety of use cases beyond simple data movement. For real-time analytics, Dataflow can ingest unbounded streams of data from Cloud Pub/Sub, perform complex transformations, and load results into BigQuery for immediate querying. In machine learning workflows, it is commonly used to preprocess and transform massive datasets stored in Cloud Storage, preparing them for training models in Vertex AI. This versatility makes it the central processing engine for modern ETL (Extract, Transform, Load) operations, streaming analytics, and large-scale data preparation within the cloud. == History == Google Cloud Dataflow was announced in June, 2014 and released to the general public as an open beta in April, 2015. In January, 2016 Google donated the underlying SDK, the implementation of a local runner, and a set of IOs (data connectors) to access Google Cloud Platform data services to the Apache Software Foundation. The donated code formed the original basis for Apache Beam. In August 2022, there was an incident where user timers were broken for certain Dataflow streaming pipelines in multiple regions, which was later resolved. Throughout 2023 and 2024, there have been various other updates and incidents affecting Google Cloud Dataflow, as documented in the release notes and service health history. The donation of the Dataflow SDK to the Apache Software Foundation was a pivotal moment, establishing Apache Beam as a unified, open-source programming model for defining both batch and streaming data pipelines. This strategic move decoupled the pipeline definition from the execution engine. As a result, developers could write portable data processing logic that was not locked into Google's ecosystem. A Beam pipeline can be executed on various runners, including Apache Flink, Apache Spark, and, of course, the highly optimized Google Cloud Dataflow service, providing flexibility and future-proofing data processing investments. == Features == Google Cloud Dataflow supports both batch and streaming data processing pipelines. It automatically handles resource provisioning, data sharding, and scaling according to workload, reducing manual configuration needed for large-scale data operations. == Use cases == Dataflow is used for ETL (Extract, Transform, Load) data pipelines, real-time analytics, and event stream processing for companies in industries such as finance, advertising, and IoT.
Geo-replication
Geo-replication systems are designed to provide improved availability and disaster tolerance by using geographically distributed data centers. This is intended to improve the response time for applications such as web portals. Geo-replication can be achieved using software, hardware or a combination of the two. == Software == Geo-replication software is a network performance-enhancing technology that is designed to provide improved access to portal or intranet content for users at the most remote parts of large organizations. It is based on the principle of storing complete replicas of portal content on local servers, and then keeping the content on those servers up-to-date using heavily compressed data updates. === Portal acceleration === Geo-replication technologies are used to provide replication of the content of portals, intranets, web applications, content and data between servers, across wide area networks WAN to allow users at remote sites to access central content at LAN speeds. Geo-replication software can improve the performance of data networks that suffer limited bandwidth, latency and periodic disconnection. Terabytes of data can be replicated over a wide area network, giving remote sites faster access to web applications. Geo-replication software uses a combination of data compression and content caching technologies. differencing technologies can also be employed to reduce the volume of data that has to be transmitted to keep portal content accurate across all servers. This update compression can reduce the load that portal traffic places on networks, and improve the response time of a portal. === Portal replication === Remote users of web portals and collaboration environments will frequently experience network bandwidth and latency problems which will slow down their experience of opening and closing files, and otherwise interacting with the portal. Geo-replication technology is deployed to accelerate the remote end user portal performance to be equivalent to that experienced by users locally accessing the portal in the central office. === Differencing engine technologies === To deliver this reduction in the size of the required data updates across a portal, geo-replication systems often use differencing engine technologies. These systems are able to difference the content of each portal server right down to the byte level. This knowledge of the content that is already on each server enables the system to rebuild any changes to the content on one server, across each of the other servers in the deployment from content already hosted on those other servers. This type of differencing system ensures that no content, at the byte level, is ever sent to a server twice. === Offline portal replication on laptops === Geo-replication systems are often extended to deliver local replication beyond the server and down to the laptop used by a single user. Server to laptop replication enables mobile users to have access to a local replica of their business portal on a standard laptop. This technology may be employed to provide in the field access to portal content by, for example, sales forces and combat forces. == Geo-replication systems ==
PagedAttention
PagedAttention is an attention algorithm for efficient serving of large language models (LLMs). It was introduced in 2023 by Woosuk Kwon and colleagues in the paper Efficient Memory Management for Large Language Model Serving with PagedAttention, alongside the vLLM serving engine. The method stores the key–value cache used during autoregressive decoding in fixed-size blocks that can be mapped to non-contiguous physical memory, borrowing ideas from virtual memory, paging, and operating system design. == Background == In transformer inference, the key–value cache grows with sequence length and the number of concurrent requests. Kwon et al. argued that earlier serving systems typically reserved contiguous cache regions in advance, which caused reserved space, internal fragmentation, and external fragmentation. In their experiments, the paper reported that the effective memory utilization of previous systems could fall as low as 20.4%. == Description == PagedAttention partitions the cache of each sequence into fixed-size KV blocks. A request's cache is represented as a sequence of logical blocks, while a block table maps those logical blocks to physical GPU-memory blocks. As a result, neighboring logical blocks do not need to be contiguous in physical memory, and new blocks can be allocated on demand as generation proceeds. The design also makes it easier to share cache state across related decoding paths. In vLLM, physical blocks can be reference-counted and shared among requests or branches, with block-granularity copy-on-write used when a shared block must be modified. The original paper applied this design to parallel sampling, beam search, and prompts with shared prefixes. == Mathematical formulation == For a query token i {\displaystyle i} in causal self-attention, the standard attention output can be written as a i j = exp ( q i ⊤ k j / d ) ∑ t = 1 i exp ( q i ⊤ k t / d ) , o i = ∑ j = 1 i a i j v j {\displaystyle a_{ij}={\frac {\exp(\mathbf {q} _{i}^{\top }\mathbf {k} _{j}/{\sqrt {d}})}{\sum _{t=1}^{i}\exp(\mathbf {q} _{i}^{\top }\mathbf {k} _{t}/{\sqrt {d}})}},\;\mathbf {o} _{i}=\sum _{j=1}^{i}a_{ij}\mathbf {v} _{j}} where q i {\displaystyle \mathbf {q} _{i}} , k j {\displaystyle \mathbf {k} _{j}} , and v j {\displaystyle \mathbf {v} _{j}} are the query, key, and value vectors, and d {\displaystyle d} is the attention dimension. If the cache is partitioned into blocks of size B {\displaystyle B} , the key and value blocks may be written as K j = ( k ( j − 1 ) B + 1 , … , k j B ) , V j = ( v ( j − 1 ) B + 1 , … , v j B ) {\displaystyle \mathbf {K} _{j}=(\mathbf {k} _{(j-1)B+1},\ldots ,\mathbf {k} _{jB}),\;\mathbf {V} _{j}=(\mathbf {v} _{(j-1)B+1},\ldots ,\mathbf {v} _{jB})} PagedAttention then performs the computation blockwise: A i j = exp ( q i ⊤ K j / d ) ∑ t = 1 ⌈ i / B ⌉ exp ( q i ⊤ K t / d ) , o i = ∑ j = 1 ⌈ i / B ⌉ V j A i j ⊤ {\displaystyle \mathbf {A} _{ij}={\frac {\exp(\mathbf {q} _{i}^{\top }\mathbf {K} _{j}/{\sqrt {d}})}{\sum _{t=1}^{\lceil i/B\rceil }\exp(\mathbf {q} _{i}^{\top }\mathbf {K} _{t}/{\sqrt {d}})}},\;\mathbf {o} _{i}=\sum _{j=1}^{\lceil i/B\rceil }\mathbf {V} _{j}\mathbf {A} _{ij}^{\top }} where A i j {\displaystyle \mathbf {A} _{ij}} is the vector of attention scores for the j {\displaystyle j} -th KV block. In the formulation given by Kwon et al., this preserves the causal attention calculation while allowing the key and value blocks to reside in non-contiguous physical memory. == Performance and use == The vLLM paper reported that, on its evaluated workloads, the use of PagedAttention and the associated memory-management design improved serving throughput by 2–4× over the compared baselines, including FasterTransformer and Orca, while preserving model outputs. In experiments on OPT-13B with the Alpaca trace, the paper also reported memory savings of 6.1–9.8% for parallel sampling and 37.6–55.2% for beam search through KV-block sharing. A 2024 survey of LLM serving systems described PagedAttention as having become an industry norm in LLM serving frameworks, citing support in TGI, vLLM, and TensorRT-LLM. == Limitations and alternatives == Subsequent work has described trade-offs in the approach. The 2025 vAttention paper argued that PagedAttention requires attention kernels to be rewritten to support paging and increases software complexity, portability issues, redundancy, and execution overhead, proposing instead a memory manager that keeps the cache contiguous in virtual memory while relying on demand paging for physical allocation. === vAttention === Unlike PagedAttention, vAttention does not introduce a different attention rule; it retains the standard attention computation Attention ( q i , K , V ) = softmax ( q i K ⊤ s c a l e ) V . {\displaystyle \operatorname {Attention} (q_{i},K,V)=\operatorname {softmax} \left({\frac {q_{i}K^{\top }}{\mathrm {scale} }}\right)V.} In the notation of Prabhu et al., the key and value tensors for a request seen so far are K , V ∈ R L ′ × ( H × D ) {\displaystyle K,V\in \mathbb {R} ^{L'\times (H\times D)}} , where L ′ {\displaystyle L'} is the context length seen so far, H {\displaystyle H} is the number of KV heads on a worker, and D {\displaystyle D} is the dimension of each KV head. In systems prior to PagedAttention, the K cache (or V cache) at each layer of a worker is typically allocated as a 4D tensor of shape [ B , L , H , D ] , {\displaystyle [B,L,H,D],} where B {\displaystyle B} is batch size and L {\displaystyle L} is the maximum context length supported by the model. vAttention preserves this contiguous virtual-memory view while deferring physical-memory allocation to runtime. A serving framework maintains separate K and V tensors for each layer, so vAttention reserves 2 N {\displaystyle 2N} virtual-memory buffers on a worker, where N {\displaystyle N} is the number of layers managed by that worker. The maximum size of one virtual-memory buffer is B S = B × S , {\displaystyle BS=B\times S,} where S {\displaystyle S} is the maximum size of a single request's per-layer K cache (or V cache) on a worker. The paper defines S = L × H × D × P , {\displaystyle S=L\times H\times D\times P,} where P {\displaystyle P} is the number of bytes needed to store one element. In this formulation, vAttention keeps the KV cache contiguous in virtual memory and relies on demand paging for physical allocation, rather than modifying the attention kernel to operate over non-contiguous KV-cache blocks.
Showcase Workshop
Showcase Workshop, also referred to as Showcase, is a SaaS company that develops a presentation-building application for business use. Users upload files and images to a web platform which generates presentations viewable on a suite of mobile apps. Showcase was founded in 2011. The company’s headquarters are in Wellington, New Zealand. == History == Showcase Workshop was originally developed in response to dynamically changing content being presented on iPads at the 2012 Olympics. After market-testing a beta version of the core application, Showcase Workshop launched commercially in 2012. In 2014 Showcase partnered with Vodafone Global Enterprise. == Product == Users upload pre-existing PDFs, videos, images and Microsoft Office documents to a secure server, building presentations or ‘showcases’ which can then be downloaded via the mobile apps. The presentations are used for mobile sales enablement, training, or operational/health and safety purposes. == Reception == Reviewers have praised the ease of use of Showcase, calling it a “better alternative to developing a native app” and “intuitive”. Criticisms include the lack of differing templates and a lack of complex customisation controls. Showcase was nominated for a Tabby Award in 2014 and won a Tabby Award in 2015 for its Windows app.