Tag: computing

El acuerdo de Power Computing con Apple inicia la era de los clones de Mac

Post author By miranda cosgrove
Post date December 16, 2024
No Comments on El acuerdo de Power Computing con Apple inicia la era de los clones de Mac

[ad_1]

16 de diciembre: Evento en la historia de Apple: Apple firma un acuerdo de clonación de Mac con Power Computing 16 de diciembre de 1994: Apple Computer firmó un acuerdo de licencia con Power Computing, lo que permite a la empresa producir computadoras compatibles con Macintosh, también conocidas como “clones de Mac”.

Con su participación de mercado en declive y su antiguo rival Microsoft avanzando gracias a una estrategia de licencias de software, los ejecutivos de Apple creían que la única forma de competir era que la empresa entregara su sistema operativo a Mac de terceros. Por supuesto, no lo parece del todo…

Power Computing y el dilema de las licencias de software de Apple

La posibilidad de que Apple concediera licencias para su tecnología se convirtió en un importante tema de discusión durante las décadas de 1980 y 1990. Como muchos han visto, Apple ha creado un software muy superior al de sus competidores como Microsoft. Sin embargo, Cupertino se quedó atrás al quedarse con el código en lugar de dejar que otros lo usen. De hecho, la única concesión de licencias que Apple otorgó durante la década de 1980 fue la concesión de licencias Acuerdo ridículamente unilateral Esto dio a Microsoft el derecho legal a utilizar elementos de la interfaz de Mac en su sistema operativo Windows, sin apenas recompensa por parte de Apple.

El hecho de que Apple no concediera la licencia de Mac OS significó que, a mediados de la década de 1990, Windows se convirtiera en el estándar. El Mac adolecía de ser un poco jugador.

A pesar de este hecho, en diciembre de 1994, Apple vinculó a Packard Bell con El mayor fabricante del mundo de computadoras personales de escritorio.. Ambas representaban el 12% del mercado, seguidas de Compaq con el 11%, IBM con el 9%, Dell con el 6%, Gateway 2000 con el 5% y Hewlett-Packard con el 3%.

Al igual que ocurre hoy con Android versus iOS, Apple dominó las ventas de hardware mientras que su software propietario siguió siendo una minoría.

El miedo a Windows 95 impulsa la estrategia de software de Apple

En algún momento de 1994, la estrategia de Apple cambió, como lo demuestra el acuerdo con Power Computing. Darse tono Miedo a Windows 95 Si lo desea, Apple de repente comenzó a negociar acuerdos con terceros para licenciar partes de su tecnología.

En diciembre de ese año, Apple firmó un acuerdo con Bandai, el mayor fabricante de juguetes de Japón, para comprar una consola de juegos Mac OS. (Manzana La consola de juegos Pippin fue lanzada en 1996. Y falló inmediatamente.)

Además del acuerdo con Power Computing que se produjo ese día en 1994, Apple también llegó a un acuerdo con Radius, lo que le permitió a esa empresa producir versiones de Mac también. Radius System 100, la primera versión oficial de Macintosh, fue lanzada el 27 de marzo de 1995.

Clonación de Mac: después de todo, no es un buen negocio para Apple

La era del “clon de Mac” que siguió a estos acuerdos de licencia de software Tuve un gran comienzo en el campo de la informática de potencia.. Pero resultó ser un desastre para Apple. En lugar de estimular las ventas de Mac, solo significó “Macs” más baratas. Como resultado, Apple sufrió un gran impacto en la cantidad que ganó por unidad.

Fred Anderson, director financiero de Apple, desarrolló más tarde la estrategia de concesión de licencias. cuesta Fondos de la empresa. La tarifa de 50 dólares que Apple cobró por cada Mac clonada no estuvo cerca de recuperar el dinero perdido de las personas que optaron por comprar Macs de terceros en lugar de los oficiales más caros.

Acuerdo de informática energética Continuó hasta el 5 de agosto de 1997.cuando Apple quedó atrapada en un enfrentamiento con el fabricante del clon de Mac. Apple hizo un poco de jiu-jitsu legal, Presentamos Mac OS 8 Por lo tanto, se argumenta que el acuerdo de licencia a terceros no se extendió más allá del Sistema 7.

Finalmente, Apple llegó a un acuerdo comprando el negocio Mac de Power Computing por 100 millones de dólares. Esto puso fin a la era de la clonación de Macintosh. También allanó el camino para El regreso de Steve Jobs en 1997 para revertir la suerte de Apple.

¿Recuerdas la era de los clones de Mac? Si es así, ¿tiene buenos o no buenos recuerdos de Mac de terceros fabricados por Power Computing y otros en ese momento? Deje sus otros pensamientos y recuerdos a continuación.

[ad_2]

Source Article Link

Tags acuerdo, Apple, clones, computing, Con, era, inicia, Los, Mac, Power

Bisnis Industri

Power Computing, el fabricante de clones de Mac, alcanza su punto máximo: hoy en la historia de Apple

Post author By miranda cosgrove
Post date June 6, 2024
No Comments on Power Computing, el fabricante de clones de Mac, alcanza su punto máximo: hoy en la historia de Apple

[ad_1]

4 de junio: Hoy en la historia de Apple: la potencia informática de Mac ha alcanzado su punto máximo y ha comenzado un rápido declive. 4 de junio de 1997: Power Computing, que creó una versión de Mac, ha alcanzado un máximo histórico, pero también es el principio del fin.

El CEO de la compañía ha llegado a un acuerdo con el CEO de Apple, Gil Amelio, sobre el próximo proyecto Mac OS 8. El acuerdo permite a la empresa comenzar a avanzar hacia una IPO como la empresa informática de más rápido crecimiento de la década. Pero las cosas no van bien.

El auge de la informática potente y sus clones para Mac

Power Computing se fundó en noviembre de 1993 y comenzó a vender copias de Mac por correo. Con el negocio de pedidos por correo de PC de Dell estableciéndose en el mercado, Stephen Kang, fundador de Power Computing, predijo correctamente que la versión para Mac haría lo mismo.

La empresa inició conversaciones con Apple en abril de 1994. A finales de ese año, Power Computing ha cerrado un acuerdo con Cupertino. (Próximamente, otro creador de clones de Mac, radioseguirá los pasos de Power Computing).

¿Por qué ofertas para permitir la clonación de Mac? Confrontación Grandes desafíos como el inminente Windows 95Apple empezó a tener dificultades en 1994. De hecho, la empresa estuvo a punto de entrar en la peor etapa de su historia. Como resultado, un Cupertino cada vez más desesperado intentó cosas que nunca antes habría disfrutado: Como hacer su propia consola de videojuegos..

Los Mac clonados son un éxito

Power Computing ha lanzado varias versiones de Macintosh. Entre ellos se encontraban el PowerTower y el PowerTower Pro de 1997, que demostraron ser impresionantemente rápidos en su época.

En 1997, al fabricante de clones de Mac le iba muy bien. El acuerdo del 4 de junio con Amelio significó que más tarde ese mes, la empresa comenzó a cotizar en bolsa en serio. (La industria tecnológica se apresuró a adoptar este tipo de estrategia, lo que contribuyó a provocar el colapso de las puntocom poco después).

como New York Times masculino En ese momento, Power Computing estaba en camino de generar 700 millones de dólares en ingresos al año. La empresa “acababa de acordar comprar 150 acres en la cercana Georgetown para construir una nueva sede por valor de 28 millones de dólares”.

Y luego caer

Sin embargo, la suerte del fabricante de clones cambió repentinamente tras un cambio en la dirección de Apple. Para julio, Amelio quedó fuera del puesto de CEO Y cofundador de Apple Steve Jobs Regresó como director ejecutivo interino. Y a Jobs no le gustó el clon de Mac.

por 5 de agosto de 1997Apple está en un punto muerto con Power Computing. El 2 de septiembre, Apple acordó adquirir la lista de clientes del fotocopiador y la licencia de Mac OS a cambio de 100 millones de dólares en acciones de AAPL y 10 millones de dólares para cubrir todas las deudas y costos pendientes. A principios de 1998, Power Computing cerró definitivamente.

¡Habla de un torbellino de altibajos!

¿Recuerdas la era de los clones de Mac? ¿Quizás tengas uno? Háganos saber sus recuerdos en los comentarios a continuación.

[ad_2]

Source Article Link

Tags alcanza, Apple, clones, computing, fabricante, historia, hoy, Mac, máximo, Power, punto

Featured

Dell cracks down on hybrid working again — computing giant is going to start color-coding employees to show who is coming back to the office

[ad_1]

Dell has issued a nineteen eighty-four-esque warning to its hybrid working employees, who shall from now on be tracked using electronic badge swipes, VPN usage and a color-coding system, in order for the company to keep tabs on office usage.

A source familiar with the matter told The Register the new initiative, spearheaded by COO Jeff Clarke, is designed to monitor the on-site presence of the company’s hybrid workers.

The system will be used to grade employees based on their presence in the office for at least 39 days per quarter, which equates to around three days per week.

Dell vs hybrid working

Starting May 13, Dell intends to provide employees with weekly updates on their onside presence through its HR software, with reports accompanied by a color-coded rating. The ratings will include green, yellow and red flags for ‘regular,’ ‘some’ and ‘limited’ on-site presence, with a special blue flag awarded to those with ‘consistent’ on-site presence.

A Dell spokesperson commented (via The Register): “In today’s global technology revolution, we believe in-person connections paired with a flexible approach are critical to drive innovation and value differentiation.”

However, workers have expressed dissatisfaction with the company’s increasingly strict measures. The news comes just weeks after the company announced it would be penalizing workers for working from home too much, with promotions being threatened.

Other workers have linked the increased surveillance with continued headcount reduction and cost optimization efforts. Dell laid off around 6,000 workers in March 2024, and a further 6,650 in February 2023. Both times, the company saw a 5% reduction in its workforce.

Intel quietly launched mysterious new AI CPU that promises to bring deep learning inference and computing to the edge — but you won’t be able to plug them in a motherboard anytime soon

[ad_1]

Intel has launched a new AI processor series for the edge, promising industrial-class deep learning inference. The new ‘Amston Lake’ Atom x7000RE chips offer up to double the cores and twice the higher graphics base frequency as the previous x6000RE series, all neatly packed within a 6W–12W BGA package.

The x7000RE series packs more performance into a smaller footprint. Boasting up to eight E-cores it supports LPDDR5/DDR5/DDR4 memory and up to nine PCIe 3.0 lanes, delivering robust multitasking capabilities.

Intel says its new processors are designed to withstand challenging conditions, enduring extreme temperature variations, shock, and vibration, and to operate in hard-to-reach locations. They offer 2x SATA Gen 3.2 ports, up to 4x USB 3.2 Gen 2 ports, a USB Type-C port, 2.5GbE Ethernet connection, along with Intel Wi-Fi, Bluetooth, and 5G platform capabilities.

Embedded, industrial, and communication

The x7000RE series consists of four SKUs, all suitable for embedded, industrial, and communication use under extended temperature conditions. The x7211RE and x7213RE have 2 cores and relatively lower base frequencies, while the x7433RE has 4 cores, and the x7835RE has 8 cores with higher base frequencies.

All four SKUs support a GPU execution unit count of either 16 or 32, and Intel’s Time Coordinated Computing and Time-Sensitive Networking GbE features. The x7000RE offer integrated Intel UHD Graphics, Intel DL Boost, Intel AVX2 with INT8 support, and OpenVINO toolkit support.

Intel says the chips will allow customers to easily deploy deep learning inference at the industrial edge and in smart cities, and “enhance computer vision solutions with built-in AI capabilities and ecosystem-enabled camera modules” as well as “capture power- and cost-efficient performance to enable latency-bounded workloads in robotics and automation.”

Chip firm founded by ex-Intel president plans massive 256-core CPU to surf AI inference wave and give Nvidia B100 a run for its money — Ampere Computing AmpereOne-3 likely to support PCIe 6.0 and DDR5 tech

[ad_1]

Ampere Computing unveiled its AmpereOne Family of processors last year, boasting up to 192 single-threaded Ampere cores, which was the highest in the industry.

These chips, designed for cloud efficiency and performance, were Ampere’s first product based on its new custom core leveraging internal IP, signalling a shift in the sector, according to CEO Renée James.

At the time of the launch, James said, “Every few decades of compute there has emerged a driving application or use of performance that sets a new bar of what is required of performance. The current driving uses are AI and connected everything combined with our continued use and desire for streaming media. We cannot continue to use power as a proxy for performance in the data center. At Ampere, we design our products to maximize performance at a sustainable power, so we can continue to drive the future of the industry.”

AmpereOne-3 on its way

Jeff Wittich, chief product officer at Ampere, recently spoke with The Next Platform about future generations of AmpereOne. He told the site that an updated chip, with 12 memory channels and an A2 core with improved performance, would be out later this year in keeping with the company’s roadmap. This chip, which The Next Platform calls AmpereOne-2, will reportedly have a 33 percent increase in DDR5 memory controllers and up to 50 percent more memory bandwidth.

However, what’s coming up beyond that, at some point in 2025, sounds the most exciting.

The Next Platform says the third generation chip, AmpereOne-3 as it is calling it, will have 256 cores and be “etched in 3 nanometer (3N to be precise) processes from TSMC”. It will use a modified A2+ core with a “two-chiplet design on the cores, with 128 cores per chiplet. It could be a four-chiplet design with 64 cores per chiplet.”

The site expects the AmpereOne-3 will support PCI-Express 6.0 I/O controllers and maybe have a dozen DDR5 memory controllers, although there’s some speculation here.

“We have been moving pretty fast on the on the compute side,” Wittich told the site. “This design has got about a lot of other cloud features in it – things around performance management to get the most out of all of those cores. In each of the chip releases, we are going to be making what would generally be considered generational changes in the CPU core. We are adding a lot in every single generation. So you are going to see more performance, a lot more efficiency, a lot more features like security enhancements, which all happen at the microarchitecture level. But we have done a lot to ensure that you get great performance consistency across all of the AmpereOnes. We are also taking a chiplet approach with this 256-core design, which is another step as well. Chiplets are a pretty big part of our overall strategy.”

The AmpereOne-3 is reportedly being etched at TSMC right now, prior to its launch next year.

‘The socket is becoming the motherboard’ — how a Samsung-backed startup wants to change computing forever by integrating everything in an elegant packaging to tackle the AI beast once and for all

[ad_1]

Many system designers are exploring chiplet-based SiPs that move beyond the limitations and costs of huge single-die implementations, but which depend heavily on silicon interposers as substrates for mounting and interconnecting the dies.

Silicon interposers deliver a higher data rate than organic substrates, but are not without drawbacks. They are costly, proprietary and limit the number of chiplets that can be placed on one substrate due to size restrictions, while increasing TCO.

Many CPUs are multi-die assemblies which make use of organic substrates (implemented using chiplet interconnect standards like UCIe and the Open Compute Project’s Bunch of Wires (BoW)), but they can’t compete performance-wise with silicon interposers.

Build a better ‘Blackwell’

Now, Eliyan has come up with what it believes is a viable solution that offers the performance advantages of silicon interposers but without their limitations. The solution to this dilemma, the firm says, is to look at the dies themselves – “at the tiny electronic circuits that drive the interconnect lines”. Most of today’s chiplet interconnect standards use a physical-layer (PHY) IP block, but Eliyan’s NuLink PHY reportedly reaches the same maximum performance levels on organic substrates that alternative PHYs can only achieve on silicon interposers.

The company says NuLink PHY, “enables systems with higher performance (more memory) and lower TCO (no interposer),” with results that deliver up to 4x the bandwidth, 4x the power efficiency, up to 4x the SiP size and up to 10x the AI performance.

Eliyan’s Co-founder and chief executive officer Ramin Farjadrad recently talked to The Next Platform about how NuLink PHY can be used to “build better, cheaper, and more powerful compute engines than can be done with current packaging techniques based on silicon interposers.”

The article, titled “How To Build A Better ‘Blackwell’ GPU Than Nvidia Did”, explains how its technology could be used to significantly improve on Nvidia’s superchip design.

It’s a fascinating read, and hints at what could be possible in the future. As The Next Platform’s Timothy Prickett Morgan sums up, “Any memory, any co-packaged optics, any PCI-Express or other controller, can be linked using NuLink to any XPU. At this point, the socket really has become the motherboard.”

There’s certainly a lot of interest in Eliyan’s NuLink PHY. The company recently closed a $60 Million Series B funding round, co-led by Samsung Catalyst Fund and Tiger Global Management. “This investment reflects the confidence in our approach to integrating multi-chip architectures that address the critical challenges of high costs, low yield, power consumption, manufacturing complexity, and size limitations,” Farjadrad said, following the funding announcement.

Apple-1 starts a computing revolution: Today in Apple history

Post author By miranda cosgrove
Post date April 11, 2024
No Comments on Apple-1 starts a computing revolution: Today in Apple history

[ad_1]

April 11: Today in Apple history: Apple-1 launches. It's the First Apple computer. April 11, 1976: Apple releases its first computer, the Apple-1.

Designed and hand-built by Steve Wozniak, the computers are sold wholesale by “Steven” Jobs. To finance their manufacturing, Wozniak sells his HP-65 calculator for $500, while Jobs sells his Volkswagen van. Years later, in 2014, a working Apple-1 will sell at auction for $905,000.

Apple-1: The first Apple computer

In terms of specs, the first Apple computer was incredibly primitive. It came with an 8-bit MOS 6502 microprocessor running at 1 MHz. It boasted 4KB of memory as standard, although expansion cards could boost this to 8KB or 48KB. Users had to add their own keyboard and monitor, although the latter could be a regular TV set, which made the Apple-1 innovative for its day. (And also, arguably, makes the Apple-1 the company’s first set-top box.)

Wozniak started working on the computer as a hobby, with no goal beyond showing it off to the people at the local Homebrew Computer Club, a hobbyist group whose meetings he attended in Menlo Park, California.

A proof of concept for the Homebrew Computer Club

“I did this computer … to show the people at Homebrew that it was possible to build a very affordable computer — a real computer you could program for the price of the Altair — with just a few chips,” Wozniak recalled in his autobiography, iWoz.

Jobs convinced Woz they would do better building and selling the Apple-1 rather than giving away the designs. Then Jobs approached Paul Terrell, who owned The Byte Shop in nearby Mountain View, California, one of the first personal computer stores.

Terrell rejected Jobs’ first suggestion that the Apple-1 should come in kit form. He told Jobs that, with computers becoming more mainstream, people wanted to buy fully assembled machines.

Apple-1 launch price: $666.66

Jobs listened and agreed. So, Terrell said he would buy 50 Apple-1 computers for $500 each, although cash would only be paid upon delivery. Terrell then marked up the computers to $666.66, or the equivalent of more than $3,500 today.

Ultimately, the Apple-1 didn’t hang around too long. Apple only built 200 or so of its first computers. The number of surviving units is significantly smaller today, due to both the computers’ age and the fact that Apple offered a trade-in deal when it launched the significantly upgraded Apple II the following year.

[ad_2]

Source Article Link

Tags Apple, Apple1, computing, History, Revolution, starts, Today

Featured

Microsoft makes major quantum computing breakthrough — development of most stable qubits might actually make the technology viable for many, but will anyone be able to afford it?

[ad_1]

Unlike traditional computing that uses binary bits, quantum computing uses quantum bits or ‘qubits’, enabling simultaneous processing of vast amounts of data, potentially solving complex problems much faster than conventional computers.

In a major step forward for quantum computing, Microsoft and Quantinuum have unveiled the most reliable logical qubits to date, boasting an error rate 800 times lower than physical qubits.

This groundbreaking achievement involved running over 14,000 individual experiments without a single error, which could make quantum computing a viable technology for various industries.

Azure Quantum Elements platform

Microsoft says the successful demonstration was made possible by applying its innovative qubit-virtualization system (coupled with error diagnostics and correction) to Quantinuum’s ion-trap hardware. Jason Zander, EVP of Strategic Missions and Technologies at Microsoft, says, “This finally moves us out of the current noisy intermediate-scale quantum (NISQ) level to Level 2 Resilient quantum computing.”

The potential of this advancement is enormous. As Zander says, “With a hybrid supercomputer powered by 100 reliable logical qubits, organizations would start to see the scientific advantage, while scaling closer to 1,000 reliable logical qubits would unlock commercial advantage.”

Quantum computing holds enormous promise for solving some of society’s most daunting challenges, including climate change, food shortages, and the energy crisis. These issues often boil down to complex chemistry and materials science problems, which classical computing struggles to handle but which would be far easier for Quantum computers to manage.

The task now, Microsoft says, is to continue improving the fidelity of qubits and enable fault-tolerant quantum computing. This will involve transitioning to reliable logical qubits, a feat achieved by merging multiple physical qubits to protect against noise and sustain resilient computation.

While the technology’s potential is immense, its widespread adoption will depend on its accessibility and cost-effectiveness. For now, though, Microsoft and Quantinuum’s breakthrough marks a significant step towards making quantum computing a practical reality.

Microsoft may have finally made quantum computing useful

Post author By lisa nichols
Post date April 3, 2024
No Comments on Microsoft may have finally made quantum computing useful

[ad_1]

The dream of quantum computing has always been exciting: What if we could build a machine working at the quantum level that could tackle complex calculations exponentially faster than a computer limited by classical physics? But despite seeing IBM, Google and others announce iterative quantum computing hardware, they’re still not being used for any practical purposes. That might change with today’s announcement from Microsoft and Quantinuum, who say they’ve developed the most error-free quantum computing system yet.

While classical computers and electronics rely on binary bits as their basic unit of information (they can be either on or off), quantum computers work with qubits, which can exist in a superposition of two states at the same time. The trouble with qubits is that they’re prone to error, which is the main reason today’s quantum computers (known as Noisy Intermediate Scale Quantum [NISQ] computers) are just used for research and experimentation.

Microsoft’s solution was to group physical qubits into virtual qubits, which allows it to apply error diagnostics and correction without destroying them, and run it all over Quantinuum’s hardware. The result was an error rate that was 800 times better than relying on physical qubits alone. Microsoft claims it was able to run more than 14,000 experiments without any errors.

According to Jason Zander, EVP of Microsoft’s Strategic Missions and Technologies division, this achievement could finally bring us to “Level 2 Resilient” quantum computing, which would be reliable enough for practical applications.

“The task at hand for the entire quantum ecosystem is to increase the fidelity of qubits and enable fault-tolerant quantum computing so that we can use a quantum machine to unlock solutions to previously intractable problems,” Zander, wrote in a blog post today. “In short, we need to transition to reliable logical qubits — created by combining multiple physical qubits together into logical ones to protect against noise and sustain a long (i.e., resilient) computation. … By having high-quality hardware components and breakthrough error-handling capabilities designed for that machine, we can get better results than any individual component could give us.”

Researchers will be able to get a taste of Microsoft’s reliable quantum computing via Azure Quantum Elements in the next few months, where it will be available as a private preview. The goal is to push even further to Level 3 quantum supercomputing, which will theoretically be able to tackle incredibly complex issues like climate change and exotic drug research. It’s unclear how long it’ll take to actually reach that point, but for now, at least we’re moving one step closer towards practical quantum computing.

This article contains affiliate links; if you click such a link and make a purchase, we may earn a commission.

[ad_2]

Source Article Link

Tags computing, finally, Microsoft, Quantum

Featured

Is the AI GPU the new mainframe? New open source tech allows users to ‘timeshare’ GPU resources for AI purposes for free — reminiscent of the days where scarce resources fosted computing elitism

[ad_1]

Without an efficient way to squeeze additional computing power from existing infrastructure, organizations are often forced to purchase additional hardware or delay projects. This can lead to longer wait times for results and potentially losing out to competitors. This problem is compounded by the rise of AI workloads which require a high GPU compute load.

ClearML has come up with what it thinks is the perfect solution to this problem – fractional GPU capability for open source users, making it possible to “split” a single GPU so it can run multiple AI tasks simultaneously.

This move recalls the early days of computing when mainframes could be shared among individuals and organizations, giving them the ability to utilize computing power without needing to buy additional hardware.

Fractional capabilities for Nvidia GPUs

ClearML says the new feature allows DevOps professionals and AI Infrastructure leaders to partition their Nvidia GTX, RTX, and data center-grade, MIG-enabled GPUs into smaller units to support multiple AI and HPC workloads, enabling users to switch between small R&D jobs and larger, more demanding training jobs.

The approach supports multi-tenancy, offering secure and confidential computing with hard memory limitation. ClearML says stakeholders can run isolated parallel workloads on a single shared compute resource, increasing efficiency and reducing costs.

“With our new free offering now supporting fractional capabilities for the broadest range of Nvidia GPUs than any other company, ClearML is democratizing access to compute as part of our commitment to help our community build better AI at any scale, faster,” says Moses Guttmann, CEO and Co-founder of ClearML. “We hope that organizations that might have a mixture of infrastructure are able to use ClearML and get more out of the compute and resources they already have.”

The new open source fractional GPU functionality is available for free on ClearML’s GitHub page.

ClearML Fractional GPU Utilization

(Image credit: ClearML)