Dynatrace

SEPTEMBER 30, 2021

Like any move, a cloud migration requires a lot of planning and preparation, but it also has the potential to transform the scope, scale, and efficiency of how you deliver value to your customers. This can fundamentally transform how they work, make processes more efficient, and improve the overall customer experience. Here are three.

Cloud 163

Cloud Traffic Best Practices Strategy 163

ACM Sigarch

APRIL 26, 2018

BPU: Brain Processing Unit is the design of the AI chips by Horizon Robotics. With its modular architecture, DLA is scalable, highly configurable, and designed to simplify integration and portability. HPU: Holographic Processing Unit (HPU) is the specific hardware of Microsoft’s Hololens. FPU: Floating Processing Unit (FPU).

Hardware 99

Hardware Artificial Intelligence Architecture Google 99

Percona

JUNE 22, 2023

Defining high availability In general terms, high availability refers to the continuous operation of a system with little to no interruption to end users in the event of hardware or software failures, power outages, or other disruptions. In addition to the disaster recovery site, this design includes an external layer of nodes.

Availability 110

Availability Database Open Source Hardware 110

Percona

OCTOBER 18, 2023

Compatibility Open source databases are designed with interoperability at their core, adhering to industry standards and offering support for a wide range of data import and export formats. Does anyone on my team require further training before we start? Look closely at your current infrastructure (hardware, storage, networks, etc.)

Open Source 94

Open Source Database Hardware Strategy 94

The Morning Paper

JUNE 11, 2019

The goal here is to reduce the training times of DNNs by finding efficient parallel execution strategies, and even including its search time, FlexFlow is able to increase training throughput by up to 3.3x FlexFlow is also given a device topology graph describing all the available hardware devices and their interconnections.

Network 81

Network Strategy Hardware Benchmarking 81

Sutter's Mill

NOVEMBER 11, 2023

This week’s meeting: Meeting #2 of C++26 At the previous meeting in June, the committee adopted the first 40 proposed changes for C++26, including many that had been ready for a couple of meetings and were just waiting for the C++26 train to open to be adopted. For those highlights, see the previous trip report. This is that library.

C++ 107

C++ Lambda Network Programming 107

The Morning Paper

JUNE 13, 2019

As we saw with the SOAP paper last time out, even with a fixed model variant and hardware there are a lot of different ways to map a training workload over the available hardware. Different hardware architectures (CPUs, GPUs, TPUs, FPGAs, ASICs, …) offer different performance and cost trade-offs.

Hardware 58

Hardware Latency Serverless Energy 58

The Morning Paper

MAY 28, 2019

The goal is to produce a low-energy hardware classifier for embedded applications doing local processing of sensor data. Race logic has four primary operations that are easy to implement in hardware: MAX, MIN, ADD-CONSTANT, and INHIBIT. One efficient way of doing that in analog hardware is the use of current-starved inverters.

Energy 52

Energy Hardware Efficiency Architecture 52

Smashing Magazine

OCTOBER 11, 2022

All industries have a part to play, and that includes web design and development. For example, changing our database architecture to be more efficient could save on server costs. On design systems, UX, web performance and CSS/JS. Tom Greenwood’s book, Sustainable Web Design is packed with advice for building low-carbon websites.

Strategy 123

Strategy Development Energy Website 123

The Morning Paper

MAY 14, 2019

Finally, we show that Seer can identify application level design bugs, and provide insights on how to better architect microservices to achieve predictable performance. A DNN model is trained to recognise patterns in space and time that lead to QoS violations. When available, it can use hardware level performance counters.

Big Data 45

Big Data Cloud Performance Hardware 45

The Morning Paper

JUNE 27, 2019

The thrust of the argument is that there’s a chain of inter-linked assumptions / dependencies from the hardware all the way to the programming model, and any time you step outside of the mainstream it’s sufficiently hard to get acceptable performance that researchers are discouraged from doing so. Challenges compiling non-standard kernels.

Systems 87

Systems Tuning Programming Innovation 87

The Morning Paper

SEPTEMBER 26, 2019

no longer hold: Security was never considered by design as historically the ability to transmit and receive wireless signals required considerable resources and knowledge. …pilots are trained and instructed to trust the instruments more than their intuition. If the instruments ask them to fly right, the pilots will fly right.

Wireless 91

Wireless Systems Airlines Hardware 91

Alex Russell

APRIL 29, 2021

Education, tools, and training all support the premise that supporting more browsers is better ( ceteris paribus ), creating a substantial incentive to grease squeaky wheels. Efficiently enables new styles of drawing content on the web , removing many hard tradeoffs between visual richness , accessibility, and performance.

Media 145

Media Games Education Engineering 145

CSS Wizardry

APRIL 2, 2018

As I currently do listen to almost exclusively digitally streamed music, it was important to design a solution that was appropriate. When I’m on a flight (as I am right now) or a train, this extra bulk doesn’t bother me too much—I get settled in anyway, so I don’t mind the extra room it takes. Requirements. Technics OTTAVA in-situ.

Website 108

Website Media Hardware Monitoring 108

All Things Distributed

MARCH 13, 2017

Digitalization offers almost endless possibilities to communicate faster, work more efficiently, and be more creative – in real-time. Previously, the top priorities for IT departments were equipping data centers with hardware, purchasing software, and further developing proprietary software.

Social Media 84

Social Media Retail Automotive Innovation 84

O'Reilly

NOVEMBER 28, 2023

Training models and developing complex applications on top of those models is becoming easier. Many of the new open source models are much smaller and not as resource intensive but still deliver good results (especially when trained for a specific application). report that the difficulty of training a model is a problem.

Open Source 108

Open Source Artificial Intelligence Google Infrastructure 108

Alex Russell

JUNE 20, 2020

In each quantum of time, hardware and OS vendors press ahead, adding features. As OS and hardware deployed base integrate these components, the set of what most computers can do is expanded. This is often determined by hardware integration and device replacement rates.

Hardware 55

Hardware Games Java Media 55

High Scalability

OCTOBER 24, 2022

There is no structural separation as there is with a train service. Each cloud-native evolution is about using the hardware more efficiently. There's a huge short-term and long-term efficiency of services that depends on the successful coordination of cloud services and infrastructure. For example, AWS created Nitro.

Cloud 187

Cloud Hardware Infrastructure Innovation 187

O'Reilly

SEPTEMBER 19, 2023

A company then needed to train up their ops team to manage the cluster, and their analysts to express their ideas in MapReduce. Doubly so as hardware improved, eating away at the lower end of Hadoop-worthy work. Google goes a step further in offering compute instances with its specialized TPU hardware.

Hardware 99

Hardware Storage Big Data Blockchain 99

Alex Russell

JANUARY 30, 2024

HTML, CSS, images, and fonts can all be parsed and run at near wire speeds on low-end hardware, but JavaScript is at least three times more expensive, byte-for-byte. Engineering is the discipline of designing solutions under specific constraints. Those folks may not find the next couple of years to their liking.

Performance 90

Performance Network Mobile Speed 90

All Things Distributed

MARCH 19, 2018

For example, training on more data means more accurate models. Last re:Invent, to make the problem of authoring, training, and hosting ML models easier, faster, and more reliable, we launched Amazon SageMaker. Machine learning models are usually trained tens or hundreds of times. In machine learning, more is usually more.

Scalability 159

Scalability Hardware AWS Tuning 159

The Netflix TechBlog

DECEMBER 23, 2020

everything from the frontend API for netflix.com, to machine learning training workloads, to video encoders. This abstraction allows the compute team to influence the reliability, efficiency, and operability of the fleet via the scheduler. We do this for reliability, scalability, and efficiency reasons.

Media 286

Media Metrics Systems Processing 286

The Netflix TechBlog

APRIL 29, 2019

Various software systems are needed to design, build, and operate this CDN infrastructure, and a significant number of them are written in Python. Such applications track the inventory of our network gear: what devices, of which models, with which hardware components, located in which sites.

Open Source 279

Open Source Network Infrastructure Big Data 279

O'Reilly

OCTOBER 11, 2022

Andrew Ng , Christopher Ré , and others have pointed out that in the past decade, we’ve made a lot of progress with algorithms and hardware for running AI. Our current set of AI algorithms are good enough, as is our hardware; the hard problems are all about data. Designing the loop needs to be part of the automation plan.

Games 116

Games Healthcare Processing Retail 116

The Netflix TechBlog

FEBRUARY 13, 2020

By Aditya Mavlankar, Jan De C**k¹, Cyril Concolato, Kyle Swanson, Anush Moorthy and Anne Aaron TL; DR We need an alternative to JPEG that a) is widely supported, b) has better compression efficiency and c) has a wider feature set. High-Efficiency Video Coding ( HEVC ) is the successor of H.264, 264, a.k.a.

Code 212

Code Metrics Efficiency Open Source 212

Smashing Magazine

AUGUST 22, 2021

One aspect of performance is about how efficiently a transport protocol can use a network’s full (physical) bandwidth (i.e. While this is conceptually possible (QUIC does exactly that), things were initially not designed like this, because we need to be able to use TCP with and without TLS on top. Congestion Control.

Performance 105

Performance Network Latency Servers 105

High Scalability

NOVEMBER 25, 2019

To help our customers manage ongoing data explosion, we have designed many of the core components over the last few years. Version7 (Future): Move all compute to the public cloud, carve out more services for impact isolation, dynamic resource pooling to manage pets and cattle efficiently. How much upfront design should you do?

Architecture 52

Architecture Cache Azure Storage 52