We help customers spend less and travel better. Hopper combines its world-class travel agency offering with proprietary fintech products to save travelers money. We aspire to be the world's best - and most fun - place to book travel. Here are just a few stats that demonstrate the company’s recent growth: - Hopper’s valuation is now $5B. The company’s valuation has more than tripled since early 2020. - The company grew its revenue by more than 300% YoY in 2021. Hopper’s revenue run rate is now 30X higher than its pre-pandemic high point, and Hopper Cloud, Hopper’s new B2B initiative, is already 15% of its revenues and growing. - According to Apptopia, Hopper was the #1 most downloaded OTA in the US in 2021. The app has over 70 million downloads. - Hopper’s recovery is faster than the market, and its share of the air travel market in North America is now 300% higher than before the pandemic, according to MIDT. - The company sells over $2B in travel and travel-related financial services annually - Hopper’s fintech offerings, such as Price Freeze and Rebooking Guarantee, now represent over 70% of its air revenue. - Given the success of its travel fintech, Hopper recently launched a B2B initiative called Hopper Cloud, where any travel provider can unlock a unique revenue stream by seamlessly integrating and offering Hopper’s travel content and fintech products. Hopper Cloud partners include Capital One, Kayak, Marriott, Amadeus, Trip.com, and MakeMyTrip. - The industry opportunity for Hopper Cloud is enormous as projections estimate that if all travel distribution channels offered travel fintech, it could increase the total consumer spend for the sector by $200 billion annually. All this recent growth is just the beginning. There are still so many new opportunities we’re excited to tackle in the next year!

Hopper

Linux is one of the most popular operating systems (technically, Linux is just the Kernel) and has many types (distribution/distro). It may be unclear to choose which one to use, from beginner-friendly distro to those used for special needs. Linux users who use Linux as a daily driver usually have tried a lot of distros before sticking to one. Distro Hoppers Distro Hoppers is a term for users who often change distros regularly. The reason can vary: try new challenges, different features, and different purposes of the distro. Distro hopping is sometimes considered bad because it requires much time and offers little benefit. This article will discuss tools and advice for Distro Hoppers. Virtual Machine: Virtualbox/KVM The virtual Machine application is useful for trying the Linux distro before you even consider installing it or burning it on your flash disk. It is important to have a taste before installing it so you're not wasting so much time if you don't like the distro in the first place. Backups Backup is important to prevent data loss after installing a new Linux distribution. There are several ways. Using backup tools, e.g., , , etc. rsync Timeshift Separate partition for data, or separate partition for directory. With this approach, you must mount that partition after installing the new distribution. /home Save all data to the cloud. Several are out there, e.g., pCloud and MEGA. Just remember only to use a provider that has a Linux Client app. providers I recommend the last approach because saving important data on a local computer is not good practice. USB Bootable: Ventoy is an open-source tool to create a bootable USB drive for ISO/WIM/IMG/VHD(x)/EFI files. Using Ventoy, we can create live USB for multiple distros in one flash disk. This is a revolutionary feature compared to other tools, e.g., Etcher, Rufus, etc., that can only create one live USB at a time. This tool is so important. Don't call yourself a distro hopper if you don't know this tool. Ventoy Package Manager: Flatpak Hopping from one distro to another with a different package manager might require some time to adapt. Using a package manager that can be installed on most distro is one way to help you get to work faster. is one of them; other alternatives are , , or . Flatpak is a good starter, and if you have a bunch of free time, I suggest trying Nix. Flatpak Snap Nix Homebrew Development: Docker With DevContainer If you are using Linux, I think it's safe to assume that your job is somewhat related to programming. Setting up a development environment for a Software Engineer can be a hurdle. The development environment might require installing many tools, e.g., programming language, database, monitoring tools, etc. Docker with DevContainer is one of the fastest solutions to create a development environment faster, even after hopping from several distros. is a containerization tool that ships your app with all the necessary dependencies. Using makes developing software inside a container possible. Several editors, e.g., Visual Studio Code and JetBrains, already support DevContainer. If you use other editors like VIM or Helix, you might develop software directly inside the container or need to find another workaround. But if you are too lazy to configure VIM like me, use VSCode. Docker DevContainer Honorable Mentions Distrobox is a tool that enables us to try Linux distro CLI, including their package manager. This requires a containerization tool (e.g., Docker). In Windows, this can be achieved using WSL (Windows Subsystem for Linux) Distrobox EasyBCD is a Windows app that can modify the Bootloader if you dual-boot with Windows. I use this tool to remove Linux Grub when distro hopping. EasyBCD Conclusion Distro Hopping is not always bad, especially if you want to learn something new. Do this responsibly so it doesn't bother your work or your time with family. Utilize the tools above to minimize the time wasted, or even better if there is a separate computer just for that. OK, that's all for now. I hope you gain some insight and become better Distro Hoppers. Thank you. Also published here

Tools for Linux Distro Hoppers: A Beginner-Friendly Guide

As we honor Women’s History Month, it’s important to remember the influence women. Of all the women who contributed to the development of technology the way we have today, one woman stands out to me the most: . Grace Hopper (Amazing Grace) Who was Grace Hopper? Source Grace Hopper was born and raised in New York/New Jersey in 1906 where she spent most of her youth. She went onto earn her B.A in mathematics and physics from Vassar College in 1928, then pursue her masters and Ph.D. in mathematics from Yale University. She taught mathematics at Vassar College as an associate professor until she enlisted to join the U.S. Navy to help in the war effort during World War II. Although she was rejected multiple times before, the U.S. military was in need of brilliant minds for the war effort. So, she was eventually accepted by the U.S. Navy to serve in the ( omen ccepted for olunteer mergency ervice) which was part of the U.S. Navy Reserve. WAVES W A V E S After serving in the U.S. Navy Reserve until she was 60, she retired with a rank of . However, she was called into active duty multiple times afterward her first retirement gaining more accolades. commodore Grace Hopper retired for the final time with the rank of a in 1986 at the age of 79 and her service to the U.S. Navy spanned more than 42 years. This makes her one of the few female admirals in the U.S. Navy which is an astounding accomplishment of its own. Rear Admiral How is she important in tech? Aside from Grace Hopper’s huge accomplishments as a commissioned officer in the U.S. Navy, she had an immense contribution to the tech world. Prior to Grace Hopper, there was a strong belief that computers can only do computations like arithmetics. However, she thought different so she wanted to create a programming language written in English that a computer can read. This led to a development of the first called . She then released some of the first compiled programming languages like and . compiler A-0 MATH-MATIC FLOW-MATIC Source This was a huge leap in computer programming, but it was only the beginning. The next leap came with the programming language called ( mmon usiness- riented anguage) which was based on FLOW-MATIC . COBOL was quickly adapted into numerous businesses and Grace Hopper took a huge part in the initial development of the language and creating some of the standardization. COBOL CO B O L COBOL isn’t a programming language that most programmers nowadays learn or even know about. You may ask: Then, why do we talk about it? According to some research by Reuters, COBOL is responsible for 70–80% of the active code in production in the world with 220 billion lines of code, powering 95% of the ATM swipes, and 80% of the in-person transactions. Whether you believe in those stats or not, these are mind-boggling numbers which shows the impact that Grace Hopper has even in the modern world. Final Thought There are many women in history who had great influences across various fields other than technology. So, let’s take the time to honor the great women who came before us, living with us, and the ones to come. I want to hear some of your thoughts. Which woman/women in the past or present do you guys respect the most? To me programming is more than an important practical art. It is also a gigantic undertaking in the foundations of knowledge. — Grace Hopper

Women in Tech: Grace Hopper

Hey hackers! Did you know that HackerNoon runs a to gauge reader sentiment on topics of interest? That’s right, every Monday readers are greeted with a new poll curated by a HackerNoon editor on some of the most trending news on the web (plus some general, wacky ones too just for fun!). weekly poll Now all that’s great, but what if we told you that there’s a way to go a step further with these polls and truly analyze readers’ sentiment? Let’s look at the : current poll running on HackerNoon Poll #1 Lesson: How to vote? This week's poll focuses on the debate over the potential ban of ByteDance’s TikTok in the US over national security concerns. Is it an overreach by the US government or a step in the right direction? To vote, pick the option that best aligns with your thoughts and compare it with the broader reader population 😎 Poll #2 Lesson: How to check past polls? Click on the “ ” to access a list of all the polls HackerNoon has run in the past. Poll of the Week Now, select a poll that has already closed and analyze the results 🥳 Poll #3 Lesson: How do I access polls' templates? Let’s look at the “ ” poll: if you click on the poll, it should show you a prompt that you can use to start analyzing the results right away! Take a look: What is your web browser of choice Got it?! AWESOME 👏 Here are some of our favorite polls that are definitely worthy of your analysis: How Much Faith Do You Have in Your Government to Effectively Regulate AI? Will AI stocks suffer the same fate as the dot-com bubble? Where do YOU think bitcoin's price is heading by the end of the year? Do You Think the "Metaverse" Will Ever Truly Take Off? Other News Around The Web Nvidia GTC Ah, Nvidia. What can be said about the company that hasn’t been said already after it became the arms dealer of choice for the generative artificial intelligence boom. Well, apparently a lot! The company its annual GPU Technology Conference, or GTC, heralding the launch of the Blackwell class of AI chips that are supposed to be better than their Hopper predecessors. But what could these new processors mean for the development of more powerful AI and what does Nvidia’s change in business direction mean for consumers at large? recently held Let us know ! here TikTok’s Future in the US Should TikTok face a ban in the US? That’s the question we’re readers this week as we highlight this important issue. The US House recently passed a bill requiring ByteDance, the parent company of TikTok, to divest from the app within 180 days or face prohibition (a ban) in the U.S. However, Senate supporters argue that this is not a ban but rather a transition of ownership to ensure national security. Is this overreaction, and perhaps, an overreach by the US government? asking and share your thoughts . Vote in the Poll here Sam Bankman-(never getting)Freed Looks like Sam Bankman-Fried’s days as a billionaire and crypto entrepreneur are over (not that there was ever a chance he would be able to dazzle in crypto ever again anyway). The once wildly successful owner of crypto exchange FTX learned this past week what financial mismanagement gets you; a quarter of a century in prison. SBF’s sentencing, which he’s going to appeal, apparently also saw a wild increase in the prices of memecoins (we’re talking thousands of %age up). In light of what’s happened, what do you of the sentencing and how do you in 2024? think feel about cryptocurrencies Optout Writing Contest HackerNoon’s partnership with Aut Labs is still going strong and we’ve seen some great entries these past few weeks! If you’d like to take a stab at thousands of dollars in USDT, simply head on down to and start writing on the Contest and tell us how YOU are opting out of a system that does not represent you. winning our template here #Optout Happy writing!

How to Use HackerNoon Polls to Your Advantage 🤔

আপনার সুবিধার জন্য হ্যাকারনুন পোল কীভাবে ব্যবহার করবেন 🤔

So nutzen Sie HackerNoon-Umfragen zu Ihrem Vorteil 🤔

Cómo utilizar las encuestas de HackerNoon a su favor 🤔

Comment utiliser les sondages HackerNoon à votre avantage 🤔

HackerNoon पोल का अपने लाभ के लिए उपयोग कैसे करें 🤔

HackerNoon のアンケートを有効活用する方法 🤔

HackerNoon 설문조사를 유리하게 활용하는 방법 🤔

Como usar as pesquisas HackerNoon a seu favor 🤔

Как использовать опросы HackerNoon в своих интересах 🤔

HackerNoon Anketlerini Avantajınıza Nasıl Kullanabilirsiniz 🤔

Cách sử dụng HackerNoon Polls để mang lại lợi ích cho bạn 🤔

如何利用 HackerNoon 民意调查为自己谋利 🤔

If you ever needed proof that is no longer the company you think it is, look no further than the company's most recent GPU Technology Conference (also known as the GTC). The conference was chock-full of announcements, but the big one was of course the unveiling of the Blackwell class of AI chips which are supposedly going to be better than the widely-in-demand, and equally as expensive, Hopper processors that have powered the likes of OpenAI's ChatGPT. Nvidia It was the demand for the H100 chips that transformed Nvidia into a company worth over $2 trillion, almost overnight in the larger scheme of things, leading some (including HackerNoon) to whether the stock was in a bubble (almost a quarter of HackerNoon readers thought that it was). But, as if Nvidia was out to prove everyone wrong, the company's last earnings release showed that not only was it still selling the Hopper processors, it was selling of them and would in the near future. question more continue to do so Which brings us to the Nvidia GTC. Nvidia Chief Executive Jensen Huang is no Steve Jobs, but he is certainly a shareholders' man, and he knows how to keep them happy – presentation skills be damned. The Blackwell announcement may be nothing akin to Apple's unveiling of the very first iPhone, but boy did Wall Street go crazy. And with crazy, comes worry. Worry about a monopoly. Worry about supply. Worry that the company may not be producing those darn AI processors fast enough. But for Nvidia, it is COVID-19 all over again: demand is up, consumers are desperate, and the company can charge whatever price it wants for its hardware. , ladies and gentlemen, is the future tech companies had hoped for when they began aggressively hiring during the worldwide pandemic only to find demand normalizing once the world opened up. This Which makes Nvidia quite possibly the only company to not only have profiteered off of consumers during the pandemic but to have hit an even bigger jackpot: the enterprise. Because face it, the $40,000 H100 chips aren't for everyday consumers — they're for the faceless corporations that can dish out the billions needed to train large language models. And therein lies Nvidia's success. You see, just a day after the GTC began, Jensen Huang held a with investors to tell them the company was poised to win a large chunk of the $250 billion corporations spend on building and maintaining data centers every year given the variety of hardware and software Nvidia has on offer. meeting Now all of this is well and good for the company, but there's one stakeholder that is hopelessly missing from all of Nvidia's future plans: the gamers. Nvidia never really started an artificial intelligence company the way OpenAI and so many others like it did. It started out as a company that produced graphic cards for video games, and for a long, long time, this is what Nvidia was known for. But as an offshoot of all the research it did in producing better and better graphic cards, it ended up creating tensor cores for real-time deep learning image enhancement and upscaling technologies for use in video games. Little did Nvidia know that this technology would then serve as the backbone of LLMs, instantly transforming it as the arms dealer of choice in corporations' pursuit of the perfect AI. as Nvidia may very well be out of touch with its roots, but that's hardly what shareholders care about. And if Nvidia shareholders don't care about the gamers, then Nvidia doesn't care about the gamers. This is why, a conference that have been about the launch of the company's next graphic card was instantly hijacked by talk of AI, because it's 2024, and AI, for all intents and purposes, sells. should Some may call it the ' ', but I'll call it what it is: the death knell of Nvidia's days as a gaming-focused company. Woodstock of AI 👋 You’re reading HackerNoon's , a weekly collection of tech goodness that combines HackerNoon's proprietary data with internet trends to determine which companies are rising and falling in the public consciousness. Subscribe to receive the newsletter in your inbox every Tuesday. Tech Company News Brief here complete Microsoft Pulls The Corporate Espionage of a Lifetime How do you kill two birds with one stone? I don't know. Ask Satya Nadella. So laser-focused is the chief executive on monopolizing his company's success in AI, that he has decided to do the unthinkable: hire its competitor's staff and hand them the reins of a newly launched division focused primarily on artificial intelligence. Microsoft I am of course referring to Mustafa Suleyman and Karen Simonyan, the co-founders of AI startup Inflection AI, which, up until recently, should have been competing against Microsoft's ward OpenAI but found itself with an offer (I assume) they could not refuse. Yes, the very same Mustafa Suleyman who co-founded DeepMind Technologies and sold it to in 2014, will be leading Microsoft AI as its chief executive with the help of Karen Simonyan as the division's chief scientist. Google Microsoft also seems to have hired "several members" of the Inflection team who have chosen to join Suleyman and Simonyan to advance the company's Copilot and other consumer AI products and research. So what does Inflection AI get in return? A whopping , apparently, and the prestige of working with Microsoft. $650 million That's a lot of money to pay to get your competitor out of the way keep OpenAI, and by extension, Sam Altman, , but as the world's most valuable company right now, Microsoft can probably get away with just about anything. and in check Microsoft ranked #6 on HackerNoon's this week. Google was on the #13 spot. Tech Company Rankings Apple Pulls the Plug on In-House MicroLED Displays First the , now the smartwatch displays. electric car In perhaps what is a period of realignment for , the company has on building its own smartwatch displays due to the cost and complexity involved. Apple pulled the plug The microLED displays were due to be introduced in a future version of the Apple Watch and were part of a push to lower the company's reliance on partners, but it seems the company is no longer planning to design and develop that technology in house, maybe focusing instead on the newly launched (and confusingly reviewed) Apple Vision Pro(s). As with the electric car, the decision to pull the plug will result in the elimination of job roles at the company. Apple ranked #3 on HackerNoon's this week. Tech Company Rankings In Other News.. 📰 SEC Seeks $1.95B Fine in Final Judgment Against Ripple — via CoinDesk AI is a data problem. Cyera is raising up to $300M on a $1.5B valuation to secure it — via TechCrunch 3 ways Apple’s monopoly lawsuit could change the iPhone experience for fans — via CNN Musk's X Corp loses lawsuit against hate speech watchdog — via Reuters Meta's broad-brush ban on Islamic term harmed free speech, board rules — via Axios Adam Neumann makes a $500 million bid for WeWork that could hit $900 million if financing and diligence firm up — via CNBC And that's a wrap! Don't forget to share this newsletter with your family and friends! See y'all next week. PEACE! ☮️ — Sheharyar Khan, Editor, Business Tech @ HackerNoon *All rankings are current as of Monday. To see how the rankings have changed, please visit HackerNoon's page. Tech Company Rankings

Nvidia Throws Gamers Under the Bus

Hey everyone! and just like you, I’ve been fascinated with the recent progress of artificial intelligence. Realizing that I needed to stay abreast with all the developments happening, I decided to embark on a personal journey of learning, thus was born! With this series, I will be learning about LLMs and share ideas, experiments, opinions, trends & learnings through my blog posts. You can follow along the journey on HackerNoon or my personal website . In today’s article, we’ll be talking about take aways from Nvidia’s , I’m Nataraj 100 days of AI here here If AI is a gold rush, Nvidia is the ultimate picks and shovels company. All things AI go through Nvidia because of its dominance as the sole company that creates the chips (GPUs) needed to deploy the computationally expensive large language models. As the models scale, it increases the need for more powerful GPUs and the entire world is relying on Nvidia to deliver on them. Its not an exaggeration to say Nvidia is at the core of all things AI and they recently had their semi-annual developer conference GTC (March 18-21). In this post I am going to summarize my key takeaways from this very special GTC conference and what it means for the gen AI space. Takeaway 1: AI is not limited to the tech sector The power of what gen AI can do and deliver has not been lost on non-tech sectors. Companies across all the sectors in the economy think that there is a need for them to adopt AI and find a way to partner, execute and stay ahead. The proof here is looking at how many non-tech sector companies have partnered with Nvidia. Companies from retail, automobile, construction, design and everything else have announced partnerships with Nvidia. Takeaway 2: We need much bigger GPUs Ever since transformers were invented, we are doubling the scale of LLMs every 6 months. For example, look at the parameter count of Open AI models below. In GPT-3 series alone the parameter count ranged from 125M to 175B. And the latest models expected to come out of Open AI are rumored to have parameter counts above trillion and soon could hit trillions in less than two years, if the rate of scaling continues. To support the increasing scale of LLMs we need increased computational power. To capture the market, Nvidia is launching a new series of GPUs called as Blackwell, named after statistician & mathematician David Blackwell. Blackwell is a new series of GPUs (GB200) but also a new super computer platform. Here’s a performance comparison of the new Blackwell GPU with its previous generation GPU Hopper. Takeaway 3: Moving from retrieval to generation In the current generation of computing, everything is about retrieving a piece of data stored in different formats and presenting it in a way that is useful to the user. But in the coming generation of computing and apps, we are about to see more of generating things at the user’s request and giving it back to the user. In case of chat-gpt for example, the answer you are getting is not stored in some database before hand, but is being generated in real time to serve the user’s question. We are about to see the generation of text, images, videos, chemicals, proteins and more. Takeaway 4: Inference is getting better Inferencing is when you ask chat-gpt a question, the model is figuring out the answer in the form of token generation and returning the answer in the form of tokens. If inference is not fast enough, then the consumers will not be able to use any of the AI applications. Inference is especially tricky because LLMs are getting larger and larger and they don’t fit in one GPU, so parallelizing the inference across GPUs is a difficult task and involves programming Nvidia GPUs to use multiple optimization techniques. With Blackwell GPUs, Nvidia is able to achieve 30x faster inference speeds. Takeaway 5: The real metaverse is NVidia’s Omniverse As I learnt more and more about Omniverse by Nvidia I am more and more bullish on the product and the idea. This is because to really leverage AI and unlock all the possibilities in automation and robotics, we actually do not have all the data needed. For example, say you want to create a robot which cooks great food of all kinds. There is no data where you have first person view of chefs cooking various kinds of dishes. We have general data of various cooking in the form of YouTube videos, but to get the robots learn the physical movements involved in cooking a first person data is needed. A 3d simulation of the world in Omniverse can help bridge that gap of data needed for unlocking all these use cases. I am extremely bullish on the idea. That’s it for Day 22 of 100 Days of AI. I write a newsletter called Above Average where I talk about the second order insights behind everything that is happening in big tech. If you are in tech and don’t want to be average, . subscribe to it Follow me on , or for latest updates on 100 days of AI or . If you are in tech you might be interested in joining my community of tech professionals . Twitter LinkedIn HackerNoon bookmark this page here

100 Days of AI, Day 16: 5 Key Takeaways From NVIDIA’s AI Developer Event

AI এর 100 দিন, দিন 16: 5 NVIDIA এর aI ডেভেলপার ইভেন্ট থেকে মূল টেকওয়ে

100 Tage KI, Tag 16: 5 wichtige Erkenntnisse aus dem AI-Entwicklerevent von NVIDIA

100 días de IA, día 16: 5 conclusiones clave del evento para desarrolladores de IA de NVIDIA

100 jours d'IA, jour 16 : 5 points clés à retenir de l'événement pour développeurs IA de NVIDIA

AI के 100 दिन, दिन 16: NVIDIA के aI डेवलपर इवेंट से 5 मुख्य बातें

AI の 100 日間、16 日目: NVIDIA の aI 開発者イベントから得た 5 つの重要なポイント

100일간의 AI, 16일차: NVIDIA AI 개발자 이벤트의 5가지 주요 시사점

100 dias de IA, dia 16: 5 principais conclusões do evento de desenvolvedores de IA da NVIDIA

100 дней искусственного интеллекта, день 16: 5 ключевых выводов с мероприятия NVIDIA для разработчиков искусственного интеллекта

100 Gün Yapay Zeka, 16. Gün: NVIDIA'nın AI Geliştirici Etkinliğinden 5 Temel Çıkarım

100 Ngày của AI, Ngày 16: 5 bài học quan trọng từ Sự kiện dành cho nhà phát triển AI của NVIDIA

AI 100 天，第 16 天：NVIDIA 人工智能开发者活动的 5 个要点

Nvidia's annual developer conference in 2024, known as GTC (GPU Technology Conference), was a pivotal event that unveiled groundbreaking advancements in artificial intelligence (AI) chips, software tools, and strategic partnerships. The conference, led by Nvidia's CEO Jensen Huang, showcased the company's commitment to innovation and leadership in the tech industry. And if you want to understand the latest trends from the world of AI - subscribe to my newsletter ‘ ‘. And the best part - it’s completely free! AI Hunters Now, onto the key takeaways from Nvidia's GTC 2024: Flagship AI Chip - B200 "Blackwell" The Flagship AI Chip - B200 "Blackwell" unveiled by Nvidia at GTC 2024 represents a groundbreaking advancement in AI chip technology. Named after mathematician David Blackwell, this new chip is a significant leap forward in performance, being 30 times faster than its predecessor, the Hopper series, particularly excelling at tasks like supporting advanced chatbots and large language models. With an impressive 208 billion transistors, more than double the previous chip, the B200 chip offers exceptional processing power. Despite its remarkable performance, the B200 chip is designed with efficiency in mind, using 25 times less energy and money than the previous Hopper series. This focus on efficiency not only enhances its cost-effectiveness but also positions it as a sustainable solution for AI applications. The innovative design of the B200 chip, which combines two squares of silicon into a single chip, allows all transistors to access memory simultaneously, further boosting productivity and performance. Nvidia's dominance in the data center AI chip market, with a market share of roughly 80%, underscores the significance of the B200 chip in shaping the future of AI technology. Major tech giants like Amazon, Google, Meta Platforms, Microsoft, OpenAI, and Tesla are expected to adopt this new chip, highlighting its importance and impact in the industry. Expected to hit the market later in 2024, the B200 chip offers advanced capabilities for AI applications. With its powerful performance, energy efficiency, and anticipated industry-wide adoption, the Blackwell B200 chip is poised to redefine AI processing and drive innovation across various sectors. Points to Reflect Upon: Performance Leap: Consider the implications of the B200 chip's 30x performance increase for AI applications and industries. Assess the feasibility of integrating the B200 chip into existing AI infrastructure for enhanced performance. Efficiency Benefits: Conduct a cost-benefit analysis to determine the potential savings from transitioning to the B200 chip. Software Tools for AI Integration Nvidia Inference Microservices (NIM) introduced at GTC 2024 represent a significant advancement in AI deployment technology. NIM is a cloud-native microservices solution that simplifies the development and deployment of generative AI applications across various infrastructures, including cloud, data centers, and GPU-accelerated workstations. It offers a streamlined path for developing AI-powered enterprise applications and deploying AI models in production, aiming to shorten time-to-market and simplify deployment processes for organizations transitioning to full-scale production deployments. NIM is designed to bridge the gap between the complexities of AI model development and the operational needs of enterprises, enabling a broader pool of developers to contribute to AI transformations within their companies. Some core benefits of NIM include its portability and control, allowing model deployment across various infrastructures, from local workstations to cloud environments. It provides prebuilt containers and Helm charts with optimized models validated across different NVIDIA hardware platforms, cloud service providers, and Kubernetes distributions, ensuring support across all NVIDIA-powered environments. Developers can access AI models through industry-standard APIs, simplifying the development of AI applications and enabling swift updates within the ecosystem. NIM also addresses the need for domain-specific solutions by packaging domain-specific NVIDIA CUDA libraries and specialized code tailored to various domains like language, speech, video processing, healthcare, and more. Leveraging optimized inference engines for each model and hardware setup, NIM ensures the best possible latency and throughput on accelerated infrastructure, reducing the cost of running inference workloads as they scale and enhancing the end-user experience. Part of the NVIDIA AI Enterprise suite, NIM is built with an enterprise-grade base container, providing a solid foundation for enterprise AI software through rigorous validation, enterprise support, and regular security updates. With support for various AI models, including community models, NVIDIA AI Foundation models, and custom AI models, NIM simplifies the AI model deployment process by packaging algorithmic, system, and runtime optimizations and adding industry-standard APIs. This comprehensive solution accelerates the deployment of scalable and customized AI applications in production, making AI more accessible and customizable across industries. Points to Reflect Upon: Business Impact: Consider how NIM can empower businesses to harness AI capabilities and drive innovation. Operational Efficiency: Evaluate how enhanced system efficiency can optimize business processes and decision-making. Strategic Partnerships and Collaborations Nvidia's expanded partnerships with industry leaders and software makers announced at GTC 2024 mark a significant milestone in the development of AI technologies and their integration into various sectors. These collaborations highlight Nvidia's strategic efforts to advance AI infrastructure, enhance AI capabilities, and drive innovation across industries. One key partnership involves Microsoft, where Nvidia and Microsoft are working together to bring the power of Nvidia Grace Blackwell GB200 and advanced Nvidia technologies to Azure. This collaboration aims to deliver trillion-parameter foundation models for tasks like natural language processing, computer vision, speech recognition, and more. Additionally, Microsoft is introducing the Azure NC H100 v5 VM virtual machine based on the Nvidia H100 NVL platform to cater to midrange training and inferencing needs. Furthermore, Nvidia's inference microservices are being integrated into Azure, enhancing AI capabilities for users by providing cloud-native microservices for optimized inference on various foundation models. This integration streamlines AI deployments and accelerates the development of performance-optimized production AI applications. In the automotive sector, Nvidia has deepened its ties with leading Chinese automakers like BYD, Xpeng, GAC Aion's Hyper brand, Zeekr, and Li Auto. These partnerships focus on integrating Nvidia's cutting-edge technologies, such as the Drive Thor in-vehicle chips, to enhance autonomous driving capabilities and infotainment systems in vehicles. Collaborations with US-based software company Cerence aim to adapt large language model AI systems for in-car applications, promising advanced voice command systems and improved user experiences. Overall, these partnerships unveiled at GTC 2024 demonstrate Nvidia's commitment to driving innovation across various industries by leveraging cutting-edge technologies and fostering collaborations with industry leaders and software makers. Points to Reflect Upon: Industry Impact: Consider how partnerships with key players can influence the adoption and integration of Nvidia's technology in various sectors. Innovation Potential: Reflect on the potential for collaborative efforts to drive innovation and create new opportunities in AI and technology. Focus on Sustainability and Sovereign AI Nvidia's emphasis on sustainability and sovereign AI initiatives at GTC 2024 underscores the company's commitment to responsible technology development. The focus on addressing environmental concerns related to AI training and data center energy consumption demonstrates Nvidia's proactive stance on sustainability issues. By prioritizing energy efficiency and sustainable practices in AI development, Nvidia is setting a standard for responsible innovation in the tech industry. Furthermore, Nvidia's dedication to sovereign AI initiatives, which enable nations to produce AI using their infrastructure and workforce, highlights the company's commitment to promoting ethical and secure AI practices globally. By empowering countries to develop AI capabilities locally, Nvidia is fostering digital sovereignty and supporting economic growth while ensuring data protection and security. Overall, Nvidia's dual focus on sustainability and sovereign AI at GTC 2024 showcases the company's holistic approach to technology development, balancing innovation with ethical considerations and environmental responsibility. This strategic emphasis on sustainability and sovereignty sets Nvidia apart as a leader in responsible AI practices and underscores its commitment to shaping a more sustainable and secure future for AI technologies. Points to Reflect Upon: Environmental Responsibility: Consider the implications of Nvidia's sustainability efforts on the tech industry and environmental conservation. Develop a sustainability strategy to align with Nvidia's initiatives. Ethical AI Practices: Reflect on the importance of promoting ethical AI practices and ensuring transparency in AI development. Global Engagement: Explore opportunities for global engagement to promote sovereign AI practices and support technological independence. Market Position and Future Outlook Nvidia's strong market position and future outlook were prominently highlighted at GTC 2024, emphasizing the company's leadership in GPU technology and AI innovation amidst competition from rivals like AMD, Intel, and emerging startups. Announcements at GTC 2024 regarding new products such as the Blackwell GPU series, DGX GB200 System, and DGX SuperPOD underscore Nvidia's focus on advancing AI infrastructure and enabling next-generation AI models. This strategic vision positions Nvidia at the forefront of technological advancements that will shape the future of AI computing and high-performance computing (HPC). Overall, Nvidia's market position and future outlook at GTC 2024 highlight its resilience, innovation drive, and strategic vision for continued growth and leadership in the evolving landscape of AI technologies. To sum up, Nvidia's annual developer conference in 2024 was a platform for unveiling cutting-edge AI technology, software tools, and strategic partnerships that reinforce the company's position as a leader in the tech industry. The key takeaways from GTC 2024 underscore Nvidia's commitment to innovation, sustainability, and collaboration, setting the stage for continued advancements in AI, GPU technology, and high-performance computing. P.S. Check out some of my previous articles at HackerNoon: Google Introduces Gemma: Revolutionary Access to Pre-Trained AI Models Sora's AI Innovation Set to Disrupt Video Content Creation Breaking Down the Next Big Thing in NFTs: ERC-404 Spotlight Gemini 1.5 Unleashes Unprecedented Context for AI Applications The AI Revolution in Journalism: A New Era of Enhanced Reporting DeFi Community Building: A Step-by-Step Guide for Crypto Startups

Nvidia GPU Technology Conference In 2024: A Deep-Dive 

A chain is as strong as its weakest link - and your AI/ML infrastructure is only as fast as your slowest component. If you train machine learning models with GPUs, then your weak link may be your storage solution. The result is what I call the “Starving GPU Problem.”  The Starving GPU problem occurs when your network or your storage solution cannot serve training data to your training logic fast enough to fully utilize your GPUs. The symptoms are fairly obvious. If you are monitoring your GPUs, then you will notice that they never get close to being fully utilized. If you have instrumented your training code, then you will notice that total training time is dominated by IO. Unfortunately, there is bad news for those who are wrestling with this issue. Let’s look at some advances being made with GPUs to understand how this problem is only going to get worse in the coming years. GPUs Are Getting Faster GPUs are getting faster. Not only is raw performance getting better, but memory and bandwidth are also increasing. Let’s take a look at these three characteristics of Nvidia’s most recent GPUs the , the and the . A100 H100 H200 GPU PERFORMANCE MEMORY MEMORY BANDWIDTH A100 624 TFLOPS 40GB 1,555GB/s H100 1,979 TFLOPS 80GB 3.35TB/s H200 1,979 TFLOPS 141GB 4.8TB/s The table above uses the statistics that align with a PCIe (Peripheral Component Interconnect Express) socket solution for the A100 and the SXM (Server PCI Express Module) socket solution for the H100 and the H200. SXM statistics do not exist for the A100. With respect to performance, the Floating Point 16 Tensor Core statistic is used for the comparison. A few observations on the statistics above are worth calling out. First, the H100 and the H200 have the same performance (1,979 TFLOPS), which is 3.17 times greater than the A100. The H100 has twice as much memory as the A100 and the memory bandwidth increased by a similar amount - which makes sense otherwise, the GPU would starve itself. The H200 can handle a whopping 141GB of memory and its memory bandwidth also increased proportionally with respect to the other GPUs. Let’s look at each of these statistics in more detail and discuss what it means to machine learning. - A teraflop (TFLOP) is one trillion (10^12) floating-point operations per second. That is a 1 with 12 zeros after it (1,000,000,000,000). It is hard to equate TFLOPs to IO demand in gigabytes as the floating point operations that occur during model training involve simple tensor math as well as first derivates against the loss function (a.k.a. gradients). However, relative comparisons are possible. Looking at the statistics above, we see that the H100 and the H200, which both perform at 1,979 TFLOPS, are 3 times faster - potentially consuming data 3 times faster if everything else can keep up. Performance - Also known as Video RAM or Graphics RAM. The GPU memory is separate from the system's main memory (RAM) and is specifically designed to handle the intensive graphical processing tasks performed by the graphics card. GPU memory dictates batch size when training models. In the past batch size was decreased when moving training logic from a CPU to a GPU. However, as GPU memory catches up with CPU memory in terms of capacity, the batch size used for GPU training will increase. When performance and memory capacity increase at the same time, the result is larger requests where each gigabyte of training data is getting processed faster. GPU Memory - Think of GPU memory bandwidth as the "highway" that connects the memory and computation cores. It determines how much data can be transferred per unit of time. Just like a wider highway allows more cars to pass in a given amount of time, a higher memory bandwidth allows more data to be moved between memory and the GPU. As you can see, the designers of these GPUs increased the memory bandwidth for each new version proportional to memory; therefore, the internal data bus of the chip will not be the bottleneck. Memory Bandwidth A Look into the Future In August 2023, Nvidia its next-generation platform for accelerated computing and generative AI - The GH200 Grace Hopper Superchip Platform. The new platform uses the , which can be connected with additional Superchips by , allowing them to work together during model training and inference. announced Grace Hopper Superchip NVIDIA NVLink While all the specifications on the Grace Hopper Superchip represent an improvement over previous chips, the most important innovation for AI/ML engineers is its unified memory. Grace Hopper gives the GPU full access to the CPU’s memory. This is important because, in the past, engineers wishing to use GPUs for training had to first pull data into system memory and then from there, move the data to the GPU memory. Grace Hopper eliminates the need to use the CPU’s memory as a bounce buffer to get data to the GPU. The simple comparison of a few key GPU statistics as well as the capabilities of Grace Hopper, has got to be a little scary to anyone responsible for upgrading GPUs and making sure everything else can keep up. A storage solution will absolutely need to serve data at a faster rate to keep up with these GPU improvements. Let’s look at a common solution to the hungry GPU problem. A Common Solution There is a common and obvious solution to this problem that does not require organizations to replace or upgrade their existing storage solution. You can keep your existing storage solution intact so that you can take advantage of all the enterprise features your organization requires. This storage solution is most likely a Data Lake that holds all of your organization’s unstructured data - therefore, it may be quite large, and the total cost of ownership is a consideration. It also has a lot of features enabled for redundancy, reliability and security, all of which impact performance. What can be done, however, is to set up a storage solution that is in the same data center as your compute infrastructure - ideally, this would be in the same cluster as your compute. Make sure you have a high-speed network with the best storage devices available. From there, copy only the data needed for ML training. Amazon’s recently announced exemplifies this approach. It is a bucket type optimized for high throughput and low latency and is confined to a single Availability Zone (no replication). Amazon’s intention is for customers to use it to hold a copy of data that requires high-speed access. Consequently, it is purpose-built for model training. According to Amazon, it provides 10x the data access speed of S3 Standard at 8x the cost. Read more about our assessment of Amazon S3 Express One Zone . Amazon S3 Express One Zone here The MinIO Solution The common solution I outlined above required AWS to customize its S3 storage solution by offering specialty buckets at an increased cost. Additionally, some organizations (that are not MinIO customers) are buying specialized storage solutions that do the simple things I described above. Unfortunately, this adds complexity to an existing infrastructure since a new product is needed to solve a relatively simple problem. The irony to all this is that MinIO customers have always had this option. You can do exactly what I described above with a new installation of MinIO on a high-speed network with NVMe drives. MinIO is a software-defined storage solution - the same product runs on bare metal or the cluster of your choice using a variety of storage devices. If your corporate Data Lake uses MinIO on bare metal with HDDs and it is working fine for all of your non-ML data - then there is no reason to replace it. However, if the datasets used for ML require faster IO because you are using GPUs, then consider the approach I outlined in this post. Be sure to make a copy of your ML data for use in your high-speed instance of MinIO - a gold copy should always exist in a hardened installation of MinIO. This will allow you to turn off features like replication and encryption in your high-speed instance of MinIO, further increasing performance. Copying data is easy using MinIO’s feature. mirroring MinIO is capable of the performance needed to feed your hungry GPUs - a achieved 325 GiB/s on GETs and 165 GiB/s on PUTs with just 32 nodes of off-the-shelf NVMe SSDs. recent benchmark Download today and learn just how easy it is to build a data lakehouse. If you have any questions be sure to reach out to us on ! MinIO Slack Also appears . here

Hungry GPUs Need Fast Object Storage

ক্ষুধার্ত জিপিইউগুলির দ্রুত অবজেক্ট স্টোরেজ প্রয়োজন

Hungrige GPUs benötigen schnellen Objektspeicher

Las GPU hambrientas necesitan un almacenamiento de objetos rápido

Les GPU gourmands ont besoin d'un stockage d'objets rapide

भूखे GPU को तेज़ ऑब्जेक्ट स्टोरेज की आवश्यकता होती है

飢餓状態の GPU には高速オブジェクト ストレージが必要

배고픈 GPU에는 빠른 개체 스토리지가 필요합니다

GPUs famintas precisam de armazenamento rápido de objetos

Голодным графическим процессорам требуется быстрое объектное хранилище

Aç GPU'ların Hızlı Nesne Depolamaya İhtiyacı Var

GPU đói cần lưu trữ đối tượng nhanh

饥饿的 GPU 需要快速的对象存储

EFFICIENCY AND CAPACITY OF BOILERS

HOPPER

hopper.com

1K-5K employees

Since 2007

COMPANY RANKING

HOPPER

EVERGREEN INDEX #155

Representing this company? Click here to claim and customize this page!

Hopper WIKI

Apply to get your Tech Company News Page live on HackerNoon.com today!

HOPPER

hopper.com

1K-5K employees

Since 2007

COMPANY RANKING

HOPPER

EVERGREEN INDEX #155

RELATED COMPANIESRANK

STORIES ABOUT HOPPER

Representing this company? Click here to claim and customize this page!

Hopper WIKI

Apply to get your Tech Company News Page live on HackerNoon.com today!